This thesis has been submitted in fulfilment of the requirements for a postgraduate degree

(e.g. PhD, MPhil, DClinPsychol) at the University of Edinburgh. Please note the following
terms and conditions of use:

• This work is protected by copyright and other intellectual property rights, which are
retained by the thesis author, unless otherwise stated.
• A copy can be downloaded for personal non-commercial research or study, without
prior permission or charge.
• This thesis cannot be reproduced or quoted extensively from without first obtaining
permission in writing from the author.
• The content must not be changed in any way or sold commercially in any format or
medium without the formal permission of the author.
• When referring to this work, full bibliographic details including the author, title,
awarding institution and date of the thesis must be given.
LIST OF ACRONYMS USED.................................................................................................... 4
CHAPTER 1: INTRODUCTION ................................................................................................ 6
1.1 INTRODUCTION .................................................................................................................. 6
1.2 HISTORICAL ACCOUNTS OF PUBLIC HEALTH ......................................................................... 6
1.2.1 Historical accounts of post-war public health .......................................................... 6
1.2.2 Ideas of the demographic or epidemiological transition ........................................ 28
1.3 HISTORICAL ACCOUNTS OF NUTRITION SCIENCE................................................................. 31
1.4 SOCIAL SCIENCE ACCOUNTS OF OBESITY........................................................................... 48
1.5 SOCIOLOGICAL ACCOUNTS OF THE ‘NEW PUBLIC HEALTH’ ................................................... 57
1.6 THEORETICAL APPROACHES FROM SCIENCE AND TECHNOLOGY STUDIES (STS) ................. 59
1.6.1 The Sociology of Scientific Knowledge, social interests and paradigms............... 60
1.6.2 Regulatory science and boundary-work ................................................................ 63
1.6.3 Previous STS case studies of nutrition science..................................................... 66
1.6.4 Defining obesity science ........................................................................................ 68
1.7 THE MAKING OF HEALTH POLICY ........................................................................................ 69
1.8 METHODOLOGICAL ISSUES ............................................................................................... 70
1.8.1 Narrative of an evolving methodology ................................................................... 70
1.8.2 Use of Documentary Sources ................................................................................ 75
1.8.3 Selection of Sources .............................................................................................. 77
1.8.4 Issues in studying the WHO .................................................................................. 79
1.9 CHILDHOOD OBESITY ....................................................................................................... 81
1.10 PRACTICAL ISSUES AND LIMITATIONS .............................................................................. 83
1.11 CONCLUSION ................................................................................................................. 83
CHAPTER 2: HEART DISEASE AND THE EMERGENCE OF RISK FACTOR RESEARCH
C.1945-1980 ............................................................................................................................ 85
2.1 INTRODUCTION ................................................................................................................ 85
2.2 THE EPIDEMIOLOGICAL TRANSITION AND THE EPIDEMIC OF HEART DISEASE ......................... 85
2.3 INVESTIGATING THE CAUSES OF CHRONIC DISEASE ............................................................ 88
2.4 THE DEVELOPMENT OF RISK FACTOR RESEARCH ................................................................ 90
2.5 THE AMERICAN HEART LOBBY AND THE FRAMINGHAM HEART STUDY .................................. 91
2.6 RISK FACTOR RESEARCH AND CAUSAL EXPLANATION ......................................................... 96
2.7 HAS THERE BEEN AN ‘EPIDEMIC’ OF CORONARY HEART DISEASE IN THE UK? ..................... 101
2.8 CONCLUSION ................................................................................................................. 103
CHAPTER 3: THE DEVELOPMENT OF PUBLIC HEALTH COALITIONS 1950 - 1999 .... 105
3.1 INTRODUCTION .............................................................................................................. 105
3.2 THE DEVELOPMENT OF AN ANTI-SMOKING PUBLIC HEALTH COALITION IN THE UK................ 105
3.3 THE EARLY YEARS OF THE OBESITY COALITION ................................................................ 110
3.4 THE COALITION’S INITIAL MODEL OF OBESITY: THE INDIVIDUAL PARADIGM .......................... 117
3.4.1 Defining and measuring obesity .......................................................................... 117
3.4.2 Estimating the prevalence of obesity in the population ....................................... 120
3.4.3 The health consequences of obesity ................................................................... 121
3.4.4 The causes of obesity .......................................................................................... 123
3.4.5 Treatments for obesity ......................................................................................... 126
3.4.6 Costs to healthcare systems................................................................................ 130
3.5 THE OBESITY COALITION GOES MAINSTREAM ................................................................... 133
3.5.1 Mapping the usage of the term ‘obesity epidemic’ .............................................. 133
3.5.2 Contemporary British obesity organisations ........................................................ 136
3.5.3 Principal sources for chapters 4, 5 and 6 ............................................................ 138
3.6 CONCLUSION ................................................................................................................. 143
CHAPTER 4: THE DEVELOPMENT OF THE BODY MASS INDEX 1972-1995................. 145
4.1 INTRODUCTION .............................................................................................................. 145
4.2 THE MEASUREMENT OF BODY FAT ................................................................................... 145
4.3 ANCEL KEYS AND GEORGE BRAY REDEFINE W/H²........................................................... 148
4.4 THE BRITISH ADOPTION OF BMI: GARROW ’S DEFINITIONS ................................................ 150
4.5 THE BRITISH ADOPTION OF BMI: OTHER AUTHORS USE GARROW ’S DEFINITIONS ............... 152

1
 4.6 THE RELATIONSHIP BETWEEN BMI AND MORTALITY .......................................................... 158
4.7 PREVALENCE AND INCIDENCE: ESTIMATING THE EXTENT OF OBESITY AND OVERWEIGHT ..... 163
4.8 CONTEMPORARY CRITICISMS OF THE USEFULNESS OF BMI .............................................. 168
4.9 CONCLUSION ................................................................................................................. 171
CHAPTER 5: THE CAUSES AND TREATMENT OF OBESITY AND OVERWEIGHT ....... 173
5.1 INTRODUCTION .............................................................................................................. 173
5.2 OBESITY AS ENERGY IMBALANCE: THEORETICAL UNDERSTANDING OF BODY WEIGHT
MECHANISMS IN EARLY OBESITY SCIENCE .............................................................................. 173
5.3 PHYSIOLOGICAL EXPLANATIONS OF THE CAUSES OF OBESITY IN LATER SOURCES .............. 179
5.4 OTHER EXPLANATIONS OF THE CAUSES OF OBESITY......................................................... 186
5.5 CONVENTIONAL TREATMENTS FOR OBESITY AND OVERWEIGHT ......................................... 193
5.6 THE DIFFICULTIES OF SUCCESSFUL TREATMENT AND A NEW APPROACH OF ‘MODEST’ WEIGHT
LOSS ................................................................................................................................... 204
5.7 CONCLUSION ................................................................................................................. 207
CHAPTER 6: ANALYSING OBESITY SCIENCE................................................................ 209
6.1 INTRODUCTION .............................................................................................................. 209
6.2 THE ADVANTAGES OF BMI FOR OBESITY SCIENCE ............................................................ 210
6.3 BMI AND THE EPIDEMICITY OF OBESITY ........................................................................... 214
6.4 LIMITS OF MECHANISTIC MODELS .................................................................................... 217
6.5 CONFOUNDING AND SIMILARITY JUDGEMENTS .................................................................. 221
6.6 THE DOMINANCE OF INDIVIDUAL EXPLANATIONS: A FURTHER ‘WRITING OUT’ OF THE SOCIAL 222
6.7 CONCLUSION ................................................................................................................. 226
CHAPTER 7: THE INTERNATIONAL ARENA: OBESITY AND OVERWEIGHT IN WHO
REPORTS ............................................................................................................................. 227
7. 1 INTRODUCTION ............................................................................................................. 227
7.2 WHO REPORTS ON CHRONIC DISEASE, DIET AND BODY WEIGHT ....................................... 227
7.3 OBESITY AS A GLOBAL HEALTH PROBLEM......................................................................... 234
7.4 PARTICULAR POPULATIONS AS EXEMPLARS ..................................................................... 239
7.5 RELATIVE RISK AND THE EXPANDED LIST OF HEALTH CONSEQUENCES ............................... 242
7.6 DISSENTING ACCOUNTS OF THE HEALTH EFFECTS OF OVERWEIGHT AND OBESITY .............. 253
7.7 CONCLUSION ................................................................................................................. 256
CHAPTER 8: THE WHO REPORT OF 2000 AND THE ENVIRONMENTAL PARADIGM OF
OBESITY ............................................................................................................................... 258
8.1 INTRODUCTION .............................................................................................................. 258
8.2 A RENEWED FOCUS ON THE EFFECTS OF MODERNISATION AND INDUSTRIALISATION ........... 258
8.3 CONTINUING PROBLEMS WITH TREATMENT ...................................................................... 267
8.4 THE GROWING IMPORTANCE OF PREVENTION IN THE 1980S AND 1990S ............................ 272
8.5 PREVENTION IN WHO TECHNICAL REPORT 894 .............................................................. 276
8.6 ESTIMATING THE COSTS OF OBESITY AND THE GROWTH OF HEALTH ECONOMICS................ 283
8.7 THE GLOBAL BURDEN OF DISEASE PROJECT .................................................................. 290
8.8 THE ENVIRONMENTAL PARADIGM OF OBESITY IN TR894................................................... 292
8.9 CONCLUSION ................................................................................................................. 293
CHAPTER 9: INDIVIDUAL AND SOCIO-POLITICAL EXPLANATIONS OF EXCESS BODY
WEIGHT ................................................................................................................................ 295
9.1 INTRODUCTION .............................................................................................................. 295
9.2 PUBLIC HEALTH EXPERTISE AND BOUNDARY-WORK IN OBESITY SCIENCE ........................... 297
9.3 THE EFFECTS OF GLOBAL HEALTH INEQUALITIES: POVERTY AS THE MISSING FACTOR.......... 299
9.4 THE SOCIAL AND THE BIOMEDICAL: COMPETING MODELS OF HEALTH AND DISEASE WITHIN THE
WHO.................................................................................................................................. 302
9.5 MAKING AN AMERICAN AND EUROPEAN HEALTH PROBLEM GLOBAL? ................................. 305
9.6 CONCLUSION ................................................................................................................. 308
CHAPTER 10 CONCLUSION ............................................................................................... 311
10.1 INTRODUCTION ............................................................................................................ 311
10.2 POLICY ACTIVITY ON OBESITY 2000 TO THE PRESENT .................................................... 311

2
 10.3 THE CREATION OF POLICY KNOWLEDGE ......................................................................... 313
10.4 BOUNDARY-WORK AND POLICY SCIENCE IN PUBLIC HEALTH NUTRITION ............................ 316
10.5 CONCLUSION ............................................................................................................... 324
APPENDIX 1 ......................................................................................................................... 326
APPENDIX 2 ......................................................................................................................... 327
APPENDIX 3 ......................................................................................................................... 330
BIBLIOGRAPHY ................................................................................................................... 348

3
LIST OF ACRONYMS USED

ACS American Cancer Society

AHA American Heart Association
ASH Action on Smoking and Health
ASO Association for the Study of Obesity
BBPS Build and Blood Pressure Study
BDA British Diabetic Association
BMJ British Medical Journal
BMI body mass index
BMR basal metabolic rate
BNF British Nutrition Foundation
CBT cognitive behaviour therapy
CCF Center for Consumer Freedom
CDC Centers for Disease Control and Prevention (of the US govt )
CPG Coronary Prevention Group
CPS Cancer Prevention Study (I & II)
CHD coronary heart disease
COMA Committee on Medical and Nutritional Aspects of Food
CVD cardio vascular disease
DALY disability adjusted life years
DHSS Department of Health and Social Security (now Dept of Health)
DoH Department of Health
FAO Food and Agriculture Organization
FDA Food and Drug Administration (of the US govt )
GBD Global Burden of Disease
HDL high-density lipoprotein
ISCSH Independent Scientific Committee on Smoking and Health
IOTF International Obesity Taskforce
LDL low-density lipoprotein
LSHTM London School of Hygiene and Tropical Medicine
MAFF Ministry of Agricultural Food and Fisheries (now Dept of
Environment Food and Rural Affairs)
MRC Medical Research Council
MONICA Multinational Monitoring of Trends and Determinants in
Cardiovascular Disease
MRFIT Multiple Risk Factor Intervention Trial
NAAFA National Association for the Advancement of Fat Acceptance
NCD non-communicable diseases
NGO non-governmental organisation
NHANES National Health and Nutrition Examination Survey
NHES National Health Examination Survey
NHI National Heart Institute
NHS Nurses Health Study
NICE National Institute for Health and Clinical Excellence
NIDDM non insulin-dependent diabetes mellitus
NIH National Institutes of Health (of the US govt)
NPHI National Public Health Institute (of Finland)
OA Obesity Association (of Great Britain)
OHE Office of Health Economics

4
OPCS Office of Population Censuses and Surveys
PAF population attributable fraction
PAL physical activity level
QUALY quality adjusted life years
RCP Royal College of Physicians
RMR resting metabolic rate
SIGN Scottish Intercollegiate Guidelines Network
SOS Swedish Obese Subjects study
SSK sociology of scientific knowledge
SSRI selective serotonin re-uptake inhibitor
TMRU Tropical Medicine Research Unit
TOPS Take Pounds off Sensibly
TPRT Tobacco Products Research Trust
TR technical report
VLCD very low calorie diet
WHO World Health Organization
WHR waist to hip ratio

5
CHAPTER 1: INTRODUCTION

1.1 Introduction

In 2000, the World Health Organisation published a technical report that labelled
obesity as an epidemic and therefore, an important global health problem (WHO,
2000). This report described how average body weights were increasing globally due
to a combination of inappropriate diet and sedentary lifestyles, and it outlined the
negative consequences of such increases in terms of health and economic costs. It
marked a significant and important shift in biomedical understandings of body weight
and its relationship to health. This thesis sets out to analyse the development of this
shift and its implications for modern public health policy.

This introductory chapter situates my thesis within several different existing

literatures. Firstly, as a way of introducing the topic, I discuss historical accounts of
post-war public health and nutrition science, and recent sociological accounts of
obesity. Then I briefly outline recent sociological analyses of public health, which I
will go on to show, in the body of this thesis, are directly pertinent to analysing the
development of thinking on obesity as a public health problem. I go on to highlight
three theoretical approaches within Science and Technology Studies that I have used
in my research, and link them to writing on health policy. I describe my research
methods and some of the issues that have arisen from doing document based research.
Finally, I conclude with a brief outline of the following chapters.

1.2 Historical accounts of public health

This section contains an account of historical writing on post-war public health,

largely focusing on development in the UK, but also including some material
discussing the WHO. It then briefly outlines Thomas McKeown’s concept of the
epidemiological transition and presents a critique of this concept.

1.2.1 Historical accounts of post-war public health

6
This section discusses and reflects upon historical accounts of twentieth century
British public health, in order to provide a context for my analysis of the development
of obesity science. It relies on two principal sources: Marketing Health: Smoking and
the Discourse of Public Health in Britain (Berridge, 2007a) which uses the history of
government policy on smoking as a lens to examine post-war British public health;
and What Price Community Medicine?; The Philosophy, Practice and Politics of
Public Health since 1919 (Lewis, 1986), an analysis of the deleterious effects of the
foundation of the NHS and its re-organisation in 1974 on the discipline of public
health. These two sources provide the most complete accounts of the changes in post-
war British public health as a profession, an academic discipline and a wider topic of
concern.

The history of post-war British public health has focused on the NHS, the majority
provider of healthcare. However, while the recent history of public health in Britain is
necessarily entwined with that of the NHS, public health has its own independent
identity and status as a profession and academic discipline which an exclusive focus
on NHS public health history risks masking. Moreover, despite the foundation of the
NHS being described as ‘a major public health achievement’ (Baggott, 2000: 45),
writing on the twentieth century history of British public health (for example Lewis,
1986, Porter, 1999, Berridge, 2007a) has largely taken place outside the study of the
NHS. The topic is not covered in depth in any of the mainstream academic histories of
the NHS: Rudolf Klein’s (1995) history of the organisation does not explicitly discuss
public health and Charles Webster’s (2002) account is restricted to a description of
the negative effects of the 1974 NHS restructuring on the profession. For this reason,
Martin Gorsky (2008: 449) describes the history of public health as a ‘subaltern
narrative’ within the historiography of the NHS. Berridge’s (1999a) Health and
Society in Britain since 1939 is a notable exception to this pattern. In it she provides a
four-stage chronology of British public health in the twentieth century that describes a
shift from an early environmental focus in the 1950s; a focus on personal prevention
from the 1960s onwards; decline in the professional standing of the discipline after
the 1974 NHS re-organisation; and a subsequent rise in its status beginning in the late
1980s.

7
As I argue later on (see section 1.8.3), there is a very limited amount of available
secondary material covering the history of the WHO. Moreover the existing histories
(for example Siddiqi, 1995, Lee, 2009) do not explicitly frame it as a public health
organisation. This is despite the fact that the WHO developed out of international
public health collaborations, such as the International Sanitary Conferences and the
League of Nations Health Organization (Lee, 2009: 2), and that much of the work the
organisation carries out - immunisation campaigns, health promotion and outbreak
monitoring - is part of mainstream public health practice. It is, therefore, not clear
why it has not been considered in this light.

Accounts of the foundation and early structure of the NHS provide an important
context for any history of twentieth century British public health. I will highlight three
aspects of the history of the NHS that are directly relevant to the history of public
health – the role of the medical profession, the dominance of hospital-based curative
medicine, and recurring attempts by British governments to cut the costs of
healthcare. Mainstream academic accounts of the history of the NHS (for example
Berridge, 1999a , Klein, 1995, Webster, 2002) all include discussions of these three
topics.

Different groups within the medical profession managed to secure varying

concessions in the protracted and fraught negotiations that led up to the foundation of
the organisation in 1948 - GPs resisted becoming government employees and retained
their independent contractor status and hospital consultants retained the right to take
on private work – but their medical expertise gave them an authority within the
organisation other healthcare professions lacked (Klein, 1995: 17). In describing the
period until the 1980s, Klein argues that

The internal political history of the NHS in the 20th century was, for most of
the period, the history of the relations between the government of the day and
the medical profession….[It] appeared to offer a paradigm of “oligarchic
elitism”, reflecting the fact that it was the child of a union between
technocratic paternalism and professional interest. Political ideology intruded
from time to time…but for much of the time arguments about policy were a
duet between the government and the medical profession (Klein, 1995: 231).

8
At its foundation, the NHS was a healthcare service dominated by hospital medicine
(Berridge, 1999a: 29, Webster, 2002:29, 38-46). It was established by nationalising
the existing hospital system and bringing it under central government control. A
tripartite structure with different systems of administration for hospitals, primary care,
and public health and social welfare was created (Webster, 2002). The historical
unwillingness of the medical professional to accept local authority control (Klein,
1995: 70) was one important cause of this centralised system and also had a profound
effect on the discipline of public health (see below). This partly reflected the power of
the medical profession: ‘much of the profession itself was at this stage at best
indifferent, and at worst actively hostile, to preventive medicine, social medicine and
public health’ and combined with the administrative structure of the NHS led to the
creation of a service ‘which favoured central rather than local control; curative over
preventive and social medicine and public health; and hospitals over health centres,
and, indeed, primary care in general’ (Stewart, 2008: 463). This model continued to
operate throughout the 1960s and 1970s, when the Hospital Plan

consisted of, essentially, a technocratic approach which strongly favoured

curative, and consequently also hospital-based, medicine. In new, and very
large, hospitals heroic surgery and newly synthesised drugs were to be the
solution to individual and national health problems (Stewart, 2008: 462).

Since its establishment the NHS has operated under significant financial constraints. It
cost more than had been anticipated in its first two years (Webster, 2002: 31) and,
despite its relative cheapness, its perceived expense continued to be a politically
contentious issue throughout the 1950s and 1960s (Webster, 2002: 32-35). Even in the
1960s, when the economy was growing and health expenditure was rising, population
change and developments in medical technology meant that ‘increased expenditure
was barely keeping pace with changing circumstances’ (Stewart, 2008: 462). Such a
system of state-funded hospital-based health care, thought to be short of resources,
made the financing of public health and preventive medicine politically problematic:

And it is clear that while the early NHS did at least acknowledge the need for
a preventive programme, it was taken aback by the demands on its curative
services which the advent of “free” healthcare engendered. However, pursuing
a social/preventive/public health agenda would have required a much more
radical approach than that adopted in the post-war era…addressing health
issues from a preventive standpoint involves not only political will, but also a

9
 recognition that its benefits may take a considerable time to come to fruition.
To put it another way, it is politically more attractive to increase short-term
hospital capacity than to institute a preventive programme via, say, health
centres (Stewart, 2008: 469).

In a healthcare system dominated by hospital consultants and funded through public

taxation, it can be much more politically expedient to promise funding for the
building of hospitals, the provision of new treatments or even the reduction of waiting
lists, rather than funding for preventive programmes whose effects will not be
apparent for many decades.

Prevention of disease was crucial for social medicine, a branch of academic medicine
which developed in the 1930s, and became very influential in British universities in
the 1940s. Its concerns overlapped with those of public health in its aim to analyse the
causes of disease in populations:

The international social medicine movement before the Second World War
aimed to create a new social role for medicine in order to grapple with
epidemiological transition created by economic and social developments in the
twentieth century. The interdisciplinary program between medicine and social
science would provide medicine with the intellectual skills needed to analyze
the social causes of health and illness in the same way as the alliance between
medicine and the laboratory sciences had provided new insights into the
chemical and physical bases of disease (Porter, 2006: 1668).

Thus, social medicine was an important body of ideas arguing for change in public
health in the 1930s and 1940s: ‘In Britain, the ideas of social medicine were taken up
by academics in medicine and the social sciences and the ideas developed as part of
the more general discussion of health planning and reconstruction during the war.’
(Berridge, 2007a: 29). In 1942 John Ryle was given the chair in Social Medicine at
Oxford University, and in 1948 the MRC funded the Social Medicine Research Unit
which became an institutional home for the research of Richard Titmuss and Jerry
Morris. Between 1942 and 1944, Titmuss and Morris had published three papers on
the epidemiology of juvenile rheumatism, the history of rheumatic heart disease and
the epidemiology of peptic ulcer that Ryle described as ‘the first practical example of
social medicine’ (quoted Oakley, 1997: 90). In this research social medicine was,
‘part of a wide and ambitious project …to reshape both health policy and medical

10
practice: the end result was to be a profound health-promoting impact on the welfare
of the community at large’ (ibid.).

However, social medicine was a poorly defined concept and competing definitions
existed (Porter, 1997a: 104). Titmuss and Morris had broad interests in the
relationship between ill-health and social environments, whereas Ryle’s narrower
understanding of social medicine saw it ‘as a kind of statistical addendum to the
clinical model of health and disease’ (Oakley, 1997: 92). The location of social
medicine also led to a narrowing of its intellectual concerns:

Increasingly, the concept of social medicine was narrowed in order to stake a

claim to academic respectability. Ryle’s own work increasingly emphasized
not only the links with clinical medicine and epidemiology at the expense of
social science and health policy, but also the importance of the study of “social
pathology” – the quantity and cause of disease – at the expense of the more
radical and difficult aim of promoting health (Lewis, 1986: 39).

As Berridge explains: ‘Social medicine in its Oxford variant and also in Birmingham
under the leadership of Thomas McKeown increasingly came to mean medical
statistics’ (Berridge, 2007a: 31). This narrowing further increased the discipline’s
isolation:

The isolation of social medicine was enhanced by the fact that its
epistemological agenda narrowed. The study of ‘population health and
disease’ was vastly reduced in scope from the holistic union of theory and
practice in the establishment of scientific humanism as a new social ethic. As
social medicine developed in Britain after the Second World War the
influence of social science became limited to the empirical methods of
quantitative social research (Porter, 1997a:113).

Proponents of social medicine also took care to differentiate it from the theories and
practice of public health, and, consequently, public health doctors’ initial enthusiasm
for the approach was replaced by outright rejection (Lewis, 1986: 41-2). The
academic nature of social medicine and its resulting disconnection from public health
practice (Lewis, 1986: 8-9) meant that it could not be used as a theoretical bulwark for
public health:

11
 the development of social medicine was conditioned by its location within the
universities and, in the search for academic credibility, it moved further away
from a concern with health policy and social science….In part because of this
and in part because of their own narrowness of vision, public health
practitioners did not take up the idea of social medicine. This resulted in a
damaging rift between the teachers and practitioners of public health (Lewis,
1986: 37).

The longer term result of the weakness of social medicine was that few departments
of social medicine were set up, and it failed to have an impact either on the medical
curriculum or on medical practice (Berridge, 2007a: 30). This weakness may also
have resulted from its unavoidable entanglement in wider debates about the nature of
health and how it should be achieved:

The tension between biological and social explanations which gave birth to
social medicine was not new…..It could be argued that part of the failure of
social medicine to survive and transform itself into the study of the social
relations of health envisaged by Titmuss and others derived precisely from its
enclosure within the debate about two opposing models of how health is
determined (Oakley, 1997: 94).

Social medicine was an international movement (Porter, 2003: 58-62, Lee, 2009: 5-9).
Its aims and ideals developed out of exchanges between Soviet doctors and public
health experts, and European and American social reformers, public health
professionals and academic researchers (Porter, 2003: 60). One of the first chairs of
social medicine was created in 1945 at the University of Brussels for the Belgian
doctor, Rene Sand, who was also active in the League of Nations Health Organization
and the WHO (Lee, 2009: 8). The WHO was founded in 1948 when social medicine
was an active and influential body of thought, and its first director general, Brock
Chisholm had links to British social medicine (Brown et al., 2006: 64). The adoption
of a broad definition of health as ‘a state of complete physical, mental and social
wellbeing and not merely the absence of disease or infirmity’ within the WHO’s
constitution is evidence of the influence exerted by proponents of social medicine in
the organisation’s early years (Lee, 2009: 16).

In the UK, public health – both as an occupation and as a scientific discipline - was in
a state of flux in the 1950s (Berridge, 2007a: 26). Public health and general practice

12
were both ‘severely demoralized’ after the foundation of the NHS (Lewis, 1999: 337).
The profession of public health, in particular

had failed to capitalize on the coming of the NHS in 1948….The pre-war

public health “empire” in the local authorities had seen public health doctors
running hospitals and a wide range of services. But this empire began to
disintegrate post-war. Lewis has argued that MoHs bear some responsibility
for this outcomes, having previously been happy to extend their activities in
whatever direction offered, without a distinct version of what “public health”
was about (Berridge, 2007a: 26).

MoHs (or MOsH – usages vary) stands for Medical Officers of Health. The Public
Health Act of 1872 made the appointment of Medical Officers of Health by local
authorities a statutory necessity in England and Wales: ‘MOsH were charged with
enforcing the public health acts in the communities, for inspecting food, sanitation
and housing, and for publishing an annual report on their activities and the state of
public health in their communities’ (Lewis, 1986: 5).

In the 1950s and 1960s, MOsHs were in a difficult position as both GPs and social
workers encroached on their professional autonomy. A concern with personal hygiene
had led to campaigns in the 1920s and 1930s to reduce infant mortality: ‘Public health
justified its increased emphasis on clinic work with mothers and children as “applied
physiology”, a new kind of preventive clinical medicine. But this philosophy could
not survive the transition to the NHS and the subsequent claim of GPs to do the work’
(Lewis, 1986: 6). There had been hostility between GPs and MOsH since the 1930s:
‘Independent practitioners regarded MOsH as a threat because of their position as
salaried employees’(Lewis, 1986: 10). After the formation of the NHS, public health
doctors shared non-hospital medical work with GPs, and in the 1950s ‘GPs and public
health doctors continued to vie with one another in their claim to do personal
prevention work’ (Lewis, 1986: 58, see also Lewis, 1999). This was one aspect of a
wider problem for MOsH, as the foundation of the NHS had removed many of their
responsibilities, and left them with a poorly defined administrative role that largely
involved planning the delivery of a wide range of health and welfare services (see
below)

13
 The logic on the one hand of strengthening the GP and promoting him as the
focus of the preventive community health services and, on the other, of
promoting social workers and social service departments as the means of
developing community care, proved the downfall of the MOsH and the public
health departments (Lewis, 1986: 92).

Public health in the late 1940s and 1950s consisted of vaccination against infectious
diseases, such as tuberculosis and polio, and the clean air campaign which had
successfully lobbied for government legislation to reduce levels of air pollution
(Berridge, 1999a: 48-9). Vaccination was carried out by the public health profession,
but the clean air campaign was not, and so anticipated later single issue campaigns
conducted by campaigning groups rather than the public health profession (Berridge,
1999a: 49). This was part of a wider shift in public health practice (see below) that
can be related both to the success of the earlier sanitarian approaches and to wider
changes in causes of disease and ill-health:

The traditional focus of the public health practitioner had been the outbreak or
the epidemic: public health practitioners looked back to the great days of
environmentalism and the fight against cholera and typhoid in the mid
nineteenth century. But this pattern of disease and disease-related mortality
began to change in the middle of the twentieth century. Between the 1840s and
1971, three-quarters of the improvements in death rates had been due to a
decline in infectious disease with non-infectious disease accounting for the
remaining quarter. But this pattern changed after the Second World War. The
old “public health” diseases like TB or diphtheria were in decline…As the
population lived longer so non-infectious causes such as heart disease, strokes
and cancer grew in importance…These changes in the patterning of disease
were important for the new role of public health, for they, along with the
modes of explanation, prefigured the important changes in the nature and
focus of public health activity in the next half-century (Berridge, 2007a: 28-9).

The British tradition in statistics was an important influence on epidemiology in the

1950s (Berridge, 2007a: 32). The MRC Statistical Research Unit at LSHTM was
internationally influential in the 1940s and 1950s, and was the base for Richard Doll
and Austin Bradford Hill’s studies into the link between smoking and lung cancer.
This research extended and refined quantitative techniques such as large population
based surveys, case control and prospective studies, and marked a shift away from
earlier studies focusing on social inequality or environmental influences (Berridge,
2007a: 31). The increasing use of social surveys during this period was another
methodological development which also replaced explanations framed in terms of

14
social structural inequality with those that emphasised social behaviour (Berridge,
2007a: 32). Smoking was the first health behaviour to be studied in this fashion, and it
became a template for research within chronic disease epidemiology on other risk
factors, such as diet or physical activity:

Smoking was a pioneer issue, but others followed. The connection between
diet and heart disease began to be outlined in the 1950s. Exercise and fitness
also started to come on to the agenda. Jerry Morris’ paper of 1953, which
showed through epidemiological methodology, the differential susceptibility
to heart disease of sedentary bus drivers and active conductors, was symbolic
of the old and the new – an occupational study which showed the importance
of lifestyle factors (Berridge, 2007a: 50).

Despite its lack of institutional success, social medicine had remained influential and
was important in the development of chronic disease epidemiology. Researchers who
were part of its network were among the first to work on the health effects of risk
factors such as smoking and physical activity (Berridge, 2007a: 30 & 50). The new
focus on individual habits meshed well with the quantitative methods increasingly
favoured by researchers working in this area: ‘The behavioural model of lifestyle
prevention became indistinguishable from social medicine as an academic discipline
throughout this period as it was increasingly grounded in a methodologically
individualist epidemiology as the dominant science of the etiology of chronic
disease.’ (Porter, 2006: 1670).

The 1974 NHS re-organisation had a detrimental effect on the already embattled
public health profession. In the early 1960s, ‘public health doctors were deemed to be
doing primarily administrative, routine work and were paid accordingly’ (Lewis,
1986: 61) according to discussions of medical salaries, and, therefore, public health
doctors ‘experience considerable tension in reconciling their role in local government
with their professional aspirations’ (Lewis, 1986: 63). Several different health-related
occupational groups – sanitary inspectors, health visitors and social workers - were
also asserting their professional autonomy at this time (Lewis, 1986: 65-6). By the
late 1960s, public health was seen to be failing and ‘the material collected by the
Seebohm Committee seemed to provide more evidence of the essential weakness of
public health departments and their failure to analyse or to deal effectively with the

15
new health problems manifested in the changing patterns of mortality and morbidity’
(Lewis, 1986: 100-1). 1

Berridge also discusses the decline of public health as a profession:

Public health as a formal medical profession had been at a cross roads at the
end of the 1960s, with the old role of the MoH in the local authorities on the
point of disappearing. The new role in the reorganised NHS was yet to be
established. Many in the public health profession welcomed the new changes
which offered the apparently tangible advantages, to a medically based
profession, of incorporation within mainstream specialist services within the
hospital, and of consultant status (Berridge, 1999a: 85).

In the 1960s, local authority health and welfare departments had expanded and as
these departments were often run by MOsH this ‘further enhanced the authority of
public-health doctors’ (Webster, 2002: 124-5). However, local government re-
organisation and the creation of unified personal social services departments proposed
by the Seerbohm Committee in 1968 meant that MOsH would lose most of their
responsibilities and the majority of their staff. These changes ‘raised doubt about the
survival of public health as an independent medical speciality’ (Webster, 2002: 125).
Three separate bodies addressed this problem and evolved a ‘rescue package’ that
involved ‘proposals for resuscitating public health as the new speciality of
“community medicine”’(Webster, 2002: 125). As a result, the Faculty of Community
Medicine was established in 1972 and changes in training ‘emphasised the
reorientation of this medical subgroup towards specialist status and towards the
concept of medical administration’ (Berridge, 1999a: 86). The 1974 re-organisation
removed public health from local authority control and, therefore, dissolved the
relationship between public-health doctors and environmental health officers
(Webster, 2002: 104). Integration into the NHS was described as in the interests of
both public health professional and administrative efficiency:

Proponents of community medicine used the argument that the barrier between
prevention and cure was crumbling to promote the integration of public health
into the NHS, with the community physician providing the information for the

1
The Seebohm Committee was an inquiry into the future of personal social services. Its 1968 report
recommended the removal of social workers from local authority public health and the setting up of
separate social work departments (Berridge, 1999a: 44).

16
 efficient and effective administration of the service and co-operating closely
with clinicians (Lewis, 1986: 11).

However, because of this, ‘the speciality was born largely of administrative fiat, with
its identity and future bound closely to that of the reorganized health service’ (Lewis,
1986: 101). The concept of community medicine was not the product of medical
advances or disciplinary strength, rather it had been developed by academics in the
field of social medicine – such as Morris2 - ‘who believed that the practice of public
health had to be reformed if it were to survive’ (Lewis, 1986: 100). In addition, there
was an important and destabilising uncertainty about the role of community
physicians (Berridge, 2007a: 106, Lewis, 1986: 12): should they act as administrators
ensuring efficient provision of health services or as watchdogs who analysed the
health of local populations?

Public health doctors accurately foresaw the problems inherent in the new
NHS structure, in particular the possibility of tension between allegiance to
their local communities and the NHS bureaucracy, and between the demands
exerted by the hospital and by the community outside it (Lewis, 1986: 102).

None of the groups involved in planning the re-organisation – academics, civil

servants or MOsH - had thought in depth about what community physicians would
actually do, and so their workload was largely dictated by their place in the new NHS
structure, meaning that predictions about their conflicting roles proved accurate
(Lewis, 1986: 102). Thus the profound disillusionment of MOsH after re-organisation
had several causes:

the transition for many, especially the more senior public health doctors, who
comprised the vast majority of new community physicians was far from easy;
secondly, the NHS was almost immediately beset by a severe financial crisis,
and community physicians found themselves at the centre of conflicts over
cutbacks; thirdly, community physicians were neither understood nor
respected by their clinical colleagues…the community physician inherited the
low status of the public health doctor; and fourthly, community physicians
experience considerable difficulty in defining both their tasks and the aim of
the new speciality (Lewis, 1986: 125).

2
Jerry Morris was an influential critic of the work of public health departments and believed that
‘public health practice should be grounded more firmly in the principles of modern epidemiology’
(Lewis, 1986: 103)

17
The result of the re-organisation was that public health as an occupation was in a state
of disarray during the 1970s (Berridge, 2007a: 161). Another NHS re-organisation
took place less than 10 years later, in 1982, which removed the middle tier of NHS
administration and renamed community physicians with management responsibilities
district medical officers (Lewis, 1986: 126). This further upheaval led to polarisation
within the profession and a fifth of public health doctors subsequently took early
retirement (Berridge, 1999a: 86). Furthermore, in the 1980s, Lewis argues that ‘the
management role of the community physician has been eroded and with it the
importance of the community physician in the eyes of the government’ (Lewis, 1986:
127). According to her analysis the specialist advisor and management roles proved
impossible to combine (Lewis, 1986: 142).

Webster argues that in the longer term, the attempt to reform public health was largely
unsuccessful and partly responsible for the continuing decline of the profession:

The rescue effort was not entirely successful; community medicine failed to
achieve the status intended by its architects and recruitment fell off. The
limited success of community medicine was reflected in the failure to reshape
the health service according to the governments stated priorities – for instance,
by redirecting resources into community care, preventive, and promotive
medicine. Also the responsibility for public health and other functions
formerly important to the MOH became fragmented with unfortunate
consequences, as demonstrated in the long and disquieting run of public-health
alarms during the 1980s and 1990s (Webster, 2002: 126)

The increasing prominence, from the early 1960s, of smoking as a public health issue,
exemplifies a shift in the theory and practice of the discipline toward personal
responsibility (see also Berridge, 1999b, Berridge, 2003b, Berridge, 2006, Berridge,
2007b, Berridge and Loughlin, 2006). The publication of the 1962 Royal College of
Physicians report Smoking and Health,

encapsulated a number of significant developments in science and in policy.

For science, it endorsed the role of epidemiology and a change of scientific
“gaze” from direct biomedically influenced causation to statistical inference.
For policy, it helped initiate a post-war health policy based round the notion of
individual responsibility and personally avoidance of risk, a policy which
skilfully combined elements of morality with the concepts of science
(Berridge, 1999a: 50).

18
This approach developed through the 1960s and, from the 1970s, individual lifestyle
change was central to public health practice and health promotion initiatives. In the
UK it was put forward as government policy in the 1976 DHSS report Prevention and
Health: Everyone’s Business. Such approaches differed sharply from both the
traditional environmental of early public health, and the focus on structural causes of
ill-health contained in social medicine and health inequalities research (Berridge,
1999a: 93-6):

The new individualisation of health issues continued. “Single issues” were

highlighted rather than broader concern for social context. Individual action
could remedy health ills; the role of women as mothers was given special
attention. These ideas were a natural reaction to a period when scientific, high-
technology medicine had been centre stage. They drew on long-standing
beliefs about health which stressed the value of a health regimen – diet,
exercise and moderate living – and intermingled powerfully with lay beliefs
about moral responsibility for health and healthy living (Berridge, 1999a: 88)

A stress on personal responsibility was combined with mass communication

techniques to produce a new approach to public health, epitomised by health
education and later health promotion:

This new public health/health promotion constituency stressed population-

based interventions, taxation as a public health tool, with a focus on control of
advertising and on public information and mass advertising to achieve
objectives (Berridge, 1999a: 89)

In the 1970s and 1980s, the media techniques of the 1960s were developed and
expanded e.g. in the advertising for the Health Education Council produced by
Saatchi and Saatchi (Berridge, 1999a: 90). As I discuss below (see section 1.3) diet
and nutrition once again became politically controversial in the 1970s and 1980s with
publication of the NACNE guidelines in 1983 and the second COMA report in 1984,
and CHD prevention campaigns such as ‘Look After Your Heart’ (Berridge, 1999a:
90). However, smoking was the first target for such approaches and anti-smoking
health education campaigns epitomised the new approach to public health:

the 1970s was a crucial decade when a new public health style and agenda
emerged …this was the decade when public health developed its twin-track
emphasis on the evidence base and health services and on the lifestyle version

19
 of public health. Smoking was the crucial “tracer issue” for the latter, the
template for other public health issues subsequently (Berridge, 2007a: 161).

This focus on lifestyle and health behaviour meant that psychology and psychological
models also became important within public health (Berridge, 2007a: 109-10).
Another discipline that became incorporated into public health at this time was
economics. The oil crisis of 1973 meant that the costs of health care services became
an increasingly important issue which led to the growth and development of the
discipline of health economics (Berridge, 2007a: 109). It became ‘a central
investigative tool for health in the technocratic era of the 1970s’(Berridge, 2007a:
128). Berridge argues that ideas from health economics influenced arguments about
anti-smoking policies, which increasingly came to be framed in economic terms, and
also that research into the link between increases in taxes and cigarette prices and
reductions in smoking rates ‘helped turn public health in a different direction -
towards high taxation policies aimed at abstention from smoking’ (Berridge, 2007a:
130).

Significant institutional changes took place within British public health in this period.
The late 1960s and early 1970s have been labelled as a period of ‘technical public
health’ (Berridge, 2007a) which was one aspect of a wider technocratic approach to
decision making prevalent within the NHS at this time (Klein, 1995). Expert
committees had been a feature of the WHO’s work since its foundation, but in the
1970s they became an increasingly important part of British public health:

The way in which researchers related to government and the mechanisms

which brought the two together became more formal and framed by
government interests. The idea of rationality, that there could be a rational
relationship between research and policy, was high on the agenda. This was a
technocratic message…The rise of the expert committee was a feature of this
period. Also characteristic of this emphasis on the bringing of expertise into
policy making were the moves to develop “rational” policies on a cross-
departmental basis, which are seen as characteristic of the late 1970s
(Berridge, 2007a: 105).

The early 1970s re-organisation of the NHS had led to a re-assessment of the network
of NHS advisory committees, and the development of new expert committees
covering specific areas, such as the Advisory Committee on Alcoholism and the

20
Advisory Council on the Misuse of Drugs. This marked a shift in relations between
government, scientific research and the medical profession, and a new way for
individual clinicians and researchers to employ their technical expertise:

Doctors were moving from influence through “outside” organizations (like the
Royal College of Physicians (RCP) committee in the case of smoking), into
positions of scientific influence which operated at the boundaries between
government and the professions. The new expert committees were more
clearly organizations which linked expertise within and outside government
and which were founded on ideas of technical expertise and of the role of
research (Berridge, 2007a: 134).

Other key developments in 1970s public health included the growth in activity of new
style pressure groups such as Action on Smoking and Health (ASH) and the rise of
state responsibility for health education (via the Health Education Council) (Berridge,
2007a: 161). These aspects were apparent for smoking and other public health
concerns such as diet and heart disease and alcohol consumption. New activist groups,
who campaigned on a single public health issue like smoking or diet and heart
disease, began to be set up in the 1960s and their numbers increased in the 1970s. The
activities of ASH exemplified this wider trend in public health and demonstrated how
‘the “health pressure group” largely replaced the formal public health occupation as a
source of public pressure on health issues’(Berridge, 2007a: 164). Such groups
largely operated on the national level using the national media:

However, their role went further than simply conveying health education.
Their ‘new radical/new social movement’ style in fact masked close
relationships with government, both strategic and through government
funding. The groups, ASH in particular, were part of an essential policy-
balancing act for government. These were government funded pressure groups
as well as new social movements and part of the “policy community” which
composed the network of influence (ibid.).

Berridge argues that these institutional shifts combined to create groups of people
working on public health topics, such as smoking or diet, whose membership had few
connections to the existing profession of public health:

The raft of new expert committees which came into being in the 1970s and
after provided a means of cementing “communities of interest” between
government and public health-medical expertise. The new health pressure

21
 groups, moving away from a mass membership model to a media and
government-based (and funded) one, also symbolized the changing
relationship between voluntarism, expertise and the state. A cadre of public
health activists who had little relationship to the formal, medical public health
profession in health services was emergent (Berridge, 2007a: 282).

In the 1970s, the social medicine approach became dominant once again within the
WHO. Despite the early influence of social medicine, for much of the 1950s and
1960s, the work of the WHO had been dominated by technologically-driven disease
eradication campaigns, such as the Malaria Elimination Programme (Siddiqi, 1995).
Political developments in the 1970s, such as the emergence of the newly independent
African countries and the growth of the Non-Aligned Movement, contributed to a
major shift in the WHO’s priorities away from ‘vertical’ disease control approaches
(Brown et al., 2006: 66) . Instead the focus shifted to primary health care, as
epitomised by the goal of ‘Health For All in the Year 2000’ announced at the Alma-
Ata conference in 1978. The Alma Ata Declaration redefined the role of the
governments towards their populations by making it ‘abundantly clear’ that the state
was responsible for the health of its citizens, a ‘redefinition of the norms and
expectations of the state role with regard to health’ (Kickbusch, 2000: 981) .
Kickbusch argues that the WHO, therefore, can be said to have ‘invented’
international health policy at this time by defining both a problem and the
organisation’s role in providing a solution:

Just as the World Bank institutionalized and internationalized the concern for
global policy and made it an inextricable part of what development was…so
did WHO for health (ibid.).

The Health For All agenda was criticised as idealistic and impractical shortly after
Alma-Ata, and an alternative programme of selective primary health care focusing on
the key techniques of growth monitoring, oral re-hydration, breastfeeding and
immunisation was adopted the next year (Brown et al., 2006: 67, Lee, 2009: 75-81).
Moreover, although it was designed to provide basic healthcare for all populations,
irrespective of income, Health For All had most impact in the WHO Europe region as
it required a well-developed health infrastructure, something lacking in most
developing countries (Kickbusch, 2000: 891).

22
In the first half of 1980s the British public health profession was still seen as in a dire
state, but, beginning in the late 1980s, its standing began to improve. Writing in the
mid 1980s, Lewis was not optimistic about the future of community medicine:

the policy documents of the 1980s – Patients First, which signalled the second
reorganization of the health service and the Griffiths Report on the
management of the NHS – have shown little awareness or appreciation of the
community physician’s contribution. As a result the position of community
medicine is currently uncertain (Lewis, 1986: 149).

Other authors describe a crisis in community medicine that led to the setting up of a
committee of inquiry chaired by the Chief Medical Officer, Donald Acheson
(Baggott, 2000: 47). Baggott lists four key problems for public health at that time:
confusion about the multiple roles of health doctors; a lack of consistent collaboration
with the other professions working within public health; the reduced independence of
public health; and the low status of and low morale within the discipline (Baggott,
2000: 87-90). This analysis echoes many of the points made by Lewis about public
health in the previous two decades (see above).

However, Berridge describes signs of a revival in ‘formal’ public health developing

from the 1980s onwards (Berridge, 2007a). This was partly due to important
developments in both epidemiology and government concerns:

The policy and scientific community changed in the late 1970s and early
1980s. Concerns about the environment re-emerged as part of public health
rather than separate from it. Epidemic disease, previously consigned to the
“dustbin of history”, suddenly made a reappearance. New alliances emerged
within the science of public health; epidemiology was no longer proof enough
and gained greater legitimacy through support from biomedicine and the
science of psychopharmacology. Occupational health revived as a public
health matter (Berridge, 2007a: 208).

Some of these new concerns - especially those about the environment – were
developed by campaigning groups outside of public health (ibid.). Policy discourses
about passive smoking exemplified these trends as they focused on the effects of the
environment on individual health and considered environmental tobacco smoke as ‘an
infective agent for the population at large’ (Berridge, 2007a: 209).

23
This renewed focus on the environment and infection within public health was
accompanied by a new militancy both in the UK and at the international level (ibid.).
From the late 1960s, there had been a developing internationalism in public health
illustrated by the World Conferences on Smoking and Health and the subsequent
involvement of the WHO European Regional Office in the development of smoking
policy: ‘This type of manoeuvring at the international level and through international
organizations was to become commonplace over the following decades; but in the
early 1970s it was something relatively new and a way of building a consensus and an
alliance for action.’ (Berridge, 2007a: 163) . In the 1980s, continued collaboration
with the WHO led to the development of extensive and influential international
networks in smoking and other policy areas such as alcohol and drugs (Berridge,
2007a: 235). Further networks were established from 1987 onwards by the WHO
Europe Office’s Healthy Cities programme, an extension of its earlier work with
cities, local authorities and universities. This aimed to disseminate the Health For All
approach and build ‘a strong lobby for public health at the local level’ (Kickbusch,
2003: 385).

In the UK, the 1988 Acheson Report ‘made a determined effort to upgrade the status
of public health’ by proposing the appointment of regional Directors of Public Health
in England and the renewed publication of annual public health reports (Berridge,
1999a: 87). However, Berridge (1999a: 86) argues that this revival left many of the
issues about the role of the community physician (outlined above) unresolved.
Moreover, even though ideas about the nature of public health had broadened in the
1970s and 1980s as did interest in the field, the profession failed to respond to these
changes:

The public health community was much broader by the 1980s; and
environmental health officers, health promotion officers and GPs all
maintained that they, too, had a role in formulating health public policies.
However, the official public health response remained a medicalised one and
attempts to broaden the Faculty of Public Health Medicine were largely
failures. Arguably primary care and the pivotal role of the GP were to an
extent taking over the community role of the old public health
physician.(Berridge, 1999a: 92).

24
As a result of the weakness of public health, the rivalry between GPs and public
health doctors (see above) seemed to be over,: in early 1980s, the idea of GP-led
primary care was attractive as it demonstrated ‘the government’s commitment to
filling the vacuum left by the collapse of the public-health medicine speciality, many
of the functions of which…could be designated as part of the expanded scope of the
primary-care team’ (Webster, 2002: 178-9).

An important development of the 1980s was the renewed attention paid to health
inequalities research. The Black Report, Inequalities in Health, was published in
1980, and became the subject of controversy as the Conservative government initially
tried to restrict the number of copies available (Berridge, 1999a: 93). However, it was
subsequently re-printed and the controversy contributed to a body of research and
debate which ‘continued to grow and expand throughout the 1980s and early 1990s’
(Berridge, 2003a: 8). This became crucial to the re-emergence of the issue as an
important policy concern in the 1990s, which culminated in the 1998 Acheson
enquiry into inequalities. Although the controversy around the Black Report was a
particularly highly-charged episode in the relationship between government and
health inequalities researchers, it did not lead to the cessation of government funding
of such research. Sally Macintyre and others (Macintyre, 2003, Webster, 2003) argue
that it was part of a long tradition of such research that had been conducted
throughout the twentieth century, and continued throughout the 1980s and 1990s.
Much of this research was done outside the public health profession (e.g. Webster,
2003: 93) – by social science researchers and campaigners - but it was sometimes
used strategically by them to argue for the importance of the discipline. For example,
in the 1980s, public health doctors promoted claims about the negative health effects
of unemployment as a way of demonstrating their knowledge about the hidden costs
of government economic policies and, thus, improving their status, particularly in
respect to health economists (Bartley, 1994, Bartley, 1996).

In the 1990s, government policy began to re-focus on the health of the population
with the setting of specific targets in the 1992 Health of the Nation White Paper
(Lewis, 1999: 339), although these were criticised for omitting controversial aspects
that related directly to health inequalities research (Berridge, 1999a: 94). At this time,

25
public health policy also began to draw upon both pharmaceutical treatments and
genetic research as ways of bolstering its authority:

As the twentieth century drew to a close, smoking epitomized the conflicting

tendencies within public health and the changes in its knowledge base, the
general tensions between environmentally conscious health promotion ideas
and the growing influence of pharmaceutical imperatives which stressed
vaccination and or drug interventions as preventive measures…The rise of the
concept of addiction to nicotine as a policy fact signified the enhanced role of
pharmaceutical interests, the role of treatment and of medicalized ideas;
treatment became a public health strategy (Berridge, 2007a: 241).

The environment also remained an important concern, within smoking policy,

illustrated by the continued focus on the effects of environmental tobacco smoke and
increasing regulation of public spaces (Berridge, 2007a: 281). This expressed in a new
form the ongoing tension between individual and population levels that has been
central to twentieth century public health.

The 1980s and 1990s were also a period of crisis for the WHO, which was confronted
with an unpopular director general, a diminishing budget, and encroachment into its
area of activities from the World Bank and later public-private partnerships such as
the Global Alliance for Vaccines (Lee, 2009: 99-115, Brown et al., 2006: 67-8). In the
early 1990s, ‘WHO began to re-fashion itself as a co-ordinator, strategic planner and
leader of “global health” initiatives’ (Brown et al., 2006: 69). Kickbusch (2000: 983)
argues that the promotion of a new method of mapping and measuring the world’s
health problems - the ‘global burden of disease’ – was key to this process of shaping
global health policy, as like the Health For All agenda it provided ‘arguments and
evidence for the need to act jointly in view of increased common threats and the need
for common action.’

Many of the shifts that I have outlined above are still visible in British public health of
the 1990s:

By the end of the century, many of the public health tactics and concepts that
were so unusual in the 1950s and 1960s were so commonplace as to be
unremarkable. People expected doctors to give advice on lifestyle conditions
and politicians to give advice on how the population should eat and drink’.
Government itself had become a public health activist. The re-orientation of

26
 epidemiology post-war to concepts of risk in relation to chronic disease were
part of everyday currency, although they had been novel and lacking in
legitimacy in the 1950s (Berridge, 2007a: 281-2).

This historical literature provides an important context for my analysis of the

development of obesity science. Histories of the NHS describe the dominance of
curative, hospital medicine and the professional interests of consultants, at the
expense of preventive medicine and the interest of other groups, like public health
doctors who practiced medicine outside of the hospital system. The work of Lewis
and Berridge documents the weakness of the British public health profession in this
post-war period. This weakness meant that issues about the relationship between diet,
bodyweight and health could be addressed by other groups of professionals, including
the group of policy-orientated nutrition researchers and their allies, whose activities I
outline in later chapters.

The 1960s and 1970s are a particularly well documented period. Berridge’s account
of developments in this period points to the increasing importance of expert
committees, whose reports form the majority of my primary sources, in providing
technical advice to governments. She, and later Kickbusch, also describes the
development of international networks in areas such as smoking or diet and health,
created by researchers’ involvement in the work of bodies such as the WHO which
are also important to my research. Furthermore, Berridge and others describe a key
shift away from environmental approaches and towards personal prevention, that
began in the 1960s with the focus on smoking as a public health issue and developed
through the 1970s to include many other everyday activities, especially those, such as
diet and exercise, that were thought to be associated with an increased risk of
coronary heart disease. These accounts describe the ways in which the concept of risk
factor from chronic disease epidemiology was incorporated into public health practice
to produce an approach to disease prevention that relied on educating individuals to
change their behaviour.

Accounts of developments in the 1980s and 1990s are less comprehensive, but they
show a renewed interest in the topic of public health and a revival in the discipline,
and maybe also in the profession of public health. Berridge also describes the
incorporation of environmental concerns into public health discourses in the 1990s

27
which is reflected in obesity sciences’ use of environmental explanations in this
period. Finally, the writing of Webster and Macintyre shows that although there was
active research into health inequalities throughout this period, it was largely
conducted outside of the profession of public health. Due to its links with the issue of
poverty, research into health inequalities was also often a source of contention
between politicians and researchers in this field.

1.2.2 Ideas of the demographic or epidemiological transition

In chapter 2 I will describe, in broad outline, the changing patterns of health and
illness that are understood to have occurred in rich, industrialised countries since the
beginning of the twentieth century. In contemporary medical history and sociology
these changes are well recognised and often labelled the epidemiological,
demographic or simply health transition. Until the late 1990s, these terms do not
appear to have been used in British epidemiological and public health writing. Earlier
commentators on secular changes in population health discussed increasing rates of
chronic disease and linked it to ‘modern’ ways of life, such as increasing consumption
of fatty and sugary foods and increasing workplace mechanisation. These authors
spoke of ‘modern epidemics’ (Morris, 1964) caused by the ‘prosperous industrial
society’ (OHE, 1969), the ‘affluent society’ (Craddock, 1978) or the ‘affluent
lifestyle’ (WHO, 1986). As Rosenberg has argued, there has long been an ongoing
discourse of the negative effects of economic development on human health – what he
labels ‘pathologies of progress’ (Rosenberg, 1998). These labels, especially ideas of
diseases of affluence, draw on this long enduring strain of thought. However, these
authors also convey an implicit narrative of improvements in average health status
due to economic development and social progress or modernity that has been shaped
by ideas of a demographic transition. Thus, there are two overlapping and partially
contradictory narratives in play - of social progress as, alternatively, improving health
or leading to disease - that are often present in discussions of chronic disease and
obesity/overweight, but rarely acknowledged.

A conventional account that draws on the first of these narratives, stresses the role of
innovations in medical practice and treatment in the late nineteenth and early
twentieth century decrease in mortality, increase in life expectancy and increase in

28
population in the UK. From the mid 1950s Thomas McKeown, professor of social
medicine at Birmingham University, began producing a series of articles and books
challenging these sorts of accounts (McKeown, 1976, McKeown, 1979, McKeown
and Brown, 1955). His argument has two major elements. First, he used British
mortality statistics and dates of important treatment innovations to argue that death
rates from the major infectious diseases – tuberculosis, typhus, typhoid, diphtheria,
scarlet fever, pneumonia – were already in decline before effective medical treatments
existed:

I conclude that immunization and treatment contributed little to the reduction

of deaths from infectious disease before 1953, and over the period since cause
of death was first registered they were much less important than other
influences’ (McKeown, 1979: 77).

In the second part of his argument, McKeown uses statistics of improving food
supply, and an assumption that fertility had reached its maximum, to argue that
decreasing mortality and population growth in population were largely due better
resistance to disease deriving from improved nutrition:

the influences responsible for the decline of mortality were environmental,

behavioural [the declining birth rate] and therapeutic. They became effective
in the eighteenth, nineteenth and twentieth centuries, and their order in time
was also that of their effectiveness(McKeown, 1979: 78)

Arguing that economic growth led to higher standards of living, better nutrition, and
so reduced susceptibility to disease was a comprehensive rebuttal of one of the central
myths of professional medicine. One historian of medicine baldly summarises
McKeown’s argument as ‘the modern demographic revolution was agricultural, not
social or medical’ (Bynum, 2008: 645).

McKeown’s arguments, although highly publicised (one of his books was a best
seller), were iconoclastic and controversial: they did not completely replace the
mainstream medical account, but they epitomise a particular faith in economic
development as the key to improving population health, and thus were politically very
useful (see below). It was not until the 1990s that McKeown’s arguments begin to be
cited and discussed in epidemiological or public health textbooks (Beaglehole et al.,

29
1993, Connelly and Worth, 1997). McKeown’s is a specific argument from historical
demography, so it is not surprising that these texts do not discuss his ideas directly,
but, they are consonant with more general ideas the role of modernity or civilisation
that do inform much of post-war British chronic disease epidemiology.

The historian, Simon Szreter has argued against McKeown’s approach (Szreter,
2007). Szreter discusses McKeown’s arguments in the context of the development of
ideas of the demographic transition:

The notion that modern economic growth has invariably been associated with
a “demographic transition” of beneficent mortality and fertility declines has
been one of the most important ideas in the liberal social sciences throughout
the entire period since 1945, endorsed in their plans by many national
governments and by all the principal global government institutions, from the
U.N. to the World Bank. It has formed an important part of the justification for
the general policy presumption that national economic growth is an invariably
and automatically desirable goal and should be encouraged everywhere as
much as possible. (Szreter, 2007: 3)

Szreter argues that McKeown’s thesis supported this orthodoxy: McKeown’s

argument that improvements in mortality were due to improved nutrition levels
brought about by increased standards of living readily translated into a policy
recommendation that improvements in health derive directly from capitalist economic
growth (Szreter, 2007: 5). His critique of McKeown’s argument questions the simple
causality at the core of the narrative of the demographic transition and re-introduces a
role for collective action to combat disease, such as the work undertaken by the
sanitary reformers of the nineteenth century (Szreter, 2007). Specifically, Szreter
argues that McKeown’s interpretation of the nineteenth century epidemiological
evidence is mistaken and that evidence of an early decline in tuberculosis mortality –
a key element in McKeown’s argument – is not supported by the data. Moreover,
Szreter proposes that any such decline would have been counteracted by a strong rise
in bronchitis-type infections due to declining urban air quality.

Szreter’s alternative interpretation of the 19th century British mortality statistics

narrative is that they show a worsening of the health of urban populations due to
increasing congestion and lack of infrastructure, which was remedied by social
interventions such as municipal sanitation: ‘the decline in mortality, which began

30
noticeably in the 1870s, was due more to the eventual successes of the politically and
ideologically negotiated movement for public health than to any other positively
identifiable factor’ (Szreter, 2007: 125). Such public health work was pioneered by
the municipal authorities of the rapidly growing new urban industrial centres of
Birmingham, Manchester, Liverpool and Glasgow. It initially included provision of
clean water, paved streets, refuse collection and a mains sewage system, and, at the
beginning of the twentieth century, expanded to cover minimum housing quality
standards, the health visitor system and regulation of the milk supply (Szreter, 2007:
125-9). Szreter argues that unchecked economic growth is profoundly disruptive for
individuals and populations –‘setting in train a socially and politically dangerous,
destabilizing, and health-threatening set of forces’ – and leads to the “four Ds” of
disruption, deprivation, disease, and death (Szreter, 2007: 204). It is only collective
action, as undertaken by local and then central government in Britain in the late
nineteenth and early twentieth centuries, that is capable of transforming economic
growth into sustained social development that improves the living standards and
health of populations (Szreter, 2007: 235). This argument contradicts McKeown’s
conclusions, and with them the more general neo-liberal faith in economic
development as the only factor necessary for improvements in health.

Szreter thus mounts a comprehensive challenge to widespread assumptions that

‘modernity’, equated simply to economic growth, leads to improvements in health
status. And in doing so, he implicitly challenges also the idea that, accompanying this
general improvement in health, we see the emergence of new ‘diseases of modernity’
that can be understood as unfortunate side-effects of economic growth and improved
material standards of living. In the course of my thesis, I will examine the
construction of an account of one particular ‘disease of modernity’, namely obesity.
And I will show that this construction was shaped by precisely the kinds of
assumptions about the beneficial and harmful consequences of economic growth that
Szreter seeks to undermine. In particular, I will show that that account was
constructed in the context of a developing field of ‘policy science’ which was in turn
shaped by wider socio-economic factors including the institutional and political
marginalization of environmentalist approaches to public health.

1.3 Historical accounts of nutrition science

31
Historians of medicine have analysed the ways in which, since its beginnings in the
late nineteenth century, nutrition science has been used in the making of public
policy.

Nutrition science has been described as a complex and heterogeneous field:

“Nutrition” covers a very wide subject area, from sociology to molecular biology,
and similarly, “nutrition scientist” can embrace a remarkably wide range of
people. The membership of the Nutrition Society (founded in 1941) has included,
for example, chemists, biochemists, physiologists, medical and veterinary
practitioners, agricultural scientists, food scientists, dieticians, sociologists,
psychologists and administrators, as well as “nutritionists”. Nutrition is also an
area in which many wider interests intersect: the implications of nutritional
research potentially impinge, not only upon the practices of doctors, dieticians and
veterinary practitioners, but also on the policies of central and local government,
the agricultural, food and pharmaceutical industries, and the domestic habits of the
population (Smith, 1997: 1).

As it overlaps with individual daily experience, ‘“nutrition experts” abound, and

nutrition scientists face more acute problems that many other scientists in establishing
their expertise’ (ibid.). These multiple and contested forms of expertise, combined
with the effects of strong commercial interests, makes nutrition a particularly valuable
case through which to study the relationship between scientific expertise and public
health policy.

From the beginnings of the discipline in the late nineteenth century, nutrition
researchers have also attempted (often successfully) to use their expertise in the
development of public policy. Naomi Aronson describes how the construction of
nutrition as a social problem resulted from ‘the entrepreneurial claims of an emerging
scientific discipline’ and demonstrates how the first generation of U.S. nutrition
scientists ‘recruited support for their research by linking nutrition to the labor
problem, one of the most visible social issues of the time’ (Aronson, 1982: 483). In
the late nineteenth century, American industrialisation was accompanied by labour
unrest. Liberal reform movements promoted the scientific management techniques of
Frederick Taylor to solve the problems of industrial capitalism.

32
 economists and statisticians hoped to end class conflict by developing
objective criteria for the adequacy of wages. They established a base line,
termed the “standard of living”, for the quality of food, clothing and other
items needed to live adequately…At this point the questions of political
economy and nutrition research coincided: what were the nutritional
requirements for human subsistence? (Aronson, 1982: 477).

The Massachusetts Bureau of Labor Statistics sponsored the first American nutritional
study in 1885 and, as part of it, Wilbur O. Atwater conducted a study of the food
consumption of factory workers. From this study he concluded that existing wages
would be adequate if workers learnt to eat according to the scientific principles of
nutrition:

Optimal nutrition would increase productivity, thereby increasing the total

social wealth to be divided between the classes. At the same time, the
application of sound nutritional principles would reduce worker expenditures
for food, thereby effectively increasing the buying power of existing
wages…Atwater concluded that the problems of poverty and labor unrest
could be solved by teaching the masses to shop and cook economically
(Aronson, 1982: 478-9).

In America, from the late 1880s onwards, there was a ‘broadly based nutrition
research movement’ including prominent individuals, philanthropic organisations and
home economists who lobbied Congress for federal funding. However, as Aronson
points out, the scientific claims of nutrition researchers were not providing a neutral
solution to the problem of low wages but providing a rationale for prevailing wage
levels. Moreover, such surveys contained a moralistic and judgmental attitude towards
the eating habits of poorer families but not those of affluent families. Nutrition
researchers’ recommendations were often disadvantageous to poor families as they
saw the consumption of oranges and green vegetables as an unnecessary luxury, and
wives were advised to spend more of their time preparing and cooking family meals
(Aronson, 1982: 481-2).

This definition of nutrition as a social problems was gradually dismantled in the US

between 1900 and 1915 due to increased private research funding - which meant that
nutrition researchers did not have to make claims of social utility to secure
government funding - and by the discovery of vitamins which ‘undermined the
scientific basis for the nutrition researchers social program by showing that there were

33
essential foods that workers couldn’t afford’ (Aronson, 1982: 483 see also
Kamminga, 2000). Although the increasingly biomedical focus of nutrition research
meant that later researchers could avoid earlier social policy orientated activities
which were continued instead by home economists (Aronson, 1982: 483) such
entrepreneurial activities continued to be a recurring aspect of British and American
nutrition science in the first half of the twentieth century.

The health problems associated with poor diet of the urban working class had also
been a subject of British scientific and political discussion since the 1890s (Smith and
Nicolson, 1995). There were often two contrasting explanations put forward:

On the one hand, nutritional problems have been ascribed to economic

handicaps, to a substantial proportion of the working class simply not being
able to afford and adequate diet. Alternatively blame has been placed upon the
perceived prevalence of educational, behavioural and moral failings among the
urban poor, especially working-class mothers…those who sought to deny that
insufficient income was the fundamental cause of inadequate diet have
routinely advocated improvements in education as the most effective remedy
for the dietary ills of the less well-off. On the other hand, those on the opposite
side of the argument, who espoused various degrees of political and economic
restructuring, have often chosen to dismiss the question of nutritional
education as merely a distraction from more important issues (Smith and
Nicolson, 1995: 290).

As an example of the second type of argument, witnesses to the Interdepartmental

Committee on Physical Degeneration in 1904 ‘complained that working-class
housewives were idle, indifferent and unduly inclined to spend their meagre resources
in fish and chip shops or on tinned food’ and ‘stressed ignorance and degenerate
habits as causes of ill-health and inferior physique among the poor much more
strongly than poverty per se’(Smith and Nicolson, 1995: 292). Twenty years later, in
the mid 1920s, the debate was still being conducted in broadly similar terms when
Edward Cathcart, Professor of Chemical Physiology at Glasgow University argued
that inadequate diet was caused by ‘bad buying and bad cooking’ (Smith and
Nicolson, 1995: 297). Such arguments, closely related to those of Wilbur Atwater (see
above), were made regularly throughout the 1930s in British debates on diet and
poverty. John Boyd Orr, who published Food Health and Income in 1936, was one of
the few high-profile and influential writers to make the opposite argument that the
poor diet of the working classes was due to lack of money rather than ignorance:

34
 His unequivocal conclusion was that it was income, or the lack of income,
which was overwhelmingly the key factor in determining whether or not
working-class people obtained a healthy diet. The income of the poorer
sections of the population was simply not sufficient to enable them to purchase
adequate quantities of nourishing food (Smith and Nicolson, 1995: 300).

Debates about the nutritional status of the poor during the 1930s were complicated by
the recent discovery of the necessity of vitamins and other micronutrients for health –
often labelled the ‘newer knowledge of nutrition’. Medical Research Council (MRC)
research into ‘protective foods’ became a source of potential public controversy in a
period of economic recession:

The practical application of many of the principles of the “newer knowledge

of nutrition” in the 1930s was controversial because, at a time of mass
unemployment, low-wage labour and fiscal retrenchment, government
ministers and senior civil servants were desperately concerned to disprove
links between malnutrition, ill-health and low income. Financial
considerations determined economic and social policies, which in turn were
promoted as being compatible with the new science. Optimism about the
nation’s improving health characterized official statements in the press and
parliament, despite private unease at the Ministry of Health (Mayhew, 1988:
462).

Research showing the importance of expensive foods such as milk and vegetables
was, therefore, difficult politically, but also scientifically, since ‘complete scientific
and medical consensus as to the practical importance of vitamins and minerals in
human diets had not been achieved’ (Smith, 1995: 280). Despite this uncertainty,
vitamin researchers in the 1930s participated in debates about nutritional policy and
the links between poverty and ill health through deficient diet, and this dissemination
and popularisation had the effect of ‘hardening’ vitamins as they became
incorporated into diverse bodies of knowledge (Kamminga, 2000). The food industry
also popularised knowledge about vitamins as companies began to develop and
market a range of fortified foods, such as infant formula and margarine (Horrocks,
1995, Horrocks, 1997). The ‘newer knowledge of nutrition’ was thus very rapidly
mobilised by a wide range of groups other than nutrition researchers.

In the case of dietary standards and other forms of nutritional guidelines, the
committee room is an important site of scientific knowledge-generation (Smith, 1995:

35
280 see also Berridge, 2005). The claims produced such contexts can be addressed to
several audiences simultaneously and perform several different functions:

scientists who meet in committees to establish dietary standards may address

themselves towards both the “research community” and nonspecialist
audiences at the same time. The knowledge generated may be subsequently
used, not only in policy-making, but also in future research….knowledge-
construction and knowledge-deployment are not always easily separated.
(Smith, 1995: 281).

Smith and other authors have analysed the workings of several British nutrition
committees between the 1930s and the 1950s (Smith, 1995, Bufton et al., 2003, Smith
and Bufton, 2004), including the two 1934 Joint Conferences involving the Ministry
of Health’s Advisory Committee on Nutrition and the Nutrition Committee of the
British Medical Association (BMA) (Smith, 1995). In 1933 there had been
controversy arising from the publication of the BMA’s calorie and protein
recommendations which were more generous than those of the Ministry of Health.
The 1934 joint conferences were a ‘damage-limitation exercise’ resulting from this
conflict between nutrition experts, and the new recommendations for calorie and
protein requirements were minor modifications of generally accepted data that
‘protected the public image and credibility of the parties involved’ (Smith, 1995:
299). This example shows both the recurring possibility of public controversy when
expert committees set nutritional limits (see below) but also the fact that conflict
between groups of expert can often be resolved by minor reformulations of existing
knowledge rather than wholesale revision of technical standards.

By the 1930s international organisations had also started to develop expertise in

nutrition research. Throughout the 1920s, bodies such as the League of Nations
Health Organisation (LNHO) and the International Labour Office, had moved from
famine relief work to supporting nutrition research programmes (Weindling, 1995).
Although research was being conducted in several European countries and nutrition
had obvious relevance to the LNHO’s social understanding of health: ‘it was not easy
to funnel these into a tangible international scheme without infringing on perceived
national perogatives’ (Borowy, 2009: 381) But for Ludwik Rajcman, the director of
the LNHO:

36
 Nutrition was to become a prime example for establishing scientifically based
policies that would transcend the sectional interests of member governments
while disseminating scientifically innovative food policies to a wider world
(Weindling, 1995: 321).

From this perspective, nutrition research was able to establish the dietary
requirements compatible with ‘optimum health’ and this technical knowledge could
be used to advise governments. Rajcman considered convening an expert committee
in 1927, but it was not until after the appointment of a young nutrition researcher,
Wallace R Ackroyd, in 1931 that regular meetings began to take place. These
meetings discussed technical topics such the uses of statistical surveys and methods
for assessing malnutrition, (Borowy, 2009: 381-5). After 1933, ‘nutrition grew into
one of the largest international health study projects of the 1930s.... producing an
impressive number of publications’ (Borowy, 2009: 386). Economic crises made such
research an urgent political priority. In an approach that parallels modern appeals to
establish global food security, the use of the technical knowledge of nutrition science
was seen by those involved in the LNHO as a way of preventing another world war by
mitigating the effects of widespread hardship resulting from worldwide recession
(Weindling, 1995: 325). However despite these activities, which included the
publication of a best selling report in 1937, the LNHO’s nutrition activities were
ultimately unsuccessful in influencing national policies. This was partly due to the
unwillingness of governments to follow the organisation’s recommendations and
enact legislation aimed at minimising social inequalities (Borowy, 2009: 391-2).

In the same period, friction developed between the British Ministry of Health and
nutrition experts in the MRC (Mayhew, 1988, Smith, 1995). British nutrition
researchers, such as Edward Mellanby and John Boyd Orr, gained legitimacy in this
institutional conflict by using the new standards and perspectives developed by the
LNHO, whilst also being heavily involved in their creation:

The MRC, in its quest not only to gain prestige and influence by fostering high
quality nutrition research but also to place social policy and lifestyles on
scientifically rational foundations, began to exercise a decisive influence in
determining the direction of the LNHO’s nutrition programme (Weindling,
1995: 322).

37
This echoes the way in which later British researchers gained legitimacy from the
standards developed by WHO expert committees whilst being active members of
these committees (see chapters 4, 7 and 8)

The BMA produced another important report on nutrition in the late 1940s, as a
response to controversy about ‘the adequacy of the nation’s food supply generally and
more particularly to what extent people were consuming enough calories for growth
and work’ (Bufton, 2005: 127). As part of the controversy government departments
had been accused of not releasing the results of anthropometric surveys and food
expenditure data (Smith and Bufton, 2004). The BMA Nutrition Committee3 was,
therefore, convened to produce a report on the health effects of food rationing but by
the time the report was released the controversy had begun to abate, and the report
was considered a success because it did not lead to further negative press coverage.
Despite their official titles, ‘observers’ from the Ministries of Health and of Food
actively participated in the production of this report:

the government officials who served on the committee, ostensibly as

observers, played important roles in providing and writing information and in
writing the report. The interaction between officials and outsiders that took
place and the release of data that was involved formed part of the process of
transition from the aftermath of war to routine peacetime government and
contributed to a modus operandi for nutrition science (Smith and Bufton,
2004: 244).

These officials ensured that a ‘nominally independent’ report ‘largely reflected

official agendas’ and ‘contributed to a consensus, not only among but also beyond
scientists and doctors that the wartime and early postwar diet, although boring, had
been reasonably adequate’ (Smith and Bufton, 2004: 271). Smith and Bufton argue
that the processes of ‘political disengagement’ that took place within British nutrition
science in the late 1940s established a precedent for many subsequent report of the
topic of nutrition and health, and this success may be an important reason why,
despite its lack of official status, the BMA nutrition committee ‘guided policy for
nearly twenty years’ (Bufton, 2005: 126).

3
Members of this committee included Jack Drummond (Director of Research at Boots), VH Mottram
(former Professor of Physiology at King’s College of Household Science), Dr Harriette Chick (of the
Lister Institute ) and Dr John Yudkin (Mottram’s successor at King’s College)

38
In 1949, before the report could be finalised, there was a new controversy about
protein requirements in the Committee that mirrors the 1934 controversy outlined
above (Bufton et al., 2003). This controversy was difficult to manage because it
involved a large element of scientific uncertainty but was also obviously closely
connected to the controversial issue of meat rationing:

The problem of the level of protein requirement facing the BMA committee in
1949, it might be suggested, falls into the category of “trans-scientific”
questions…According to Weinberg, such questions are those that that can be
asked, but cannot be answered, by science. He pointed out that in debates
about trans-scientific questions, credibility becomes as important as scientific
competence… None of our scientific actors was really sure what to
recommend about protein requirements. None was specially competent to
settle the question, and the dispute and its settlement became concerned
largely with the credibility of alternative formulations (Bufton et al., 2003:
488).

Weinberg’s account is plausible, especially when combined with Jasanoff’s argument

that British policy participants are often chosen for their personal credibility, not just
their technical expertise (see section 1.6.2 below).The protein controversy was fuelled
by the political inclinations of individual committee members, but these were also
tempered by their professional ambitions and knowledge of the potential damaging
effects of a public controversy (Bufton et al., 2003: 491-2). As in 1934, consensus
was eventually reached in order to preserve the credibility the report and the
committee producing it.

In the interwar period, malnutrition amongst colonial populations had become

recognised as an important problem for the British government (Worboys, 1988).
Worboys argues that it was the use of the established technique of the nutrition survey
in a new context that led to this ‘discovery’:

The “discovery” of colonial malnutrition was a result of the direct transfer of

the “dietary survey” from the centre to the periphery. The important point was
that the transfer was direct and did not involve the creation of an
exceptionalist, tropical nutrition science. Rooted in the basic sciences of
physiology and biochemistry, nutritional science allowed a common approach
to problems worldwide (Worboys, 1988: 222).

39
Despite this reliance on biomedical research data to establish the existence of a
problem, malnutrition became officially defined as an economic, rather than a health,
concern: ‘when the British government officially took up the matter of nutrition in the
colonies it defined the issue structurally, as one of agricultural and economic
development, rather than of medical services and public health’ (Worboys, 1988: 216-
7). This framing resulted in part from the government’s priorities in the context of a
worldwide recession, but it may also reflect the ambiguous status of nutrition science
itself.

The claims made [about malnutrition] signalled the impending emergence of

nutrition as a discipline in its own right and showed that a number of people
were thinking about new professional identities for nutrition within public
health. However, the professional identity and allegiance of “nutritionists” was
by no means clear for it was maintained that the subject was “an economic,
agricultural, industrial and commercial problem, as well as a problem of
physiology”. At an intellectual level nutrition was clearly interdisciplinary in
character, embracing these and other fields. But at a practical and political
level, it was intricately bound up with economic concerns and political
interests that the idea of disinterested professional expertise was always
questionable (Worboys, 1988: 213-4).

This intellectual and practical intertwining of the scientific and the economic within
the discipline of nutrition allowed the emergence of two distinctive understandings of
the problem of malnourished colonial populations to emerge. This split between
economic and health framings of under- and mal-nutrition is also illustrated by the
earlier interventions made by nutritionists, such as Atwater, into debates on wages
levels.

The research outlined above also shows that setting recommended levels of protein
consumption had involved nutrition researchers in controversies since the end of the
nineteenth century. Sufficient protein, the most expensive and desirable
macronutrient, was seen as an important benchmark of an adequate diet. From the
1950s onwards, as British researchers started to study malnutrition in more detail,
protein deficiency became framed as an urgent global threat. This framing was largely
due to the work of the Protein Advisory Group, a network of nutrition researchers
who became influential in three UN agencies: the Food and Agriculture Organisation
(FAO), the World Health Organisation (WHO) and the United Nations Children’s

40
Fund (Unicef) (Ruxin, 2000: 151). The PAG was partially funded by US philanthropy
and depended on an important breakthrough in nutrition research:4

In 1957, research by J C Waterlow and Nevin Scrimshaw helped consolidate

the protein field by showing that kwashiorkor in Africa and Latin America
were indistinguishable…Fuelled by data indicating that kwashiorkor was
prevalent and preventable, WHO’s commitment to protein began to shape
something of a priesthood of nutritionists, exercising substantial influence
over policy (Ruxin, 2000: 153).

This ‘priesthood’ became very influential within the UN:

PAG and its associates made remarkable progress in encouraging interest in

protein among policy makers and politicians. In 1963, FAO concluded from
its Third World Food Survey that since in developing countries “the level of
animal protein intake is only one fifth of that of the more developed areas”
world food supplies would have to rise by 50 per cent by 1975 (Ruxin, 2000:
155).

By the late 1960s the problem of how to fill “the protein gap” had become a critical
issue for the PAG and associated UN bodies, and two years later a UN report written
by members of PAG called for ‘the last great push for protein’ (Ruxin, 2000: 156).
However, not all UN agencies accepted these claims. Despite the organisation’s
earlier support, FAO officials often resisted attempts by PAG to expand their
expertise. By the early 1970s, they did not fully accept the term “protein gap”
because FAO surveys showed that there was more protein in the world than was
required to meet basic human needs (Ruxin, 2000: 157-8).

In 1974, concern about a global food crisis made the protein gap seem like a less
important issue, and subsequent discussions focused on food production and hunger
rather than specific nutritional deficiencies. Critics within nutrition science also
argued against the emphasis put on protein malnutrition and kwashiorkor at the
expense of other forms of malnutrition:

McLaren published a stinging condemnation of protein policies in The Lancet

entitled “The Great Protein Fiasco” in 1974. He argued that there was no
4
The Josiah Macy Foundation sponsored a joint FAO/WHO conference on protein malnutrition in
1953 and the research of a sub-group of the PAG – the Committee on Protein Malnutrition – was
jointly funded by the Rockefeller Foundation and Unicef (Ruxin, 2000: 153).

41
 “protein gap” and that PAG had been perpetuating a myth…decades of protein
obsession (Ruxin, 2000: 159).

The disavowal of claims about widespread protein deficiency led to a wider mistrust
of nutrition researchers:

In the minds of many policy makers, the nutritionists had been unable to guide
or create successful projects. Knowledge of the complexities of good
nutritional status was increasing but there was a feeling that past results of
nutrition programmes had not fulfilled expectations. On protein policy makers
felt misled (ibid.).

Over the next two years, the PAG lost support and in 1977 it was replaced by the
Advisory Group on Nutrition (AGN). Ruxin (2000: 164) argues that this
organisational shift ‘symbolised a fundamental shift downwards of the status of the
outside nutrition expert’. This downward shift presumably affected the work of
nutrition experts in WHO committees on the relationship between diet and chronic
disease that began to operate from the late 1960s onwards (see section 7.2)

The work of other British expert committees in the post-war period, such as the
Committee on Medical Aspects of Food Policy (COMA) and the National Advisory
Committee on Nutritional Education (NACNE) has also been analysed by historians
of medicine (Bufton and Berridge, 2000, Bufton, 2005). Bufton and Berridge analyse
the post-war work of COMA (Bufton and Berridge, 2000). They outline how British
food and nutrition policy was seen to have improved the population’s health during
WW2 (see above) and how after rationing ended there were significant increases in
levels of sugar and fat consumption. By 1954 the government was aware of this trend
and linked it to increasing rates of heart disease:

Explanations for the rise, whether real or apparent, began to focus on diet. The
explanation which was to emerge as dominant was that of dietary fat, and
saturated fat in particular, was a central cause. This first emerged in the early
1950s through the work of Ancel Keys (Bufton and Berridge, 2000: 209).

In a 1952 paper in The Lancet, Keys linked age-related increases in heart disease in
US men to increasing blood cholesterol levels, comparing them to lower levels in

42
Italian men, and five years later he suggested that substituting vegetable for animal
fats might lower cholesterol levels:

Keys’ research was influential in an emergent international consensus. During

the 1960s, a number of western governments commissioned expert committees
to investigate the causes of coronary heart disease…Most advocated that
individuals should lower their sugar and saturated fat intake and increase their
polyunsaturated fat intake and it was a Scandinavian report which helped to
trigger developments in the UK. The diet and heart disease thesis gained
credibility (Bufton and Berridge, 2000: 210).

Other nutrition researchers, including John Yudkin in the UK, contested Key’s
claims, but ‘the diet and heart disease thesis gained ground in many industrialised
countries around the world with various expert committees giving support to the
idea.’ (Bufton, 2005: 131). In response to this increased interest in diet and nutrition,
in 1957 the UK government revived the Standing Committee on Nutritional and
Medical Problems, renaming it the Committee on Medical and Nutritional Aspects of
Food Policy (COMA). It was chaired by the Chief Medical Officer. In 1959 COMA
investigated the fat content in milk in relation to coronary heart disease (CHD) and
ten years later formed a panel on diet and heart disease to investigate the issue further.
The panel considered 400 articles over 4 years and produced its report in 1974:

Its basic recommendations were that the amounts of saturated fat and sugar in
the diet should be reduced. These recommendations were quite limited and
basic because the panel members had found it difficult to agree. Yudkin had
disagreed with the most basic of recommendations…He had argued that his
colleagues had exaggerated the role of dietary fat in causing heart disease and
underestimated the role of sucrose (Bufton, 2005: 131-3).

Many of the members did not accept the Report’s recommendation that individuals
should reduce the amount of fat they consume, and John Yudkin disagreed with this
so strongly that he wrote a note of reservation, arguing that the report minimised the
role of sucrose in causing heart disease (Bufton and Berridge, 2000: 212). He also
gave newspaper and television interviews to outline his disagreement with the
developing consensus:

Those scientists, like Yudkin in the case of diet and heart disease, whose
theories were not adopted as part of the agreed consensus, increasingly used

43
 the media to advance their scientific case; and the media was important, too,
when committee findings were ignored by government. (Berridge, 2005b: 25)

This recourse to the media meant that journalists became key intermediaries in the
science/policy process. Such a lack of consensus amongst the panel was one of the
reasons why the report had minimal effects:

The main impact of its report seems to have been on advertising. Here it was
used as a reference point. If a proposed advertisement made health claims,
unsupported by the conclusions of the COMA report then it was turned down
(Bufton and Berridge, 2000: 212).

In the second half of the 1970s the issue of diet and heart disease again began to
attract political and media attention. The RCP and British Cardiac Society produced a
joint report on the topic in 1976. This was not an official report but as it was sent to
all doctors, it ‘did in all probability influence the dietary advice some doctors gave to
their patients’ (Bufton, 2005: 133). It also contained advice to substitute
polyunsaturated fats for some saturated fat that contradicted the COMA report’s
recommendations. The Coronary Prevention Group was also formed in 1976 and
began to ‘campaign for greater public awareness of coronary heart disease and policy
initiatives from government to reduce its incidence’ (Bufton, 2005: 134).5 Such
initiatives led to the establishment of a second COMA panel on diet and health. By
the early 1980s there were two expert committees operating in this area: the second
COMA panel and the National Advisory Committee on Nutritional Education
(NACNE) which had been established in 1979. NACNE has been described as ‘quasi-
official’ (Bufton, 2005: 134) whereas COMA was an ‘insider’ body whose approach
has been contrasted with NACNE’s more ‘policy-activist’ orientation (Berridge,
2005b: 25). A few individuals, such as Philip James and JL Mann, were members of
both committees and the CPG.

The 1983 NACNE report recommended significant dietary change, notably that
individuals should reduce their average total fat intake and their consumption of
sugar, salt and alcohol. This received a hostile reaction from the DHSS and food
industry groups (Cannon, 1987) and through the British Nutritional Foundation (BNF)

5
Berridge (2005) argues that the Coronary Prevention Group was explicitly modelled on ASH (see
section 3.2).

44
the food industry tried to block its publication (Bufton, 2005: 134). Government
ministers were also accused in the medical and popular press of obstructing the
report’s production and restricting its circulation (Bufton and Berridge, 2000: 214).
A year later, COMA set up an expert panel to update the findings of its 1974 report,
and Philip James, who was chairman of the NACNE report committee, argued that
this was to provide a mechanism for the government to distance itself from the
NACNE findings, and that ‘the COMA panel felt under pressure to produce a report
acceptable to the government’ (ibid.). COMA took four years to produce this report
which included the review of 600 scientific articles and preparation of 40 working
papers and like the other committees, this panel also had difficulty in reaching
consensus. Bufton (2005: 135) describes its recommendations to individuals as ‘more
circumscribed than those of NACNE’ and they focused on improved food labelling in
the hope that consumers would purchase lower fat foods. There was controversy
around the advice to reduce consumption of saturated fats as one committee member,
JRA Mitchell, thought this was unsupported by the available evidence: as active
members of the CPG, Philip James6 and JL Mann disagreed. The viewpoints these
expert committee members might have reflected their disciplinary backgrounds so the
epidemiologists were more in favour of the diet-heart disease hypothesis than
cardiologists and physicians as Bufton (2005: 139) argues. However, this does not
explain the energy and activism of James (a nutrition researcher) and Mann (a
physiology researcher).

Despite the fact that the 1984 recommendations were similar to those of NACNE, the
COMA report was released without a significant controversy. According to Bufton
and Berridge, there were a number of reasons for this change: the chairman and
committee producing the COMA report wanted it to make an impact; the earlier
highly publicised controversy around the NACNE report had rehearsed many of the
arguments and alerted the press to the possibility of government hostility; and its
recommendations also fitted well with a new emphasis on prevention within primary
care. The report was also congruent with a new approach in public health:

6
‘James was considered by one COMA panel member to be fanatical about the diet-heart disease
thesis, and one former government minister thought him a maverick.’ (Bufton, 2005: 135)

45
 The diet and heart disease consensus was also part of a wider public health
consensus established in these years round the role of individual lifestyle and
the need for population-based policy interventions. Policy based on such an
approach rather than government attempts to regulate food consumption more
directly, is also likely to be less threatening for the food industry and may
even create new opportunities (Bufton and Berridge, 2000: 216).

In this period members of the COMA panel regularly appeared in television

programmes discussing the relationship between diet and heart disease giving the
debate a high public profile.

The scientists on COMA panels during the 1980s frequently appeared on

television arguing for their interpretation of the facts or contesting the science.
Some, such as W.P.T James and G. Rose were also involved in voluntary
pressure groups such as the CPG. James, who some saw as mercurial and
outspoken, was prominent in putting forward his views. This contrasts with
some other policy arenas where those who were outspoken, used the media
and held dissident views did not get access to expert medical committees and
were effectively excluded from the corridors of expert power (Bufton, 2005:
139).

Despite this level of activity (and activism) it is unclear what effect researchers’
campaigning efforts had in this area: ‘Scientists became activists, but it is difficult to
gauge if any of this influenced national policy formation.’ (Bufton, 2005: 140).

The next COMA report in 1994 made similar recommendations without dissenting
experts or a widely publicised disavowal by the Department of Health: ‘the report was
more detailed and closely argued than the previous reports. However, unlike earlier
panel reports, there seems to have been more consensus around the recommendations
and less dissent about dietary factors and their relationship to heart disease.’ (Bufton,
2005: 138). Another mark of increasing scientific consensus in this area was that by
1994 four voluntary groups in the area of diet and health - the British Cardiac Society,
British Heart Association, British Hypertension Society and British Diabetic
Association - who had previous been unable to agree on the content of dietary advice
for the public, had developed a common set of dietary recommendations (Bufton and
Berridge, 2000: 217). Bufton and Berridge argue that differing responses to these
reports were ‘not just a matter of “delay” in responding to scientific “truth”’(ibid.) but
rather the outcome of complex processes involving the workings of expert
committees, including the influence of individual members, reporting in the popular

46
press and long-term changes in public health policy. The process of producing these
reports had also become professionalized – it was more complex and time consuming,
more transparent and individuals were now paid for their work, and therefore had
more to lose by becoming embroiled in protracted controversies. However, the
consensus in this area was still a partial one:

Researchers are still publishing strongly argued papers and books questioning the
strength of the diet and heart disease thesis. Indeed one recent article reviewed
Yudkin’s theory in new form. This would seem to suggest that unanimity on the
issues is still not complete. (Bufton, 2005: 138)

The controversy over the nutritional value of high protein diets, and the Atkins diet in
particular, suggests that this consensus can be challenged by strong scientific or
commercial interests. Nutrition researchers were engaged in the making of public
health policy from the very beginnings of the discipline and this engagement resulted
in important controversies. Some of these, like the one over appropriate levels of
protein in the diet, recur regularly throughout the post-war period. During this period
a consensus developed about the relationship between consumption of high levels of
dietary fat and an increased risk of heart disease, but it took a great deal of work to
achieve (Garrety, 1997). Perhaps, this investment of time and effort was one of the
main reasons why, despite recurring criticism, the diet-heart disease consensus has
remained largely intact within academic medicine and public health practice from the
1970s to the present day.

This body of research demonstrates some of the recurring and difficult features of
incorporating nutrition science into British public health policymaking. Firstly, there
was very little consensus amongst experts in this field – controversies regularly broke
out in crucial areas such as protein requirements or the role of fat in causing heart
disease. Secondly, because of widespread public interest in such debates, individual
researchers with unorthodox points of view often used the media to gain more
credibility and enhance their ability to negotiate in private. Finally, as Worboys points
out (Worboys, 1988) there were frequently two distinct models of undernutrition
circulating within these debates: an economic model which saw it as an issue of
industrial development and a health model which saw it as an issue of medical
provision and public health.

47
1.4 Social science accounts of obesity

This section describes a collection of recent social science writing about obesity that
has informed my work. Just as medical and health policy discourses have increasingly
focused on obesity and overweight in the last ten years, a social science commentary
has developed that analyses the biomedical claims made about obesity and the
implications of such claims, both for public policy and wider society. Between the
1970s and the mid 1990s, the main non-medical accounts of social understandings of
excess body weight came from feminists and historians of eating or the body. The
subsequent development of social science writing on the topic has led to a growing
literature that analyses the media coverage of the ‘obesity epidemic’ and discusses the
extent and implications of medical accounts of large body size. It has also generated a
new sub-discipline called critical obesity studies or critical weight studies7 that rejects
many of the claims made by biomedical discourses about the extent of excess body
weight and its negative effect on health.

As I mentioned above, feminist authors have been writing about eating disorders,
body weight and body image since the 1970s (Orbach, 1982, Chernin, 1981, Wolf,
1990, Bordo, 1993), analysing obesity as the result of disordered eating deriving from
social pressure for women to maintain an acceptable body size and shape. Susie
Orbach, a practicing psychotherapist, sees obesity as an unconscious rebellion against
such pressures and a way of maintaining an alternative identity out-with the narrow
range prescribed within orthodox femininity (Orbach, 1982: 18-21). Naomi Wolf
argues that eating disorders are caused by dieting and that the ideal of thinness is a
political reaction to women’s increasing autonomy and political power (Wolf, 1990:
196-7). Susan Bordo sees eating disorders as a psychological consequence of the
contradictions of modern societies: ‘the coexistence of anorexia and obesity reveals
the instability of the contemporary personality, the difficulty of finding homeostasis
between the producer and the consumer sides of the self’(Bordo, 1993: 201). Jeanine
C Cogan (1999) echoes this earlier body of writing. She calls for a paradigm shift

7
Some authors use the term ‘critical weight studies’ to describe their approach because ‘obesity’ is a
medicalised term that is rejected by some activists, ‘fat’ is offensive to some, and their work is
concerned with many types of bodies (Monaghan et al., 2010).

48
away from ‘the weight-centred approach toward health’, contending that current
cultural meanings of obesity and eating disorders are based on a selective focus on
key pieces of scientific research that ignore the potential health benefits of moderate
obesity and the long-term failure rates of diets. Cogan argues that a rejection of the
current ‘thinness ideology’ is necessary to ‘prevent another generation of children,
women and men from developing eating problems, loathing their bodies, engaging in
risky weight loss strategies, and dying to be thin’ (Cogan, 1999: 202).

Related to and overlapping with feminist accounts, there is also a much smaller body
of writing that comes out of the fat oppression and size acceptance movements.
Explicitly modelling itself on other rights-based movements, it describes and argues
against the discrimination against people of large body weight, and challenges
medical orthodoxy on the health consequences of obesity (Schoenfielder and Wieser,
1983). Sociologists such as Sobal (1999b) have also examined the significant
stigmatisation experienced by those considered to be overweight or obese.

The most descriptive accounts in this area are historical or geographical surveys of the
phenomenon of obesity, overweight and large body size which describe changing
conceptions of body size in different countries and historical periods (Schwartz, 1986,
De Gaurine, 1993, Sobal, 1999a, Stearns, 2002).8 These authors describe social
contexts where large body size is valued as a sign of health and prosperity, and the
development of contemporary norms of slenderness that largely pre-date the
establishment of contemporary obesity science.

After 2000, the existence of an ‘obesity epidemic’ was widely reported in the British
and American popular press. The topic was used in many news stories and became the
subject of much commentary. Media studies researchers, particularly those in
America, have analysed this reporting. In one of the earliest analyses of the media
coverage of the obesity epidemic in the US, Lawrence (2004) argues that a vigorous
framing contest took place between 1985 and 1996 between arguments emphasising

8
Jeffrey Sobal is a sociologist who has been writing about obesity, weight and nutrition since the late
1980s, and has co-authored articles with medical experts such as Albert Stunkard, a psychiatry
professor at the University of Pennsylvania, and Kelly Brownell, a professor of epidemiology and
public health at Yale University (see section 5.4).

49
personal responsibility and those emphasising the ‘social environment’ which, for her,
results from corporate and public policy. As reflected in news discourse

popular understanding of the causes of obesity has moved from the

individualized and medical realms of biology and personal behaviour toward
the realm of environmental causation (Lawrence, 2004: 69).

I would argue that the shift that she identifies derives largely from the sort of policy
documents that I analyse in the following chapters. Some of them, in particular the
WHO technical report on obesity as global epidemic, were extensively cited in both
the popular and medical press (Fletcher, 2007). In contrast Samantha Kwan’s (2009b)
interview-based research suggests that ideas of individual responsibility and moral
models of fatness remain dominant, despite alternative accounts that stress the role of
a ‘toxic food environment’ (Brownell and Horgen, 2004). However, the difference
between these two sets of results is probably due to their differing methods.

Abigail Saguy and Kevin Riley (2005) also analyse claims made in the media
distinguishing between four different groups – antiobesity researchers, antiobesity
activists, fat acceptance researchers and fat acceptance activists – and describing how
each framed the causes of obesity differently. For Saguy and Riley, struggles over
framing and morality are actually conducted through arguments about scientific
method and facts:

in the case of obesity, debates over the nature of the condition have largely
hinged upon underlying moral assumptions about fat individuals and their
behaviours. To date, medical arguments about the health risks of obesity have
been effectively used to stymie political arguments about rights for fat
individuals (Saguy and Riley, 2005: 871, emphasis authors’ own).

Natalie Boero (2007) argues that there has been a ‘remarkable’ increase in scientific
research on obesity and its treatment since 1990 which has led to new forms of
medicalisation where ‘common knowledge’ about the causes of excess body weight
has been co-opted by the medical profession (Boero, 2007: 53). According to Boero,

The obesity epidemic is not a traditional epidemic of contagion and mass

death. Rather it is what I call a “post-modern epidemic”, one in which
unevenly medicalized phenomena lacking a clear pathological basis get cast in

50
 the language and moral panic of more “traditional” epidemics (Boero, 2007:
41).

The concept of moral panic (Cohen, 2002 [1972] ), is one that recurs in critical
analyses of obesity epidemic discourses. For example, Paul Campos and his co-
authors rebut claims that obesity is epidemic, that obesity is contributing significantly
to increased mortality rates, or that long-term weight loss is medically beneficial
(Campos et al., 2006). For them and other authors working in the field of critical
obesity studies (see below), the obesity epidemic is a moral panic rather than a
genuine public health problem. Such media analyses point to the moral and political
assumptions embedded in public health discourses about weight.

Several recent accounts use ideas of medicalisation (Conrad, 1992) to focus on the
implications of defining obesity as a disease, rather than an alternative body size or a
risk factor for other diseases. Annemarie Jutel (2006) argues that ‘overweight’ can be
distinguished from ‘fatness’ by the use of measurement. As a result of contemporary
medicalisation, obesity is now quantified and ‘Weight becomes the diagnosing tool,
but…it also becomes the basis for a pseudo-disease in and of itself’ (Jutel, 2006:
2270). According to Jutel, this change is due to a convergence of a belief in the
neutrality and objectivity of practices involved in counting and measuring bodies, and
the strength of normative ideas about modern conceptions of health. She describes
this change as a process of ‘creating a non-disease’ and concludes that

Overweight is not a disease anymore than slenderness is an indication of

health. Like baldness it is a description of physical appearance. It may or may
not be the cause of, related to, or a risk factor. It is not, on its own and outside
of the context of individual risk factors, beliefs and practices, meaningful
(Jutel, 2001: 2275).

Such accounts of medicalisation can be linked to the more historical approaches with
which I began. Jeffery Sobal argues that the medicalisation of obesity has been taking
place since the beginning of the twentieth century, and that

fatness has moved from a moral conception of fat as badness, to the

medicalization of obesity as sickness, to the demedicalization of large body
size as politically acceptable (Sobal, 1995: 67).

51
However, in analysing the development of public health policy a more fine grained
approach is necessary. For example, Chang and Christakis (2002) assess the claim
that obesity has been progressively medicalised using a content analysis of entries on
obesity in a bestselling American medical textbook. They argue that between 1927
and 2000 these entries show a change in ‘the social appraisal of obesity’

Obese individuals are progressively held less responsible for their condition in
successive editions of the text. Initially cast as societal parasites, they are later
transformed into societal victims (Chang and Christakis, 2002: 151).

They question whether this amounts to medicalisation since research has shown that
both patients and doctors are sceptical of the value of medical interventions.
Commenting on the work of one critic of theories of medicalisation, they argue that

Strong (1980) found that physicians generally dislike treating alcoholics,

preferring to manage problems that are more straight-forwardly ‘biological’,
or easily susceptible to abstraction from social context, problems for which
they have a clear-cut expertise in etiology, diagnosis and effective treatment.
Given that obesity shares many of these features with alcoholism on the
physician side, and given the emergence of resistance in the lay populace
against the medical management of obesity on the patient side…one might
follow Strong’s reasoning on the limits of medical imperialism and
hypothesise that the medicalisation of obesity is subject to important doctor
and patient constraints (Chang and Christakis, 2002: 167).

Given the difficulties of treatment, the parallel between obesity and alcoholism is
persuasive. Such arguments lead to the questioning of simple claims about the
medicalisation of obesity, and show the necessity of further empirical research to
establish in what ways obesity is being medicalised, and how these processes may be
resisted by both doctors and patients.

Economic historian Avner Offer (2001, 2006) provides an alternative account of

increasing average body weights in the US and UK that focuses on the role of
declining self-control. He argues, against theorists such as Norbert Elias and Pierre
Bourdieu, that, in the second half of the twentieth century, affluence, or economic
development, is leading to decreased self-control and, therefore, increased ill-health:
‘If affluence is generally associated with increasing well-being, the increase in
bodyweight above normative levels presents an unwelcome paradox.’ (Offer, 2001:

52
79). The concept of myopic choice is important to his argument that many people
irrationally privilege short-term rewards over the (health, economic and
psychological) advantages of delayed gratification:

Our hypothesis is that myopic choice accounts for the reversal of the historical
trend towards self-control….Affluence may be characterized as a flow of new
and inexpensive rewards. If these rewards arrive faster than the disciplines of
prudence can form then self-control will decline with affluence (Offer, 2001:
85).

New rewards were thrown up by affluence faster than it took to learn to cope
with the previous ones, so that overall, despite growing wealth, self-control
declined. Obesity shows how abundance, through cheapness, variety, novelty
and choice, could make a mockery of the rational consumer, how it enticed
only in order to humiliate (Offer, 2006: 169).

Offer identifies many important trends found in the public health literature: increasing
availability of food, the manufacture of a wider variety of more highly processed
foods (including snack foods and fast food) and more opportunities to eat away from
the home. He also identifies the ways in which a reliance on individual self-control
may actually reinforce health inequalities, by arguing that the resources an individual
has at their disposal affect their capacity to adapt their behaviour:

As body weight began to rise, it stimulated an effort to recapture self-control.

As in other dimensions of self-control, those with more at stake, and with
more resources, have been more successful. Women, with more at stake than
men, maintained lower weights; the well-off were more successful than the
poor. The repertoire of reactions included food choice, exercise, eating
disorders, normative defiance, and acceptance (Offer, 2001: 100)

However, his argument cannot adequately explain differential rates of weight gain
among different social groups – the concept of ‘mate competition’ which is central to
his argument cannot adequately explain why women from affluent social groups find
it more possible to exercise self-control than their male partners, or other groups of
women. It is also inadequate to argue that high rates of male incarceration and
employment in the military leads to reduced competition for men explains higher rates
of overweight/ obesity amongst African-American women (Offer, 2006: 166). This
limitation arises partly because his argument is explicitly based on a mainstream
economic model of individuals as rational consumers who attempt to maximise their

53
resources. There is little attempt in Offer’s writing to understand the complex
mixtures of individual and social factors that lead individuals to make ‘irrational’ or
‘myopic’ choices.

Since 2005 a group of social scientists have developed an alternative account of the
obesity epidemic that they call critical obesity studies. This body of writing brings
together many of the strands I have outlined above, especially a concern with
processes of medicalisation, and a focus on the potential for discrimination against
those who are labelled overweight or obese by the ‘rhetoric’ of an obesity epidemic.
One author, Lucy Aphramor, an NHS dietician as well as a social scientist, questions
the research linking weight loss with improvement in health, and argues that the
energy balance model omits too many variables to satisfactorily explain individual
weight gain. In particular, she questions whether it is ethical to recommend weight
loss treatments, given a 95% failure rate (Aphramor, 2005: 319). Another sociologist,
Lee Monaghan, argues that the concept of ‘excess’ weight ‘does not correspond with
epidemiologic evidence’ since the association of ‘overweight’ with excess mortality is
equivocal (Monaghan, 2005: 305). He contends that an important reason for this
inadequacy is methodological inadequacy on the part of existing research that does
not adequately account for the effects of physical activity on morbidity and mortality
(Monaghan, 2005: 307). For Monaghan

the highly publicized war against fat is about moral judgements and panic
(manufactured fear and loathing). It is about social inequality (class, gender,
generational and racial bias), political expediency and organizational and
economic interests. For many everyday people, including men and boys (but
more often women), it is about striving to be considered good or just plain
acceptable in a body-orientated culture (Monaghan, 2005: 309).

In a more recent discussion Monaghan and his co-authors have developed the
concepts of obesity epidemic entrepreneur (and entrepreneurship) in order to analyse
the actors, interests and practices involved in ‘constructing medicalised fatness as a
social issue or crisis’ (Monaghan et al., 2010: 38). According to this account,
‘creators’ such as epidemiologists and public health scientists draw on a wider
cultural fear of fatness to ‘actively define and redefine the “benchmarks” of “excess”
weight, which leads to obesity being constructed as a “chronic disease” that has
reached “crisis” proportions’ (Monaghan et al., 2010: 47). A combination of

54
professional interests and ‘highly lucrative connections between obesity experts,
organizations like the IOTF9 and the pharmaceutical industry’ (Monaghan et al.,
2010: 49) has led to the science behind the obesity epidemic being contaminated by
commercial interests and moral concerns. One example of the ‘fragility and morality’
of this science is the downward revision of American body mass index (BMI)
thresholds in 1998 by National Institutes of Health (NIH) which has been attributed
by Campos (2004) to the influence of vested interests in the weight loss industry.

Others also describe how biomedical narratives dominate discussions of obesity,

despite the large uncertainties in the basic scientific research. This dominance, they
argue, leads to a lack of discussion about the effects of such medicalised
understandings of body weight on individuals.

This omission constitutes an ‘exclusion fallacy’ where what we ‘choose’ not

to discuss…is assumed to have no bearing on the issue. In other words, the
stereotyping of fat, the feelings of guilt and shame that are produced through
this discourse, and the tendencies towards a culture of healthism10 and
individualism, are regarded as secondary to the primary concern to develop
scientific evidence to understand the causes of and treatment for the obesity
epidemic (Rich and Evans, 2005: 344).

The most significant piece of writing that has come out of critical obesity studies is a
book by two physical education academics containing a very detailed critique of the
concept of an obesity epidemic (Gard and Wright, 2005). These authors analyse
material from the scientific publications, the mass media and popular science accounts
of the obesity epidemic. In the process they assess the current state of ‘obesity
science’ including the use of the body mass index as an index of overweight, and the
epidemiological evidence linking large body size with negative health consequences
(see chapter 4). Gard and Wright largely reject the arguments of obesity science, on
the grounds of scientific adequacy and social equity. They argue that when scientists
attempt to explain the cause of the obesity epidemic in terms of changing diets or
childhood inactivity, they ‘draw on often conservative and age-old ideas about social
and moral decline’. They continue:

9
IOTF stand for the International Obesity Taskforce (see sections 3.5 and 7.2).
10
Robert Crawford (1980) defines healthism as a new form of medicalisation that has resulted from an
increasing stress on preventive behaviour and the increasing popularity of alternative health treatments
that see health as the result of individuals choosing to lead healthy lifestyles.

55
 Despite an almost complete dearth of compelling evidence, obesity scientists
regularly propose wide-ranging measures to ‘cure’ this decline, from what
children should learn at school, how parent’s should manage children’s lives,
to the ways in which buildings should be constructed. In short there appear to
be few areas of our lives which obesity scientists would leave untouched in the
“war on obesity” (Gard and Wright, 2005: 15).

They use analyses from feminism and the sociology of health and illness to provide
alternative approaches to the obesity epidemic, and conclude by arguing that
addressing the problem of the body weight of Western populations involves an overtly
moral and ideological project that would need to tackle issues such as ‘economic
disadvantage, the workings of capitalism, increasingly deregulated labour markets and
the imperative of companies, particularly, but not only, those that sell food, to be
profitable’ (Gard and Wright, 2005: 190).

An earlier article by Bryn Austin also critically analyses the ‘obsessive concerns with
food, fat and diet’ found in nutritional science and public health discourses (Austin,
1999). Austin argues that

Contemporary nutritional public health’s profound allegiance to the

biomedically conceived body and naivete about the centrality of cultural
meaning-making regarding eating, food, fat and gender do more harm than
simply hobble efforts to promote more healthful behaviour. By failing to
consider the intersection of food, bodies and diet in its cultural complexity,
public health gives scientific credibility to our society’s obsession with dieting
and loathing of fat and is implicated in the promotion of a cultural climate that
generates eating disorders (Austin, 1999: 245-6).

Austin’s analysis takes him into the processes of modern epidemiological fact
construction to argue that non-scientific beliefs about appropriate body weight are
distorting the epidemiological analyses, such as the link that has made between
increasing risk of cardiovascular disease (Austin, 1999: 255 - 8). In addition, critical
obesity studies, like the earlier feminist writing, begins to address the power relations
involved in public health discourses around excess body weight.

In my own work, I have not been able to consider some of these issues directly
because of the nature of the documents I analyse, and I briefly want acknowledge two

56
important areas of omission. The first is the voices of individuals labelled as
overweight and obese. They are not included in the reports or textbooks that I analyse.
Secondly, feminist analyses also point to contradictory gender aspects to
contemporary obesity science. Due to the predominance of male doctors and female
patients, obesity science is part of a long tradition of male medical advice aimed at
women.11 More importantly, twentieth century concern about increasing rates of heart
disease, out of which concern for excess body weight partially developed (see chapter
2), was framed as a problem of excess heart disease among men (Riska, 2000). It is
still not usually understood as a disease that affects women, despite, for example,
being the leading cause of death for women in the US and the UK (Pollock, 2010: 79-
81). Women were thus being expected to practice preventive strategies initially
designed to address the health problems of (middle aged) men.

1.5 Sociological accounts of the ‘new public health’

This section outlines a further body of literature relevant to the construction of ideas
of an obesity epidemic which provides an influential critique of the focus on
prevention within modern public health.

Medical sociology has considered the post-war growth in preventive medicine largely
through analysis of the activities of health education and health promotion. A related
approach that is more directly relevant to my research is a critique of the ‘new public
health’(Lupton, 1995, Petersen and Lupton, 1996). The ‘old public health’ is defined
as the 19th century sanitary reforms that attempted to control infectious disease,
whereas ‘new public health’ is a late 20th century individualised approach that uses
techniques such as health promotion, social marketing, diagnostic screening and
community participation in the attempt to control health problems seen as deriving
from ‘lifestyle’ factors such as diet, exercise and smoking (Lupton, 1995). Deborah
Lupton describes how ‘the new public health’ is

typically represented as a reaction against both the individualistic and victim-

blaming approach of health education and the curative model of biomedicine.
It is heralded as a return to the concern with environmental factors that first

11
Men are seen to be more prone to obesity but less prone to overweight.

57
 generated the public health movement of the nineteenth century’ (Lupton,
1995: 50).

Alan Petersen and Deborah Lupton draw on the work of Michel Foucault and Ulrich
Beck to argue that the contemporary focus on lifestyle and the management of risk,
embodied in health promotion, is ‘at its core a moral enterprise that involves
prescription about how we should live our lives and conduct our bodies, both
individually and collectively’ (Petersen and Lupton, 1996: 174). In particular they
discuss the development of the idea of the ‘entrepreneurial self’: ‘the self who is
expected to live life in a prudent and calculating way, and to be ever vigilant of risks’
(Petersen and Lupton, 1996: xiii), arguing that such a self is the product of neo-liberal
discourses about the duties of citizenship. They conclude that, ‘The new public health
can be seen as but the most recent of a series of regimes of power and knowledge that
are oriented to the regulation and surveillance of individual bodies and the social body
as a whole’ (Petersen and Lupton, 1996: 3).

Petersen and Lupton also discuss the application of epidemiological knowledge in the
development of public health policies: ‘Given its close alignment with policy
processes, it is surprising that epidemiology as a discipline has remained generally
impervious to the type of critical scrutiny to which other sciences have been treated
by sociologists of science’ (Petersen and Lupton, 1996: xiii – xiv). Their description
of the socially constructed nature of epidemiological data is heavily indebted to
theoretical approaches from STS: they are particularly interested in the role of
quantification in the construction of epidemiological facts (echoing Hacking, 1990,
Hacking, 2006b, Hacking, 2006c), and the relationship between epidemiology and
risk discourses. Two important issues for analyses of this kind of scientific research
emerge from their account. Firstly the sheer complexity of the physiological and
social processes being studied makes epidemiological research very complex: because
of this they argue that it is ‘beset by its reliance on probabilities and post-hoc
observational studies that attempt to relate health outcomes … to exposure to
hypothesised “risk factors” that precede the outcome’(Petersen and Lupton, 1996: 45).
Secondly, by definition epidemiological research applies to populations, yet much of
the new public health attempts to apply this knowledge to changing the behaviour of
individuals. People who are the target of such initiatives are well aware of this gap in

58
scientific knowledge and often have their own lay epidemiological explanations
which may or may not overlap with ‘scientific’ epidemiology (Petersen and Lupton,
1996: 50-52, see also Davison et al., 1992).

One weakness from which this analysis of the ‘new public health’ suffers is a deficit
of empirical evidence demonstrating the consequences of such changes in discourse.
Nor does it explain the historical development of these changes in public health (see
chapter 10). The analysis of ‘the new public health’ as a moral enterprise, however, is
particularly relevant to the topic of obesity since moralistic tendencies have never
entirely disappeared from advice about appropriate body weight (Austin, 1999, Gard
and Wright, 2005). Another important aspect of this approach is its focus on changing
definitions of health and disease. In preventive medicine, individuals are no longer
categorised into either the sick or the healthy, instead there has been a blurring of
these categories and the development of a new category of individuals who are
deemed ‘at risk’: ‘The notion of “health” therefore has become somewhat of an
abstract and liminal category in epidemiology, as all people, whether or not they are
experiencing symptoms, may harbour “risk factors” potentially leading to illness’
(Petersen and Lupton, 1996: 48). As we will see in the course of this thesis, this
aspect of the emergence of the ‘new public health’ is clearly evident in the case of
changing definitions of obesity.

1.6 Theoretical approaches from Science and Technology Studies (STS)

The media analyses and critical obesity studies approaches that I outline in section 1.4
above analyse the ways in which scientific claims are used in the construction of
social problems. In this sense they are social constructivist accounts. However a
critical weakness of this body of literature is that it does not analyse ‘science’ and ‘the
social’ symmetrically: scientific knowledge is either ‘black boxed’ so that its content
is not subjected to sociological analysis or it is implied better and more objective
science could exist, if all of the social contaminants were removed.

The key contribution of my thesis is to address these weaknesses in the existing

literature in this field by applying STS theory as an analytic framework through which
to analyse the content of policy-oriented biomedical accounts of excess body weight.

59
The crucial insights offered by STS are that that all knowledge is produced within,
and contains the traces of, particular social contexts and that it should be analysed on
this basis. In this section I will discuss four themes from STS that are important to my
research: the sociology of scientific knowledge (SSK) which I have used as an
umbrella approach that analyses scientific knowledge in terms of the operation of
social interests; Thomas Kuhn’s concept of scientific paradigms; Sheila Jasanoff’s
understanding of the distinctive character of policy science and Thomas Gieryn’s
concept of boundary-work. I will then outline two previous STS case studies of
nutrition science and conclude by defining how I use the term ‘obesity science’ in the
chapters that follow.

1.6.1 The Sociology of Scientific Knowledge, social interests and

paradigms

Within STS, the strong programme in the sociology of scientific knowledge (SSK)
represents a central theoretical approach to the analysis of the construction of
scientific knowledge (Bloor, 1976). The central contention of SSK is that, because
scientific accounts and judgments are inevitably underdetermined (i.e. it is always
possible to produce more than one account of an empirical phenomenon), the role for
sociological analysis is to explain why some knowledge claims – rather than others -
become established as scientific ‘fact’ (Barnes et al., 1996). This interest in ‘causality’
represents one of the four basic tenets of SSK.

The other three key tenets are that, in developing sociological explanations for the
development of scientific knowledge ‘impartiality’, ‘symmetry’ and ‘reflexivity are
vital (Bloor, 1976: 4-5). Impartiality and symmetry require that all knowledge is
explained on the same terms disregarding labels such as ‘true’ and ‘false ‘or ‘rational’
and ‘irrational’. The relativism which follows from SSK’s emphasis on impartiality
and symmetry is particularly important when studying processes of concept
formation. It permits the development of an analysis of the construction of pieces of
scientific knowledge using largely actor’s categories and prevents any essentialist
presuppositions about what constitutes obesity, as well as ensuring that the past is not
analysed as though the present state of knowledge were inevitable. Finally, the notion

60
of reflexivity means that the tenets of SSK apply to sociological knowledge as well as
scientific knowledge.

Canonical SSK studies utilise historical scientific controversies and offer explanations
for the ‘closure’ of these controversies (i.e. the establishment of particular knowledge
claims as objective fact) in terms of the social interests of the actors concerned.
However, the use of interest-based explanations by authors such as Shapin (1982) has
been criticised on the basis of the difficulties of identifying members of particular
social groups and the impossibility of demonstrating causal links between actors’
beliefs and actions (Sismondo, 2004: 46-7). However, when an analysis of social
interests takes place mostly at the level of a ‘professions’s’ rather than an
‘individual’s’ interests (as in the case of my research), these criticisms have much less
force. They do, however, act as a reminder to use interest-based explanations with
care, rather than automatically ‘reading off’ an individual’s motivations from their
occupation.

Unlike canonical SSK studies, my research analyses the contemporary history of

policy science, and the development of a scientific consensus rather than a
controversy.12 Nonetheless, for the reasons that I have outlined, it is a useful
theoretical approach in the context of my research. Steve Shapin argues that

An empirical sociology of knowledge has to do more than demonstrate the

underdetermination of scientific accounts and judgements; it has to go on to
show why particular accounts were produced and why particular evaluations
were rendered; and it has to do this by displaying the historically contingent
connections between knowledge and the concerns of various social groups in
their intellectual and social settings (Shapin, 1982: 164).

My research has begun to do this for contemporary public health policy in the area of
diet and nutrition. In the following chapters I will show why particular accounts of
obesity were produced and what models of health they contained. I will also describe
the contingent connections between scientific knowledge about obesity and the
concerns of the different social groups and settings involved in its production.

12
As I have described (see section 1.4), there is controversy about the nature of obesity and existence
of an epidemic of obesity. However, this has formed a small part of my thesis.

61
Although interest-based explanations allow for the influence of wider social processes
on the formation of scientific knowledge, the actors who are considered in most detail
in such historical case studies are usually individual scientists and their peer groups.
This narrow focus is problematic in the context of the present study, because, as I will
go on to illustrate, scientific knowledge about obesity is constituted through several
different fields of science, including biomedical laboratory research, clinical
knowledge and large-scale epidemiological studies. My research thus requires me to
draw upon additional theoretical approaches which broaden SSK’s conceptual focus
to consider how and why the relationships between these different fields of research
are implicated in the construction of obesity science.

One useful method of conceptualising these links is to employ Thomas Kuhn’s

concept of a paradigm (Kuhn, 1970). Kuhn describes scientific research as consisting
of periods of ‘normal science’ where activities are guided by a shared model of the
phenomena being studied and the methods of study, interspersed with scientific
revolutions where paradigms break down and have to be replaced. I have described
the two models of obesity that I identify and analyse in the following chapters as
paradigms because I think the concept captures the combination of theoretical and
practical knowledge contained within these models of excess body weight. The
individual and environmental models of obesity are simultaneously shared bodies of
knowledge about the condition and its causes, which link to laboratory research
science, and collections of standard treatments or policy solutions.

Kuhn is acknowledged to have used the term paradigm in many different senses.
According to Margaret Masterman (1970: 65) these different usages can be labelled
as philosophical (a set of beliefs), sociological (a universally recognised
achievement), or as an artefact or a construct, (an actual textbook or piece of work). I
use the term in the following chapters, largely, in her second sense of a recognised
scientific achievement. Each of the models of obesity that I outline is a widely
accepted body of knowledge about the condition. However, these two models are not
incommensurable in the way that paradigms are sometimes understood to be
(Sismondo, 2004: 17-18), and, in fact, they currently co-exist. This can occur because
the causes of obesity they discuss, and the solutions they advocate, operate at different
levels of social organisation, namely the individual and the population. Finally, these

62
paradigms develop incrementally (Barnes, 1982) – there have been no revolutions in
twentieth century public health policy in the area of diet and nutrition. Instead each
piece of writing, a report or a textbook, can be seen as developing the models of
obesity.

1.6.2 Regulatory science and boundary-work

Another way in which the field of science that I am studying differs from canonical
SSK studies is that it concerns ‘policy science’. Policy science is a distinctive form of
scientific knowledge in that it is explicitly oriented towards problems requiring some
kinds of government solution and it is largely a product of appointed experts and
expert committees. Sheila Jasanoff has written extensively on the relationship
between science and policy processes (Jasanoff, 1990, Jasanoff, 2005). She explicitly
rejects technocratic models of the science policy relationship that separate technical
considerations from social or political ones:

The notion that the scientific component of decision making can be separated
from the political and entrusted to independent experts has effectively been
dismantled by recent contributions to the political and social studies of science
(Jasanoff, 1990: 16).

Advisory (or policy-orientated or regulatory science) is, therefore, ‘a hybrid activity

that combines elements of scientific evidence and reasoning with large doses of social
and political judgement’ (Jasanoff, 1990: 229). For Jasanoff, such scientific advice is
a process of negotiation and boundary-work (Gieryn, 1999). In outlining
constructivist approaches to the study of science, Thomas Gieryn describes how

the boundaries of science are episodically established sustained, enlarged,

policed, breached, and sometimes erased in the defense, pursuit, or denial of
claims of epistemic authority. As knowledge makers seek to present their
claims or practices as legitimate (credible, trustworthy, reliable) by locating
them within “science,” they discursively construct for it an ever changing
arrangement of boundaries and territories and landmarks, always contingent
on immediate circumstances (Gieryn, 1999: xi).

Gieryn labels the activities involved in building, maintaining, adapting and policing
these discursive perimeters ‘boundary-work’: ‘Boundary-work occurs as people

63
contend for, legitimate, or challenge the cognitive authority of science’ (Gieryn, 1995:
405). His historical case studies illustrate the argument that such tasks are
fundamental to the credibility of scientific knowledge. 13

Such boundary-work is central to Jasanoff’s understanding of the interaction between

scientific expertise and policymaking. She argues that there is a fluid boundary
between science and policy, and that scientists draw and re-draw this boundary in
order to exempt ‘technical’ matters from political control:

By drawing seemingly sharp boundaries between science and policy, scientists

in effect post “keep out” signs to prevent non-scientists from challenging or
reinterpreting claims labelled as “science”. The creation of such boundaries
seems crucial to the political acceptability of advice. When the boundary
holds, both regulators and the public accepts the experts’ designation as
controlling, and the recommendations of advisory committees, whatever their
content, are invested with unshakeable authority (Jasanoff, 1990: 236).

However, this fluid boundary can also be used by expert scientists to define the limits
of their authority and avoid their recommendations becoming politically controversial.
Scientific advice about diet and body size has often been highly charged due to the
operation of extensive commercial interests in this area and the potentially wide
ranging political implications of public health policy. Researchers serving on expert
committees in this field have adopted differing definitions of the boundaries of
science (and therefore their expertise) in different contexts. For example, as they have
come to understand the causes of excess bodyweight as resulting from a damaging
environment, rather solely from individual failings, nutrition experts have acquired
the authority to condemn the un-healthiness of many aspects of modern industrialised
life (see section 8.2).

As I will go on to illustrate through my empirical analysis, the obesity epidemic

provides an important example of policy institutions as knowledge makers –
mandated or regulatory knowledge called into being for statutory purposes (Jasanoff,
1990). It is an example of knowledge making in the committee room, rather than the

13
One case study is of the botanical research of Albert and Gabrielle Howard who developed the
Indore system of composting – the horticultural compost heap - and are considered founders of modern
organic farming (Gieryn, 1999).

64
lab or the clinic (see section 1.3 above). The knowledge created in these processes of
policymaking has a distinctive character:

What emerges from a successful recourse to scientific advice, then, is a very

special kind of construct: one that many, perhaps most, observers accept as
science, although it both shapes and is shaped by policy. That such constructs
sometimes breakdown under political pressure is hardly surprising. Their
frequent durability is the greater puzzle, for they are founded neither on
testable, objective truths about nature, as presupposed by the technocratic
model of legitimation, nor on the kind of broadly participatory politics
envisaged by liberal democratic theory (Jasanoff 1990: 234).

This is partly because of the many overlapping roles that individual members and
committees can play whilst providing expert advice to governments:

Protected by the umbrella of expertise, advisory committee members in fact

are free to serve in widely divergent professional capacities: as technical
consultants, as educators, as peer reviewers, as policy advocates, as mediators,
and even as judges. Though their purpose is to address only technical issues,
committee meetings therefore serve as forums where scientific as well as
political conflicts can be simultaneously negotiated. When the process works,
few incentives remain for political adversaries to deconstruct the results or to
attack them as bad science (Jasanoff, 1990: 237).

The activities of post-war British advisory committees on nutrition that I outlined in

section 1.3 provide an illustration of many of these different roles and the processes of
negotiating scientific and political conflicts that Jasanoff outlines in this paragraph.
According to Jasanoff participants are aware that they are making claims to expertise
that go beyond the purely scientific: ‘the experts themselves seem at times painfully
aware that what they are doing is not “science” in any ordinary sense, but a hybrid
activity that combines elements of scientific evidence and reasoning with large doses
of social and political judgement’ (Jasanoff, 1990: 229). Reading reports on obesity
and overweight does not give a sense of ‘painful’ awareness, but does demonstrate the
hybrid nature of such knowledge as authors attempt to combine their research results
with an awareness of both the politics of food production and the effects of
industrialisation and modernisation on daily life.

Jasanoff also argues that there are distinctive aspects to British policy making and
expert knowledge creation. As it often takes place in private, British policy making is

65
a ‘more insulated process’ that places greater faith in the trustworthiness of
recognised experts (Jasanoff, 2005: 262) than the US political system. Secondly, in
the UK policy making involves a specific kind of expert who has earned their status
through sustained work on public issues:

To a remarkable extent British expertise remains tied to the person of the

individual expert , who achieves standing not only through knowledge and
competence, but also through a demonstrated record of service to society. It is
as if the expert’s function is to discern the public’s needs and to define the
public good as much as it is to provide appropriate technical knowledge and
skills for resolving the matter at hand. In this cultural setting individuals
ranging from Prince Charles and Baroness Warnock to Julie Hill of the Green
Alliance and various academic social scientists can all emerge as authoritative
policy actors (Jasanoff, 2005: 268).

The relatively private nature of British policy making combined with a reliance on
personal credibility means that, in some situations, expert scientists can credibly make
recommendations going well beyond the areas of their research expertise (see chapter
8).

1.6.3 Previous STS case studies of nutrition science

As noted in the introduction to this section, the key novel contribution of my thesis is
to apply STS theory to an analysis of obesity science. However, in concluding this
discussion of my utilisation of STS it is important to draw attention to two important
pieces of STS writing which overlap with the empirical focus of my research. Few
STS authors have addressed nutrition science or the relationship between diet and
health. Two authors that do write about this topic are Stephen Hilgartner and Karin
Garrety. Garrety (1997) discusses the development of a consensus in American health
policy of the 1970s, about the links between cholesterol and heart disease. She
describes the complex negotiations that resulted in accepted knowledge about the
links between coronary heart disease, cholesterol and dietary fat, in the absence, she
argues, of ‘definitive’ experimental evidence:

Even before the new knowledge escaped from the laboratory, actors in several
social worlds promoted conditions which helped the initially unstable hypothesis
to survive and thrive. They created both the social problem of heart disease and its
solution – a massive medical research effort (Garrety, 1997: 757).

66
The organisers of the 1984 Consensus Conference on Lowering Blood Cholesterol to
Prevent Heart Disease recommended that the entire US population over the age of two
should reduce its consumption of dietary fat and cholesterol, despite results from a ten
year trial that were only marginally statistically significant, and criticised within the
medical literature (Garrety, 1997: 732). In describing how an influential and relatively
durable consensus was achieved, Garrety demonstrates how different actors such as
the American Heart Association, physiological and epidemiological research
scientists, journalists from the non-scientific press, the food industry, the American
Medical Association, politicians and the FDA became involved in the processes of
creating knowledge about diet and health:

Between 1960 and 1984, while scientists were endeavouring to construct the
definitive evidence, many social worlds continued to use and shape various
interpretations of the cholesterol hypothesis…the cholesterol hypothesis
became increasingly popular in many social worlds. Most medical
organizations, Western governments and numerous food industry and
consumer groups supported dietary change. There were only a few individuals
and groups interested in pointing out the “gaps in knowledge” (Garrety, 1997:
746).

She argues that the collection of knowledge claims linking dietary fat, cholesterol and
heart disease acted as a boundary object (Gieryn 1999) ‘with different meanings uses
and implications in different social worlds’ (Garrety, 1997: 755). Her conclusion is
that the popularity of the cholesterol hypothesis led to boundary struggles as its
opponents attempted to reverse the scientific consensus. Attempts to adjudicate these
opposing claims led to sites where ‘the worlds of science and policy merged’ as the
‘proponents of dietary change eventually jettisoned their requirements for “definitive
evidence”’ (Garrety, 1997: 757).

Stephen Hilgartner also writes about the public role of science (Hilgartner, 2000,
Hilgartner and Nelkin, 1987). He has used a description of the public controversies
around three of the National Academy of Science reports on diet and nutrition to
illustrate his analysis of scientific advice as a form of performance.

The ability to offer authoritative advice is obviously not an entitlement,

automatically bestowed on any group that seeks it, but something that advisors

67
 must actively assert, cultivate and guard – sometimes in the face of intense
opposition. How do advisory bodies lay claim to the cultural authority of
science? How do they cast themselves as trustworthy advisors? And how do
they create credible voices for themselves? (Hilgartner, 2000: 5).

In this way he addresses both the development of a consensus and the outbreak of a
controversy, since for each report he shows how the credibility of scientific advice
was produced, challenged and sometimes sustained. His detailed description of the
production and reception of these three reports demonstrates the techniques of
persuasive rhetoric, stage management and information control that underpin
successful production of scientific advisory reports:

this study suggests that struggles over the enclosure and disclosure of
information play a far more important role in stabilizing (and destabilizing)
scientific texts and knowledge than most recent work in science and
technology studies has recognized (Hilgartner, 2000: 149).

All attempts at authoritative public statement of science-based advice will face

similar problems. Consequently, Hilgartner’s analysis is relevant to the reception of,
for example, the obesity policy advice which I analyse in this thesis. However,
Hilgartner’s research was carried out in America, where policy processes are
relatively transparent, and after a high profile public controversy where material was
leaked to the popular press. This gave him access to a lot of ‘backstage’ material
about the production of each report which is not available in all institutional contexts
(see section 1.8.3).

1.6.4 Defining obesity science

In the case of public health policy on diet and health the resulting scientific
knowledge is a new construct that draws from many areas of biomedical research –
physiological research into the regulation of body weight in rodents and humans,
clinical research into the effectiveness of specific weight reduction treatments and
large scale epidemiological studies into the links between lifestyle factors such as
diet, physical activity and smoking and chronic disease. As Jasanoff (1990: 227)
argues it is a hybrid form that is both profoundly shaped by public health policy and
has an important influence on subsequent policy. Drawing on these ideas of policy
science as specific kind of construct, throughout the following chapters, I use the term

68
‘obesity science’ to refer to this particular body of knowledge that is the product of
scientific advisory processes involving a variety of different social actors and
incorporating results from many different areas of contemporary biomedical research,
as well as other disciplines such as health economics and social psychology.

1.7 The making of health policy

Because of their focus on shared knowledge and the social processes that establish
and maintain it, STS approaches mesh well with empirical analyses of the making of
health policy that highlight the role of networks and communities. Gill Walt writes
about the role of policy communities which are

networks of individuals from various institutions, disciplines, or professions,

and in the health field may be practitioners (health professionals) researchers
(academic epidemiologists or parasitologists) or commentators (medical
journalists). The health policy community might also include pharmaceutical
companies, hospital administrators, any interest groups and members of
government (Walt, 1994: 110).

She describes how policy communities form a nexus for the exchange of information,
some of which will reach government policy makers:

Policy communities provide a number of different fora in which the early

stages of opinion formation and consensus building among experts takes place
(scientific meetings, journals, newspapers) although it may take years for ideas
to diffuse broadly, especially when they are critical of current policy (Walt,
1994: 110-1).

These policy communities may come together to form larger groupings that she labels
policy networks, citing the example of the international network of development
economists, family planning advocates and health professionals who developed and
promoted the idea of primary health care (Walt, 1994: 111). Other authors extend this
terminology to talk about policy sub-systems, issue networks and ‘iron triangles’14
(Buse et al., 2005: 111). Membership of the larger discourse or epistemic
communities that form interest networks is defined by ‘shared political values and a
shared understanding of a problem, its definition and its causes, though usually

14
The authors define ‘iron triangles’ as very stable small sub-systems of such networks.

69
marked by detailed disagreements about policy responses’ (ibid.). On the international
scale there are also global policy networks and which unite the domestic bureaucrats,
elected officials and interest groups of the iron triangle with international
organisations and global civil society (Buse et al., 2005: 153). Analysing the
development of public health policy in these terms encourages a focus on coalitions
that form around particular health issues, and the shared ways in which members of
these coalitions frame their concerns and make their arguments.

1.8 Methodological issues

This section begins with a narrative of the ways my methodology evolved during my
research. I then discuss my use of documentary sources, outline the criteria I used in
the selection of my sources and conclude by describing some of the issues I
encountered when studying the WHO.

1.8.1 Narrative of an evolving methodology

The starting point for this research had been my sense that the term ‘obesity epidemic’
marked a new way of understanding the health consequences of increasing rates of
excess body weight. Therefore, the first stage of my research involved mapping the
usage of the term ‘obesity epidemic’ – who used it and where? – and beginning to
locate the documents that would become my key sources. The use of documents as
my primary source material seemed appropriate at this stage as it would enable me to
produce a detailed account of specific stages in the development and spread of this
new body of knowledge. It would also provide this knowledge within its
contemporary context, rather than recounted from the context of the present. I began
by searching the popular and scientific press15 and looking for the sources of their
coverage of this topic. I quickly realised that both the popular press and the scientific
press were reflecting and recycling the language of official reports and policy
documents, most notably the WHO technical report of 2000. This was the first stage
in my understanding that the discourse of an obesity epidemic was a body of
knowledge created in the making of public health policy – it was not coined by the
popular press re-interpreting the technical language of the scientific press, nor did it

15
The results of some of these searches are discussed in chapter 3

70
arise directly out biomedical research. The obesity epidemic discourse was also a
scientific consensus that very rapidly came to be part of a taken-for-granted public
discourse.

Using library searches and following citations, I used a snowballing strategy to gather
more published sources and become familiar with this body of literature. I was struck
by the increasing importance given to excess body weight within chronic disease
epidemiology and public health policy. Judging only by the conditions of their
production and their contents, the sources I have assembled demonstrated that,
between the 1970s and the 1990s, excess bodyweight became more intensively
studied and more widely accepted as a significant public health problem. At the same
time I realised that twentieth century medical interest in excess bodyweight initially
derived from its role as risk factor for coronary heart disease (see chapter 2). This led
me to investigate earlier official reports on coronary heart disease and historical
accounts of the development of chronic disease epidemiology in the first half of the
twentieth century, in order to understand the pre-history of obesity science.

I drew up a timeline to key events and documents in order to understand the

development of both chronic disease epidemiology and obesity science. This led me
to understand obesity science and discourse within a broad periodisation whereby
obesity and overweight were first studied within chronic disease epidemiology as a
risk factor for heart disease, then, as obesity science developed, as a condition in their
own right, and finally as a rapidly increasing and globally threatening disease. As my
research progressed, I learnt that the distinction between a disease and a risk factor
can be more fluid than I first understood it to be. However, this does not
fundamentally invalidate my description of the changing status of excess bodyweight
or the observation that these changes were closely associated with the increasing
attention paid to the problem of excess bodyweight within British public health policy
in the 1990s.

In the course of this mapping process, I found that certain names recurred. There were
a few very active individuals – George Bray, Philip James and John Garrow – and a
wider group of regular participants that evolved gradually over the thirty year period,
as individuals retired or focused on other aspects of their research. I began to see that

71
a concern with excess bodyweight had united a group of researchers in the fields of
nutrition and chronic disease, and that these researchers had formed a network or
coalition (see section 3.2) to persuade others of the importance of this new public
health problem and lobby governments for action in this area. By using key word title
searches of library catalogues and databases (such as Ingenta and Web of
Knowledge), checking bibliographies, references to previous reports and the activities
of key individuals, I continued to assemble the collection of key sources listed in
Appendix 1 (see also section 1.8.3). This process was aided by witnessing a
presentation by Philip James at a national conference in 2003 where he listed all the
reports on obesity that he had been involved in producing over the last 30 years.

Philip James was a key participant in the network or coalition that developed around
excess bodyweight but I did not restrict my searches to reports he had worked on. I
also followed the activities of other researchers. As part of this tracking process, I
examined the continuities between these reports and the earlier ones into coronary
heart disease and diet-related chronic disease – both in terms of individual participants
and framings of the material in the reports. I used obituaries and accounts of particular
projects, such as the Seven Countries Study or the WHO MONICA project (see
section 2.5), as a way of finding out more about the work of particular individuals and
groups.

In order to analyse the developing understanding of excess bodyweight in this period,

I decided to analyse the documents I had assembled in more detail. The background
reading I had done in the early stages of my thesis research led me to outline a series
of key changes that had taken place in understandings of excess bodyweight during
this period. These changes took place in the areas of definition, measurement, health
effects, causes, treatment and economic costs of excess bodyweight. I began to see
that these two models relied on very different kinds of scientific data and had very
different implications for treatment and prevention recommendations. Because of
these differences and the fact that they largely operate on different levels, I argue that
collectively they amount to a shift from one paradigm of excess bodyweight to
another (see section 8.8). I have tried to encapsulate these shifts in the six questions
listed below:

72
1. What is the prevalence of excess bodyweight - relatively low and stable or high
and increasing?16
2. How is excess body weight defined and measured - using Metropolitan Life
Insurance Company ideal weight charts and a range of indices or using the Body
Mass Index (BMI) and standard cut-off points?
3. What are given as the health consequences of excess body weight –a relatively
short list of complications, such as an increased risk of diabetes, CHD and
mechanical problems such as osteoarthritis, or a much longer list of conditions,
including some cancers, where the relative risks are elevated?
4. What are seen as the causes of excess bodyweight – is it the result of individual
failings such as over-eating and inactivity or is it the symptom of the widespread
adoption of inappropriate diet and sedentary lifestyle?
5. Should the therapeutic focus mostly be on individual treatments such as weight
loss diets and drugs, or on population-level preventative measures such as
taxing/banning certain foods or redesigning urban environments, aimed at
populations?
6. Who pays for obesity – are the costs of obesity primarily born by individuals or by
healthcare systems?

Because of their either/or format these questions are obviously simplified and
schematic, but they were helpful as a framework to analyse the contents of my chosen
sources and my results are summarised in the table provided in appendix 2. These
results show a pattern of incremental change as successive enactments of this body of
knowledge shifted the categories used, incorporated more research results and
enlarged the scope of the pubic health problem of excess bodyweight and the potential
solutions offered.

In the final stages of my research, I went back to my library catalogue and database
searches to make sure that I hadn’t missed any important reports. Once I was sure I
had achieved saturation, I stopped searching. In addition, I searched a complete list of
the WHO technical reports in order to check my understanding of the timing of its
activities in the area of chronic disease epidemiology. I also decided that for periods
such as the mid 1970s and late 1980s, when there were no reports on obesity
published I would analyse the contents of textbooks as they can be considered as a
comparable source of authoritative knowledge (Kuhn, 1970).

16
This is my interpretation of the epidemiological data provided: all the authors I studied agreed that
excess body weight was an important public health problem, and because of this basic understanding of
the problem, they argued that rates are high and increasing even in the absence of large-scale
prevalence data demonstrating such a pattern.

73
At this point, I also began to think more concretely about the status of the knowledge
that I had derived from these documents. Analysing this collection of documents has
demonstrated the evolution of a body of policy-orientated biomedical knowledge
about the causes, consequences and possible policy responses to increasing rates of
excess bodyweight. It showed a shift between two understandings of excess
bodyweight that I label the individual and the environmental paradigms of obesity.
Each of these texts can be considered as an enactment of a body of shared knowledge,
as a policy science paradigm. Each contains an authoritative and contemporary
account of knowledge about this subject – they were usually consensus statements of
what could be said about particular topics or areas of research – and therefore
functions as an exemplar within this area of research. Successive documents re-
iterated this body of knowledge and developed it so that the understanding of excess
bodyweight mobilised by obesity science slowly evolved. By the 1990s this evolution
had led to the development of a new understanding of obesity and overweight as a
population health problem with environmental causes, rather than an individual
problem with solely behavioural causes.

However, there are important limitations of such documentary based analysis. Firstly
analysing documents can give very little information about the conditions of their
production. One way to overcome this would have been to combine such an analysis
with archival research or interviews of key participants. At the time of my fieldwork I
was unaware of the existence of the possible UK-based archives, such those of the
Department of Health, and so did not consult them. For the WHO documents, archival
research would have been of limited use since the organisation has a twenty year
embargo in place. I did conduct a few interviews in the early stages of my research
but I used them to inform my research rather than a source of research data. On
reflection, I eventually realised that it might be useful and interesting to produce a
more detailed account of the negotiations behind scientific consensus around the
obesity epidemic. But I also realised that this would require other methods and further
research. Interviews would have potentially offered a valuable source of additional
data. They could have provided information about the negotiations involved the
production of these reports, which would have usefully supplemented my textual
analysis. This layering of perspectives would have amounted to a form of
triangulation (Ritchie, 2003: 43-44), and added depth to my findings (see also section

74
1.8.4 below). However, by that time I had more than enough research material for a
PhD and insufficient time to interview such elite government scientists.

1.8.2 Use of Documentary Sources

There are several methods I could have used to investigate the changing discourses
around obesity. I considered carrying out my research by interviewing research
scientists and epidemiologist to analyse whether their understandings of obesity have
changed and what effect these changes might have. Instead, I have chosen to make my
research a primarily document-based case study that analyses writing about this
issue.17 Documents18 and interviews can be seen as epistemologically equivalent
forms of evidence. That is to say neither provides a complete and authoritative
account of a social phenomenon such as obesity science (Kvale, 1996, Platt, 1981).
Like interviews, documents represent a partial account of social phenomena which
must be analysed within the social and historical context where they are produced and
the uses to which they were intended to be put.

I chose the method of a document-based case study because I was interested in

providing a detailed account of the development of a body of shared and widely
accepted knowledge. I believe that document-based research is particularly useful in
tracing processes of concept formation and development. Actors’ recollections of
even the recent past can be inaccurate due to the effects of overlaying present ideas
onto past events; consequently, interview evidence about past events, and especially
about past thought processes, can often be unreliable. Moreover, recollection of
scientific developments, in particular, tends to be heavily influenced by positivistic
ideas of truth and discovery (Lowy, 1990). Document-based analysis can overcome
this effect, and therefore provide a more nuanced account of how concepts developed
and specific details of how they were mobilised at particular times and in particular
contexts.

17
I have conducted interviews with 6 individuals – epidemiologists, public health workers and social
scientists who have conducted research in the area of obesity. These were helpful in developing my
knowledge of more recent framings of specific issues, but I did not make use of them directly in the
following chapters.
18
I define document as including writing in the scientific and popular press, official reports and
government publications and the contents of relevant websites.

75
Writers on documentary analysis argue that the major issues to be considered by
researchers are authenticity, credibility and representativeness (Platt, 1981, Scott,
1990). My research is based on contemporary documents and so authenticity is not a
problem. Expert committee reports and other policy science publications are produced
with the aim of being authoritative accounts of particular bodies of knowledge, and
this gives a particular status to the knowledge contained within them, and a high
degree of credibility. These reports are consensus documents (see below) and are,
therefore, central to the charting and explanation of changes in the policy science
consensus about the obesity epidemic. Textbooks were a less important source in this
context but I used them to supplement the available reports because they are
considered paradigmatic accounts of disciplinary knowledge (Kuhn, 1970: 43).
Meeting or symposium proceedings and reports (in this field at least) function in a
similar fashion. All these publications are repositories of a common body of
knowledge about a particular topic, which makes them central to the analysis of the
development of a shared body of consensus knowledge. Interviews with participants,
such as expert scientists or civil servants and archive research would have enabled me
to describe the micro-politics behind the policies being recommended but would have
also diluted my focus on the development of a shared body of scientific knowledge
about obesity and overweight.

I have treated the reports I study as ‘consensus reports’. Hilgartner’s description of the
production of National Academy of Science reports explains how such reports are
produced:

Each consensus report is prepared by an expert panel, expressly convened for

that purpose. These expert committees do not perform new research, but
review the literature, hold workshops and symposia, and consult with relevant
specialists. The Academy instructs its committees to strive to reach unanimous
agreement on their findings and recommendations (Hilgartner, 2000: 23).

In my analysis of the reports I consider, I have assumed that they are produced in a
similar fashion, so that they have been jointly authored and produced on a consensus
basis. I think they can be considered, in a very specific sense, as a lowest common
denominator - a body of consensus knowledge that can be agreed upon by all
participants. I address issues of representativeness in the next section.

76
1.8.3 Selection of Sources

I have focused on policy writing in English. Despite the parallels that are drawn
between increasing rates of obesity and overweight in Britain and America, I have
looked at the development of obesity science primarily from a UK perspective, only
using US and other international sources where these represent an important input into
or extension of the activities of UK obesity scientists. This was for both practical and
academic reasons: I thought a wider perspective would hugely increase the size of my
data sample, and I did not think I had sufficient knowledge of American policy
processes and institutions, at this stage in my career, to provide a properly nuanced
analysis of such material.

Obesity science is also an international discourse and some of my material reflects

this, such as the WHO technical reports and the heavy reliance on American data in
earlier publications (see chapter 3). Given the international nature of contemporary
biomedical science, this is to be expected, and it also means that a strictly
geographical analysis is not possible. Medical writing in English is an international
discourse, although much of it still appears in journals published in the US and UK.
However, international professional networks and research collaborations mean these
articles are reporting on research that is carried in many different countries. Such
considerations mean that it would be difficult to use geographical categories to
exclude authors or research in any meaningful fashion. Therefore, I have analysed this
discourse as one entity using its existence in English as a method of demarcation,
rather than any hypothetical country of origin. Moreover, it would be
counterproductive to attempt to exclude American research from my analysis, since it
was such an important resource for British authors.

A bibliography on the topic of obesity published in 1974 (Whelan and Silverstone,

1974) begins with a publication from 1964, but the majority of its entries date from
after 1969. This is consistent with my searches of UK academic libraries. From the
late 1960s onwards, a small number of publications begin to appear; several of these
are reports of symposium proceedings, but there is also one publication from the
Office of Health Economics and a couple of textbooks (see below). A search of the

77
COPAC database of British research library catalogues 19 using the keyword ‘obesity’
gives 8 publications from the 1950s, 20 publications from the 1960s and 80
publications from the 1970s. Because of this pattern of growth, I have used two
different criteria in my selection of sources. For the 1960s, because writing on this
topic is relatively uncommon, I have tried to find early reports, conference reports and
textbooks.20 From 1970 onwards, as there is much more material available, I selected
all the official reports on the topic of obesity and textbooks produced by high profile
individuals who were regularly participating in the production of obesity science,
usually as members of expert committees.

In 1969 the Office of Health Economics (OHE)21 published the earliest British report
on the topic of obesity I have been able to locate. This report was produced when
relatively little was being written on the topic of obesity and appears to have been
largely ignored by researchers and practitioners. I have considered it, therefore, not as
an example of an influential source, but to provide an example of what could be said
about the topic at this time. The OHE produced another report on obesity in 1994,
again as part of their series on the economics of diseases which now appears to be
moribund. Many of these seem to be areas where new drug treatments could be
expected to be developed, and in this context their 1969 report seems like an
opportunistic, and unsuccessful, attempt to contribute to a developing discourse.

The next report produced jointly in 1977 by the Department of Health and Social
Security (DHSS) and the Medical Research Council (MRC) considered research into
obesity. It was a survey of an emerging field that focused heavily on physiological
research into the human metabolism and endocrine system. In 1983, the Royal
College of Physicians produced a report on the topic that demonstrated increasing
interest by clinical medicine practitioners in this area. Then there was a gap of 11

19
This count omits duplicates, non-English writing, unpublished theses and self-help/diet books in
order to try and give an estimate of scientific writing on the topic in English. Using a keyword search
should compensate for the fact that some writing about obesity is given titles referring to advances in
nutrition, dietary therapy or energy balance, rather than obesity, and it shows a similar pattern of
increasing publications on the topic.
20
An earlier symposium on obesity took place in 1963 at the Royal College of Surgeons but I have not
been able to locate a copy of the proceedings.
21
The OHE is a research organisation founded in 1962 by the Association of the British
Pharmaceutical Industry. Its mission is to ‘Support better health care policies by insightful economic
and statistical analysis of critical issues’ by providing ‘authoritative resources, research and analyses in
health economics and health policy’ (OHE, 2009a).

78
years (from 1983 to 1994) in the sequence of reports on obesity. However, in this
period scientific research on excess bodyweight and the links between diet and
chronic disease was well established, even if obesity was not seen as a politically
important health problem. Moreover, policy concern about healthy eating kept ideas
about the relationship between diet, health and body weight circulating in the non-
scientific press. In 1994 the UK government set up two taskforces on nutrition and
physical activity and in 1995 a report on obesity was published by the Department of
Health. In 1996 the Scottish Intercollegiate Guidelines Network produced a report on
the topic and the British Nutrition Foundation produced one in 1999.

The WHO reports I have included form a different series with a different rationale
behind their selection. The technical report of 2000, Obesity: Preventing and
Managing the Global Epidemic was the starting point of my research. It was a widely
cited authoritative source for the idea that obesity had become a major public health
problem. Having identified this and heard Philip James give a presentation about the
reports on obesity he had been involved in producing, I tracked back to the earlier
British reports that I have noted above. However, as my research progressed, I
realised that earlier WHO technical reports about cardiovascular disease and the links
between chronic disease and health had also been incorporated into obesity science
and I included technical reports from 1982, 1986 and 1990 on these topics. The 1995
technical report on anthropometry was included because Philip James participated in
its production and because its treatment of the topic of excess body weight is very
close to that of the 2000 report.

Finally I want to note that this series of publications illustrates the success of the
public health coalition that formed around obesity and overweight. The first report
was produced by a think-tank and ignored, and in the 1970s and 1980s the topic was
written about by an ad-hoc committee and professional medical organisations.
However, by the mid 1990s, reports were being published by the Department of
Health and, in 2000, by the WHO. This ascent continued in the next decade as obesity
became a policy issue spanning many UK government departments (Government
Office for Science, 2007).

1.8.4 Issues in studying the WHO

79
In some senses the WHO is a well-documented organisation. It has always produced
reports of its activities and much of this output is well-documented, readily available
and now often online. Three official histories of the organisation have been published
(WHO, 1958, WHO, 1968, Litsios, 2008) covering the period 1948 to 1978.
However, official histories give an authorised account of events that does not dwell on
internal politics or controversial aspects of the organisation’s activities. There is also
some scholarly writing analysing the founding of the organisation and its history
(Siddiqi, 1995, Lee, 1998, Staples, 2006, Lee, 2009). These authors have provided
much useful information about the workings of the organisation’s structure, and the
history of high profile campaigns such as the Malaria Eradication Programme
(Siddiqi, 1995, Lee, 2009). However, I have been unable to find any secondary
sources describing the production of technical reports in general, or specific technical
reports such as the one on obesity published in 2000. Walt (1994: 144-5) writes of the
existence of international policy networks and the role of international non-
governmental organisations - such as the International Obesity Taskforce – but in the
absence of information about the production of these reports it is possible only to
speculate about the influence of particular individuals and groups.

I have also not been able to gather much about the organisation’s work in the area of
chronic disease epidemiology, beyond what is outlined in the official histories
referred to above and an official history of the WHO Regional Office for Europe
(Kaprio, 1991). This is because the WHO, like other UN bodies, does not yet appear
to function in a very transparent manner. For example, unlike in government or
academia, it is often difficult to establish where even relatively high profile
individuals are currently employed within the organisation, or what posts they have
held in the past. Again some of these issues derive from my use of exclusively
documentary sources, rather than supplementing them with interviews. Because the
WHO is such an opaque organisation, and the development of health policy in Britain
has traditionally relied on the recommendations of expert committees who deliberate
in private (Jasanoff, 1997: 228), access to the kind of ‘backstage’ material that
Hilgartner (2000) uses would have involved a different research approach based on
interviews and archive material. Interviews with key participants would have
provided me with greater understanding of the processes involved in the production of

80
these documents (see section 1.8.1 above) and the additional perspectives generated
by such interviews would have allowed for triangulation of my data (Denzin cited in
Lewis and Ritchie, 2003: 276). However, as I outlined above (see section 1.8.2
above), I initially set out to examine an agreed on body of policy science knowledge
and how it changed over time, so, at first, I was less interested in the kind of
‘backstage’ detail that would have been available from interview-based research.

1.9 Childhood obesity

I have explicitly excluded consideration of childhood obesity from this thesis. Since
2000, rising rates of excess body weight amongst children has become a high profile
issue in health and education policy (SIGN, 2003, WHO, 2006). It is also an issue that
was discussed in some of the reports I analyse in later chapters (DHSS/MRC, 1976,
RCP, 1983, SIGN, 1996, WHO, 2000). Excess body weight in children was seen as
important on the grounds that fat babies become fat children and then fat adults.
However, reports from the 1970s described the evidence for this assertion as
‘fragmentary and sometimes conflicting’ (DHSS/MRC, 1976: 11, see also RCP, 1983:
23). But by the late 1990s, the authors of another report could describe infancy and
childhood as one of the ‘critical period’ for the development of obesity and cite data
demonstrating the link between obesity in childhood and as an adult (BNF, 1999: 49).
Moreover, the definition and measurement of overweight and obesity amongst
children and young adults was a contested issue. In 1983, one author argues that there
is a need for a standard measure using a reference population so that studies of
different populations can be compared. The WHO had collected such data for the
American population, but there was still the problem of how to express individual
children’s weights and define obesity and overweight (RCP, 1983: 18-19). By the mid
1990s, British reference growth charts for children had been developed and limited
data suggested that 20% of children and adolescents in one area of Scotland might be
overweight (SIGN, 1996: 4,9). However, at the end of the decade, a WHO report
stated that there ‘has not been the same level of agreement over the classification of
overweight and obesity in children and adolescents as there has been in adults’
(WHO, 2000). Part of the problem in drawing up such classifications was that, unlike
in adults,

81
 In children….BMI changes substantially with age, rising steeply in infancy,
falling during the preschool years, and then rising again during adolescence
and early adulthood. For this reason, child BMI need to be assessed using age-
related reference curves (WHO, 2000: 12).

Differences between populations and lack of national data were also confounding the
task of constructing a globally applicable childhood reference population (WHO,
2000: 12-3). Despite these problems, the report’s authors argued that

whatever method is used to classify obesity, studies of this disease during

childhood and adolescence have generally reported both a high prevalence and
rates that are increasing (WHO, 2000: 32).

The results of available research studies showed that the prevalence of overweight had
doubled amongst 5 to 24 year olds in Louisiana between 1973 and 1994; the rates of
obese schoolchildren in Japan aged 6 to 14 had increased from 5% to 10% between
1974 and 1993; and the prevalence of obesity among Thai children aged 6 to 12
increased from 12% in 1991 to 16% in 1993 (ibid.). This was similar to the situation
for adult rates of overweight and obesity (see chapter 4) and, in this case, these very
scattered results provide a weak evidential basis for inferring the existence of a global
public health problem.

Because of their dependency on adults for food and the structuring of their time by
compulsory education, the causes of overweight and obesity amongst children were
framed differently those for adults. They were discussed in terms of infant feeding
patterns (DHSS/MRC, 1976: 18-19), low levels of physical activity (BNF, 1999: 90,
120-2, WHO, 2000: 121), television viewing (WHO, 2000: 121) and food advertising
(WHO, 2000: 133). And finally, due to their ongoing development, the suggested
treatments for overweight and obese children and young people were also different
from those for adults. Because they are still growing, children need adequate
nutrition, which is difficult to reconcile with treatment for excess weight. One author
considered that between the ages of 5 and 10, measures to limit weight gain rather
than lose weight would be sufficient due to further growth (BNF, 1999: 145). A later
report accordingly describes treatment as involving reducing energy intake, improving
dietary quality, increasing physical activity and reducing sedentary behaviour –
although pharmaceutical and surgical treatments could be considered for children with

82
‘potentially fatal complications of obesity’ (WHO, 2000: 226-8). These are measures
that are considered to prevent weight gain rather than promote weight loss, as would
be appropriate for adults, but they are also described as preventive of adult weight
gain since ‘childhood obesity substantially increases the risk of adult obesity’ (ibid.).

1.10 Practical Issues and Limitations

The main constraint that I have encountered has been obtaining information of
processes within the WHO (see above). If I had managed to interview high profile
participants, like Philip James or John Garrow, this might have been less of a
problem, but I would have had the additional issue of integrating their retrospective
accounts with the material from my documentary sources. This would be an important
and interesting task, possibly for future research now that I have a firm grasp on the
development of the field of obesity science. Other limitations of this research are that
it focuses on a literature of public health policy advice and recommendations
contained within it. Such writing cannot be assumed to reflect clinical medical
practice or the planning of health care. Again further research would be required to
establish whether such writing has concrete effects, and what they are in different
areas of medicine.

Primarily, I intend this to be a piece of research for its own sake that adds to
knowledge in my chosen field. I did not plan to carry out the kind of research that
feeds directly into policy making. As it provides a historical account, rather than an
analysis of contemporary initiatives, it may also be of less interest to policy makers,
despite arguments for the use of history in this context (Berridge, 2000, Berridge,
2008). However, obesity science and the links between diet, nutrition and health
remain a topic of both political and media interest and so I anticipate that this research
may be seen to have some relevance for public health policy. If that is the case, then I
hope that my research can provide a different perspective on the health implications
of excess body weight, and demonstrate links between this and other contemporary
policy issues, such as health inequalities.

1.11 Conclusion

83
To conclude this introduction I will give a brief outline of the chapters to follow.
In chapter 2, I outline the history of post-war chronic disease epidemiology which
developed out of changing patterns of population health and a growing concern with
rising rates of coronary heart disease. I show how this led to a new understanding of
chronic diseases, the risk factor approach. In chapter 3, I discuss the formation from
the 1950s of a public health coalition around one such risk factor – smoking – and
describe how nutrition researchers and their allies began to develop another coalition
around the topic of excess body weight. I do this by analysing a series of reports and
textbooks on the topic of obesity that were published between 1969 and 2000. I
outline this coalition’s initial model of the condition, which I label the individual
paradigm. In chapters 4 and 5, I describe two important areas of obesity science – the
development of the BMI and BMI-based models of overweight and obesity, and the
energy balance model and its explanations of the causes of obesity. Chapter 6
analyses these topics in more detail, discussing the advantages of BMI for obesity
science and using approaches from medical sociology to outline the limitations of the
energy balance model. In chapter 7, I move to the international forum and discuss the
development of knowledge about chronic diseases and excess body weight in World
Health Organization (WHO) publications from the early 1980s onwards. I
demonstrate how increasing incorporation of the results from epidemiological studies
and a new institutional context led to changes in the shared model of obesity and
overweight. Chapter 8 analyses a WHO technical report of 2000 on obesity as a
global epidemic, explaining further changes in the understanding of obesity that
constitute a new model I label the environmental paradigm. In chapter 9, I analyse this
new model, discussing the framing of obesity as a global health problem despite the
paucity of data and the implications of explanations that rely so heavily on accounts
of the effects of modernisation, rather than other plausible causes such as poverty.
Chapter 10 concludes with a discussion of the conclusions, both empirical and
theoretical, that can be drawn from this case study of the development of public health
policy knowledge.

84
CHAPTER 2: HEART DISEASE AND THE EMERGENCE OF RISK
FACTOR RESEARCH C.1945-1980

2.1 Introduction

This chapter discusses a profound shift that took place in the medical research and
practice of industrialised countries after the Second World War. It describes the
growth of concern about rates of chronic disease, such as heart disease and cancer,
and the new research methods that were developed in order to understand the causes
and treatments of these conditions. The most important of these new methods was the
large scale prospective epidemiological study, as exemplified by the Framingham
Heart Study which was set up in 1948 to investigate the prevalence and development
of heart disease (Dawber, 1980: 239). This study is still running and has had a major
impact on contemporary medical understandings of the causes of heart disease.

Using a model borrowed from the insurance industry, Framingham researchers

developed an understanding of heart disease as a multi-factorial condition that was
caused by a set of lifestyle-related attributes such as high blood pressure, cigarette
smoking and elevated blood cholesterol levels. These attributes quickly became
known as risk factors and this approach became a crucial element of the development
of the post-war epidemiological paradigm of coronary heart disease (CHD)
(Aronowitz, 1998: 113). This paradigm, and the wider field of chronic disease
epidemiology, is one important location from which much biomedical concern with
excess body weight – usually labelled as overweight or obesity – developed.
Overweight and obesity became the subject of twentieth century medical research due
to their status as risk factors for heart disease. This chapter, therefore, functions as an
outline of the immediate context from which British (and American) obesity science
developed.

2.2 The epidemiological transition and the epidemic of heart disease

Mortality statistics collected in developed countries, such as the UK and the US, show
a striking change in disease prevalence from the middle of the nineteenth century
onwards. Overall mortality rates declined, due to a decrease in death rates from

85
infectious diseases such as cholera, typhus, typhoid, diphtheria and tuberculosis that
had been the major cause of mortality and morbidity. However, this decline was
perplexingly associated with an increase in rates of chronic or non-communicable
diseases (NCDs). Certain types of these diseases, including heart disease, stroke,
cancer and diabetes, became seen as important sources of mortality and morbidity.
Discussing the natural history of chronic diseases – ‘the term that has become used,
rather loosely, for the malignant, metabolic and mental disorders that present mostly
in the second half of life’ - one British epidemiological textbook explains this new
understanding as follows:

In the ageing population of an affluent society, which has mastered many of

the infections and malnutrition, and has high standards of maternal and child
care, these chronic diseases increasingly dominate the practice of medicine
(Morris, 1964: 133).

This change is currently understood to have taken place across many different
countries and has been labelled the ‘epidemiological transition’ (Omram, 1971,
Susser, 1985b: 149-51, Szreter, 2007: 4-5). During this transition, societies are
understood to have shifted from a pattern of high mortality and high rates of
infectious diseases to one of low rates of mortality and higher rates of chronic
diseases (see section 1.2.2). Over a period of a hundred years, between the early
nineteenth and early twentieth centuries, such a shift was associated with dramatic
increases in average life expectancies in the industrialising economies of the UK,
USA and Western Europe.

As rates of infectious disease fell, widespread medical (and public) concern developed
about high rates of heart disease amongst certain populations: ‘the great epidemic
disease of the twentieth century’ (Rothstein, 2003: 191). Rothstein argues that despite
changes in classification and inconsistency in death certificates, the diagnosis of
coronary heart disease was rare in the late nineteenth century, and only became
sufficiently common to be included in medical textbooks in the 1920s: ‘Expressions
of concern and confusion during the 1920s indicated that they viewed coronary heart
disease as a new and growing problem in medicine’ (Rothstein, 2003: 199, see also
Lawrence, 1997). In the next thirty years, coronary heart disease was to become
recognised as an increasingly common cause of mortality and morbidity, especially

86
amongst middle aged men. By the 1950s chronic diseases were seen as a major health
problem in the US, and there was a developing consensus that the British population
was also suffering from rising rates of coronary heart disease:

It is widely accepted by those involved in health education, community

medicine and even medical sociology that since the end of the Second World
War, Britain has suffered an epidemic of heart disease. Epidemiologists tell us
that “coronary” or “ischaemic” heart disease is both the largest single category
of fatal illness among men and also the disease responsible for most of the
inequality between social classes in life expectancy (Bartley, 1985: 289).

This new concern with heart disease as an important health problem redirected
medical attention from infectious to chronic disease, leading to the production of
further evidence for the growing prevalence of CHD. The evidence derived from two
series of autopsy studies. One series took place in the 1930s, as doctors tried to gather
information about the prevalence of heart disease and the reliability of pre- versus
post-mortem diagnoses (Rothstein, 2003: 200 - 202). Such studies aimed to improve
diagnosis in the living, but a later series of post-mortems carried out in 1953, on
soldiers killed in the Korean War, was conducted in order to establish the prevalence
of atheroschlerosis (narrowing of the arteries, seen to be a precursor of CHD). The
results showed that the majority had some degree of hardening, which had been not
expected in such a group of relatively young men (Atrens, 1994).

Thus a large number of studies using different methods and populations all
found a real increase in coronary heart disease death rates that began about
1920. The increase also occurred at the same time in other westernized
countries. Unquestionably, at that time coronary heart disease began its rise to
become the great pandemic disease of the twentieth century in all advanced
countries (Rothstein, 2003: 203).

From the early days of cardiology, middle aged male businessmen had been seen as
particularly vulnerable to heart disease: ‘Business men leading lives of great strain,
and eating, and drinking, and smoking to excess, form the largest contingent of angina
cases’ (William Osler [1914] quoted in Rothstein, 2003: 206).22 Such accounts
focused on the dangers of affluent living and the stresses and strains of modern city
life. Studies based on Metropolitan Life Insurance Company data from the 1920s and

22
These ideas fed into theories of the Type A personality that were developed by American
sociologists from the late 1950s onwards (Riska, 2000).

87
1930s showed that coronary heart disease rates were higher for men than for women
with a ratio that decreased with increasing age (Rothstein, 2003: 206). However,
despite doctors’ willingness to believe that coronary heart disease was a disease of the
affluent, American and British studies in the 1950s showed that lower socio-economic
groups had higher rates of heart disease: ‘Coronary heart disease was similar to
practically all other major causes of death in attacking the poor with greater frequency
and severity than the rich’ (Rothstein, 2003: 208). But the stereotype meant that high
mortality from heart disease was seen as an important economic problem, since it
killed middle aged men who were key wage earners.

2.3 Investigating the causes of chronic disease

Mid-twentieth century doctors and policymakers proposed various explanations for

the epidemiological transition. The decline in infectious disease was usually attributed
to a combination of medical and social progress: improvements in public health, such
as sanitation and clean water, improvements in medicine, such as vaccination and new
drugs like sulphanilamide and penicillin, and rising standards of living were all seen
to have played a role (Susser, 1985b: 150). During this period, doctors and medical
scientists had developed a clear explanatory paradigm for infectious diseases based on
laboratory-based scientific knowledge of the infectious and pathological role of
bacteria and viruses. In the second half of the nineteenth century, the work of Joseph
Lister, Louis Pasteur and Robert Koch changed understanding of disease (Worboys,
2000) and ultimately led to the development of effective treatments for many of the
major infectious diseases. An important element of this new understanding of
medicine and sickness was the idea of specific causes of disease, since it built on
existing ‘germ’ theories to argue that infection by specific micro-organisms caused
disease (Weindling, 1993:193), rather than the environment or the patient’s individual
disposition. This meant that the search for the causes of, and cures for disease, should
be conducted in bacteriological laboratories. Because this ‘bacteriological revolution’
was so successful, it eclipsed other explanations for disease for much of the next fifty
years: ‘The cultural power of the bacterial model within medicine was such that
alternative, nonbacterial theories of infection ceased to be articulated’ (Worboys,
2000: 280). Even in disciplines such as epidemiology, which had strong links to

88
public health and traditions of social medicine, bacteriological explanations were
dominant in this period (Amsterdamska, 2005).

In contrast, the causes of chronic diseases were not known. Because such diseases
develop over several decades, researchers could not readily identify specific
aetiological factors as had been the case for infectious diseases (Rothstein, 2003: 5).
Elizabeth Fee describes the post-war situation as one of confusion and feuding
between medicine and public health:

The problem was that nobody knew how to prevent cancer or heart disease.
There was little agreement or clarity about the relevance of nutritional,
occupational or environmental health – or about any other aetiological factors.
The only approach to prevention upon which everyone could agree was the
need for screening and early diagnosis. The chronic diseases – cancer,
hypertension, diabetes and others – could neither be prevented nor cured on
the older public health and medical models; at best, they could be controlled
through screening, education and medical supervision (Fee, 1996: 250).

This explanatory gap meant that doctors tended to invoke other explanations,
especially ones that involved assumptions about medical and social progress, though
in often ambivalent or contradictory ways. The rise in chronic diseases was often
attributed to the fact that, due to the decline in rates of infectious disease, people lived
longer. Conditions such as heart disease and other cardiovascular diseases were
simply the result of older bodies or individual organs wearing out – hence the label of
‘degenerative diseases’. Distinguishing ageing from the diseases that often
accompanied it led to the hope that these conditions might be treatable:

Increasingly chronic disease was thought to result from specific and not
necessarily inevitable individual and environmental factors…This conceptual
change from degeneration to specific mechanism, along with the growing
optimism that specific policy and clinical interventions might prevent or
ameliorate the onset and course of chronic disease, prepared the way for
embarking on costly, large-scale population studies of CHD risk (Aronowitz,
1998: 122).

An alternative, more pessimistic, understanding drew on narratives of the problems of

social progress - what Rosenberg (1998) labels the ‘pathologies of progress’ and
others ‘diseases of civilisation’ – to argue that diseases such as CHD resulted from
changes in lifestyle resulting from increased wealth – hence the term ‘diseases of

89
affluence’ (Morris, 1964). The excesses of eating drinking and smoking that Osler
described as habits of wealthy businessman were thought to be becoming
commonplace, due to economic growth and increasing prosperity.

2.4 The development of risk factor research

Meanwhile, certain strands of research were developing new ways of investigating the
aetiology of chronic diseases, especially CHD, which replaced laboratory research. A
key resource was the actuarial investigations that had been taking place within the
American insurance industry. In the early twentieth century, American insurance
companies began to develop mortality tables, in order to be able to insure individuals
at above average risk, known as “impaired lives” or “substandard risks” (Rothstein,
2003: 63). Due to increasing prosperity, life insurance was a rapidly expanding
industry during the late nineteenth and early twentieth centuries and increasingly
precise physical examination was also being used by other institutions such as
railroads, schools, police forces and the army (Davis, 1981: 393). Insurance
companies employed local doctors as medical examiners who were instructed to
measure and examine potential policyholders with a variety of tools, including (in
1891) a set of scales, tape measure, ophthalmoscope, stethoscope, urinary test
apparatus and a sphygmomanometer (Davis, 1981: 398). It is therefore not surprising
that these companies ‘most systematically and effectively accumulated relevant
physical examination data and related to the health and disease potential of the largest
number of people in the United States prior to World War I’ (Davis, 1981: 394).
Insurance companies, especially the Metropolitan Life Insurance Company, had the
most comprehensive collection of data on trends in body weight and its relationship to
health until at least the 1960s, when governments started to collect such information
(Oddy et al., 2009: 225).

In 1903 the Specialized Mortality Investigation was published by the Actuarial

Society of America. This investigation of the feasibility of insuring ‘high risk’ or
‘impaired’ policyholders was based on data from 38 countries. The Joint Committee
on Mortality, another insurance company body, was founded in 1909 and published a
series of studies between 1912 and 1939 that identified new medical factors which
increased the risk of mortality. Rothstein argues that these studies are the origin of the

90
modern concept of risk factors for diseases such as heart disease and cancer. The
following excerpt summarises this process for one particular risk factor, body weight,
which is central to my research:

The most unexpected finding of the original 1903 study was the discovery of a
strong relationship between mortality and ‘build’, a construct that combined
height and weight. Using industry data gathered between 1909 and 1928 for
all male policyholders ages 40 to 49 and taking 100 as the average mortality
rate, those who were at least 25% overweight for their height had a relative
mortality rate of 141, those of average weight had a relative mortality rate of
86, and those who were 5% to 14% underweight had a relative mortality rate
of 77…..The low mortality rates of slightly underweight policyholders
astonished physicians who were busy treating gaunt patients sick or dying of
tuberculosis, the “wasting disease”. They viewed ruddy and rotund persons as
the epitome of good health and underweight ones as suspect. Although the
medical directors had no theory of disease etiology that explained the
statistical relationship, they accepted it unequivocally and made it a key factor
in selection [of policyholders] (Rothstein, 2003: 64).

Medical directors of insurance companies were more concerned with profits than
aetiology and so they accepted this new and unexpected relationship between build
and increased mortality. Their use of weight in medical examinations, the selection of
policyholders and the construction of mortality tables, made weight, or build, one of a
set of important medical risks that also included high blood pressure, diabetes and
kidney disease.

2.5 The American heart lobby and the Framingham Heart Study

A major research effort was seen as necessary to investigate the precise causes of the
epidemic of CHD and provide information on how to prevent cases in the future. The
actuarial method, with its focus on identifying risk, was recognised as a powerful new
method for investigating heart disease. During the 1940s, the American Heart
Association drew on fears of the growth of heart disease to raise funds and win
resources, leading in 1948 to the passage of the National Heart Act and the creation of
the National Heart Institution (NHI) within the National Institute of Health (NIH)
(Fye, 1996). The AHA had been founded in 1924, as the professional organisation of
the emerging speciality of cardiology. By the early 1940s the organisation was well
established but had chronic financial problems, and had limited success in raising

91
research funds from the insurance industry. In 1937, the National Foundation for
Infantile Paralysis had successfully used celebrity endorsements and political allies in
the White House to raise enormous sums of money in its March of the Dimes
campaign (Fye, 1996: 94-7). The success of this model of fund raising, combined with
the founding of competing organisations in the area of heart disease and hypertension,
led to the transformation of AHA into a voluntary health organisation that raised
funds for research into heart disease:

After World War II, AHA leaders joined with a group of legislators,
government officials and concerned citizens to form the “heart lobby” which
used a military metaphor to describe their goal. They told the American people
and their leaders that heart disease could be conquered if enough money was
given to support cardiovascular research. The success of the Manhattan project
to build an atomic bomb and recent medical advances, such as antibiotics,
helped them make their case (Fye, 1996: 86-7).

In the 1930s and 1940s there was a significant increase in funding for medical
research, firstly from private bodies such as the Rockefeller Foundation, and, later,
from the US government:

The AHA leaders looked beyond public-fundraising, corporate donations, and

philanthropists to support their twin causes: cardiovascular research and
researchers. They knew that the federal government was the most promising
source for research funding…Gradually a coalition of politicians, federal
officials, scientists and academic physicians succeeded in getting the
government to liberally fund research projects in private institutions (Fye,
1996: 102).

Leaders of the AHA were aware that they were competing with groups lobbying for
funding for other diseases, especially cancer. After the National Cancer Institute Act
was passed in 1937, the heart lobby used cancer lobbying as a model (Fye, 1996:
103). They recruited allies, such as Mary Lasker the wife of a rich industrialist, and
several politicians including the Surgeon General, Leonard Scheele. As a result of
intensive lobbying, the National Heart Institute (NHI) was founded in 1948 and the
First National Conference on Cardiovascular Diseases (sponsored by the AHA and
NHI) took place in Washington in 1950. The NHI continued to lobby Congress for
increased funding. In 1953, its director told Congress that 10 million Americans
suffered from cardiovascular diseases and that ‘[H]eart disease exacts an enormous

92
economic cost in medical and institutional care, in military management, and in
industrial production’ (quoted Fye, 1996: 160). In such lobbying, the newly
established heart lobby presented cardiology research as the search for a cure for
heart disease, an approach that united the academic and practitioner groups within the
AHA. This increased funding for medical research post-WW2 led to a dramatic
expansion in cardiovascular research activity, and US cardiology continued to expand
throughout the 1950s and 1960s (Fye, 1996: 181).

Also in 1948, the NHI and the University of Boston set up the Framingham Heart
Study. The researchers decided to locate the study in Framingham, a small town in
Massachusetts, because of its size and proximity to medical facilities in Boston, its
homogeneous population23, the existence of an annual census and previous successful
participation in a tuberculosis study between 1917 and 1923 (Dawber, 1980). In the
1950s Framingham researchers borrowed the idea of risk factors from the insurance
industry, and from 1961 the terms ‘factors of risk’ or ‘risk factor’ were being used in
study publications (Rothstein, 2003: 283). The first risk factors identified were age
and sex; subsequently, already identified risk factors of hypertension, raised blood
cholesterol, increased body weight and smoking were also re-identified (Dawber,
1980), and the researchers of the Framingham study were ‘among the first to
emphasize the dangers of smoking for coronary heart disease’ (Rothstein, 2003: 283).

An international consensus about the relevant risk factors to investigate began to

develop relatively fast. A 1957 WHO technical report on atheroschlerosis and heart
disease discussed the role of environmental factors such as calories, obesity and
overweight (labelled as dietary factors), physical activity, stress, alcohol and tobacco
and infections (WHO, 1957). In this report, the authors state that ‘obesity in itself is
not a primary factor in producing ischaemic heart disease’ (WHO, 1957: 17). In 1962,
in another technical report on the prevention of arterial hypertension and ischaemic
heart disease, factors indicating ‘exposure to increased risk’ were listed as high blood
pressure, high blood cholesterol, obesity, diabetes, heavy cigarette smoking and a
family history of cardiovascular disease (WHO, 1962: 19).

23
According to Rothstein the population was largely of European descent, more than half of whom
were of either Italian or Irish ancestry (Rothstein, 2003: 280).

93
The link between increased risk of heart disease and hypertension, raised blood
cholesterol and smoking seemed relatively straightforward, but the relationship
between diet and blood cholesterol was contested (Dawber, 1980: 141), and in turn
the link between raised blood cholesterol and increased body weight was found not to
be strong (Dawber, 1980: 129). Although men 20% or more above the median weight
had double the risk of heart disease (Dawber, 1980: 146), this uncertainty about the
relationship between raised blood cholesterol, diet and weight, gave overweight and
obesity an uncertain status as a risk factor for heart disease. This uncertainty persisted
in later publications where obesity is described as undesirable as it is associated with
more important risk factors (DHSS, 1974: 11, RCP/BCS, 1976: 2).

The success of the Framingham researchers in identifying risk factors for heart
disease led to the setting up of other large scale prospective epidemiological studies.
In 1958 Ancel Keys of University of Minnesota, with support from the WHO and,
once again, the NHI, set up The Seven Countries Study. Seven Countries developed
from earlier prospective studies into the risk factors for heart disease among middle
aged men, including a study of Minnesota business men begun in 1948 and two
studies on the epidemiology of coronary heart disease that began in Italy and Spain in
1952 (Keys, 1980b: 1). The study focused on diet as a risk factor for cardio-vascular
disease: building on his previous research, Keys and his collaborators hypothesised
that ‘differences among populations in the frequency of heart attacks and stroke
would occur in some orderly relation to physical characteristics and lifestyle,
particularly composition of the diet, and especially fats in the diet’ (Blackburn, n.d.).
The final version of the study included eighteen cohorts of men aged 40 to 59 of
varied occupations living in rural and urban areas, from Finland, Greece, the
Netherlands, Italy, Japan, the US and the former Yugoslavia. These areas were
selected for a combination of pragmatic reasons, such as research contacts who were
willing to collaborate at low cost, and dietary or epidemiological variability: Japan
had very low rates of heart disease at the time and Finland very high ones, and Greece
had high levels of consumption of olive oil (Keys, 1980b: 7).

Later high-profile studies that also adopted this new research method included the
Nurses Health Study and the Multiple Risk Factor Intervention Trial (MRFIT). The
Nurses Health Study was set up in 1976 by the NIH and the Harvard School of Public

94
Health, initially to investigate the health effects of long term usage of oral
contraceptives, and is currently still running having recruited a new younger cohort of
participants. MRFIT began in 1972. It was funded by the National Heart Blood and
Lung Institute (previously the NHI) to test the effect of interventions on several
different heart disease risk factors. This very costly study is widely agreed to have
shown that the expensive treatments given to the ‘special intervention’ group had only
a marginal effect on cardiovascular mortality rates. The organisers relied on
participants presenting themselves at screening centres, and volunteers in such studies
are invariably healthier than non-volunteers: ‘Thus a study designed to lower
mortality rates in a high-risk population used a sample with much lower-than-average
mortality rates’(Rothstein, 2003: 309). Rothstein argues that this high profile failure
was an example of a more general problem in clinical trials of risk factor behaviour
changes as they often relied on atypical and highly motivated participants.

The increased amount of American research funding available in the 1950s and 1960s
(Fye, 1996) meant that this first group of studies were organised by American
research institutions. However, in the 1970s, European bodies, such as the WHO
Europe Office, began to build on their involvement in the international projects to set
up their own studies. The most famous of these was the North Karelia Project, a
collaboration between the Finnish government, the Finnish Heart Association, the
University of Kuopio and the WHO, which ran from 1972 to 1977. This project
extended the methods of studies such as Seven Countries24 by including investigations
into the effects of interventions on key risk factors. An important component of this
study was an attempt to reduce the prevalence of risk factors, and therefore cardio-
vascular disease, in a population chosen because they had very high rates of those
diseases. The risk factors identified were smoking, blood cholesterol levels (or diet)
and hypertension. The project strategy was explicitly based on risk factors as the
following except makes clear:

1. There is no natural limit between a normal and an abnormal risk factor

level. The risk for coronary attack increases as the level of any risk factor
increases. One feature of the Finnish and North Karelian population was that
the average level of all risk factors was high (Rimpela et al. 1974). Practically
none of the adult population had a “safe” level of all risk factors.

24
North Karelia was one of the Finnish regions included in the Seven Countries study.

95
 For this reason, the intervention aimed at general changes of the risk factors
among the population (i.e. changes in the distributions and means) instead of
measures restricted to “high risk” groups (National Public Health Laboratory
of Finland, 1981: 18).

The North Karelia Project is still routinely described as one of the very few successful
health interventions in the area of cardiovascular disease.25 Its outcome is
summarised in the project report as follows: ‘The estimated CHD risk reduction from
all risk factors in North Karelia was 17.4% among men and 11.5% among women.
The greater change among men is in accordance with project goals.’ (National Public
Health Laboratory of Finland, 1981: 303)

A final large-scale epidemiological study that falls directly in this tradition of research
is the WHO MONICA project (MONitoring of trends and determinants in
CArdiovascular diseases). MONICA was a European prestige project set up to rival
Framingham: it was a self-conscious attempt by the European research community to
conduct a large prospective epidemiological study into cardiovascular risk factors,
modelled on the studies that had been running in the US since the 1950s. It began in
1978 and ran for 23 years, collecting data on 38 populations in 21 countries. Like
North Karelia, MONICA had strong links with the WHO Regional Office for Europe
in Copenhagen, although the project director was Professor Hugh Tunstall-Pedoe, a
cardiologist based at Dundee University.26 One commentator has described the
establishment of the ‘first global epidemiological study’ as ‘an amazing feat’ and
states that ‘it is almost unimaginable that an enterprise of this scale will ever be
conducted again’ (McKee, 2003: 613). I will discuss MONICA at greater length in
Chapter 8 as it was a very important source of data on increasing average body
weights in European populations.

2.6 Risk factor research and causal explanation

Risk factor analysis represented a new paradigm in research into disease and ill-
health, displacing laboratory investigations as the main mode of research and criterion

25
It has even been regularly described as the only successful public health intervention aiming at
reducing body weight, even though this was not one of the project’s aims and there is no available data
to demonstrate this success.
26
One of the other members of the MONICA steering committee was Professor Jaakko Tuomilehto
who was centrally involved in the North Karelia Project.

96
of proof. Its development in the UK in the 1950s has been described as a “paradigm
shift” that led to the ‘growing acceptability of epidemiological rather than the
biomedical, laboratory-based mode of proof’(Berridge and Loughlin, 2006: 957).
Partly, this was due to the exemplary status of Framingham: ‘The probabilistic yet
“hard” data produced by studies such as Framingham and their wide dissemination in
a variety of clinical, public health, cardiology, and other journals...beginning in the
mid-1950s played an important part in gaining visibility and acceptance for risk factor
ideas’ (Aronowitz, 1998: 121). Individual susceptibility to heart disease could still not
be satisfactorily explained and Aronowitz argues that this is another reason for the
spread of the risk factor approach:

Risk factors provided a new scientifically rationalized framework for

managing the increasing uncertainty associated with the occurrence of CHD
by providing an overarching, consoling, meaning-giving framework. Risk
factors provided a reassuring explanatory framework because they gave some
sense of who was at greatest risk and what one might do to decrease risk. At
the same time, risk factors embodied the cultural and medical ideals of
precision, specificity and quantification (Aronowitz, 1998: 125).

Risk factor analysis, being based on the identification of statistical correlation, does
not identify causes at all – merely associations. In particular, it does not throw any
light on biological causes. Luc Berlivet outlines the use of epidemiological evidence
in the 1950s debate on the links between smoking and lung cancer:

In the absence of comprehensive and accurate laboratory explanations, the

statistical-probablistic approach worked like a “black box” – the input i.e.
smoking, and the output, i.e. cancer of the lung, were associated through a
method that could not shed any light on the biological processes which lay
behind (Berlivet, 2003: 52).

Moreover, a number of historians have noted that the identification of risk factors as
causes has also tended to be distinctly selective. This warrants some explanation.

In principle, risk factor analysis is open to the possibility that multiple causes or
predisposing factors – including social and environmental factors – may contribute to
the production of disease and that such causes may interact holistically. This kind of
analysis was apparent for instance in early life insurance industry work to identify risk
factors. Prospective epidemiological studies such as Framingham borrowed the

97
concept of ‘risk factor’ from the life insurance industry which had been analysing
information on the mortality rates of their policyholders in terms of ‘risk factors’
since the beginning of the twentieth century. Rothstein argues that

Several basic differences existed between the old life insurance risk factor and
the new medical risk factor popularized by the Framingham study. The life
insurance risk factor was conceived in terms of a gradient of risk depending on
its level, while medical risk factors were often dichotomized into healthy and
unhealthy levels. Each life insurance risk factor was related to all other risk
factors, while each of the new medical risk factors was considered separately.
Last, the life insurance risk factor emphasized both the social and the medical
characteristics of the applicant, while the medical risk factor was restricted to
medical characteristics (Rothstein, 2003: 285).

However, Rothstein notes that subsequent risk factor research, from the Framingham
study onwards, epidemiologists of heart disease have tended to write out social and
environmental risk factors.

Despite some promising early findings, they disregarded social characteristics

such as education, income, occupation, living conditions, usual sources of
health care, marital status, place of birth, and family structure. Yet social
characteristics are as important as physiological ones in clinical
decisions…The narrow focus of this pioneering study established an
unfortunate precedent for most subsequent studies (Rothstein, 2003: 285).

How, and why, this occurred remains somewhat unclear, however.

Aronowitz attributes it in part to historical accident, resulting from the initial framing
of the Framingham study. One of the important and novel features of Framingham
was that it became a prospective study i.e. one looking for the factors that preceded
the development of heart disease. As originally conceived, the aims of the study were
far less ambitious than they subsequently became, and were oriented towards
identifying risks in the sense of clinically useful measurements that could be used to
predict the likelihood of heart disease in particular individuals (Aronowitz, 1998:120)
– i.e. not concerned with larger social and environmental factors. Only subsequently
was it decided to turn Framingham into a prospective study to identify risk factors
more generally. And although the study period was extended beyond its initial
duration of five to ten years for this purpose, the individualised perspective on risk
factors was retained. Despite the fact that the methodological innovations making it

98
into a large scale cohort study relating exposure to outcome of chronic disease were
later additions to the study, they were to be highly influential.

Additionally, Aronowitz identifies in the Framingham study a tendency towards

biological reductionism, in the sense of a tendency to see risks as real only insofar as
they can be explicated in terms of underlying biological causal processes

But not every association can be expressed in risk factor terms. Putative risk
factors need to meet certain conditions. They need to be measurable and
specific characteristics of individuals in order to fit into the risk equations that
express the results of epidemiological trials…Nonspecific and less
individualistic variables, even if they could be measured and manipulated as if
they were specific characteristics of the individual, have not been easily
assimilated into mechanistic models of disease and mainstream clinical and
public health approaches. Such variables generally lack a direct biological
mechanism by which coronary artery pathology develops in the individual.
Although knowledge about risk factors is almost entirely developed from
epidemiological observation, risk factors have been understood – and
legitimated – only as they contribute to the specific, localized
pathophysiological processes that result in disease (Aronowitz, 1998: 132-3).

Indeed, identification of risk factors for heart disease commonly led to clinical and
laboratory research aimed at identifying causal biological processes – e.g. Ancel Keys
conducted research on the human physiology of cholesterol consumption, and Pekka
Puska on the relationship between nutrition and hypertension, focusing on the effect
of blood lipids.

At the same time, this biological understanding of risk factors facilitated their
incorporation, at least in certain instances, into clinical medicine, and indeed to
certain risk factors coming increasingly to be seen as diseases in their own right.
Within clinical medicine, the acceptance of risk factor frameworks was less related to
overall funding and more to individual doctors’ economic and professional interests –
the availability of specific tests (and payment for performing those tests), the
availability of drug treatments, expert committee reports producing guidelines for best
practice, and campaigns by governments and health groups promoting preventive
health (Aronowitz, 1998: 127). A case in point is hypertension, long recognised as a
risk factor in heart disease. With the advent of effective drugs to treat hypertension,
clinicians began to pay greater attention to the condition. Initially, only those patients

99
showing unmistakable and acute symptoms were treated (Timmermann, 2006). But as
the idea of risk factors became increasingly widespread within the medical profession
and in policy circles, hypertension was redefined with the establishment of the
concept of “malignant hypertension” not just a risk factor, but as itself an occult form
of pathology, defined by guideline thresholds and made visible through routine
surveillance with the sphygmomanometer (Timmermann, 2006: 245-6). Similarly,
routine screening and treatment for hypercholestrolaemia in the US only developed in
the 1980s, despite the medical profession having recognised its importance as a risk
factor since the 1950s. This change resulted from campaigns by the National Institutes
of Health and the American Heart Association who had developed consensus
recommendations and treatment protocols, and pharmaceutical companies who had
developed new drug therapies (Aronowitz, 1998: 127).

Ironically, this narrowing of the scope of what kinds of factors can be considered
causal has actually weakened the explanatory power of the risk factor approach. As a
number of epidemiologists have pointed out, risk factor analysis, as it has developed
since Framingham, is only capable of explaining a certain proportion of variation in
rates of heart disease. In the 1970s, one American epidemiologist argued that current
risk factors were not capable of explaining the patterning of heart disease distribution
by geography or gender (Winkelstein quoted Aronowitz, 1998:133). The disregarding
of wider social and environmental causes of disease, such as the effects of bad
housing, may be one reasons why the risk factor approach is acknowledged to be able
to explain only a proportion of the variation of heart disease rates within populations
(Rose and Marmot, 1981).

Berlivet suggests that this exclusionary approach to risk factors is deeply embedded in
the methodological suppositions on which current risk factor research is founded,
particularly the criteria for attributing causation that were articulated by the British
statistician Austin Bradford Hill in the 1950s. At the end of his discussion of the
development of ideas of causality in chronic disease epidemiology, he asks

Could poverty be a “cause” of disease? Many, including several outstanding

epidemiologists would agree; still, [Bradford] Hill’s criteria [of causation]
accommodate more easily tobacco smoke, infectious agents or chemical
carcinogens than social and economic deprivation (Berlivet, 2003: 61).

100
But others attribute it to a systematically political orientation permeating biomedical
and especially epidemiological thought. Bartley describes how problems which are
‘too big to solve’ (Bartley, 1992: 220) can end up being sidelined or ignored in the
policy process. Due to their focus on a limited number of behavioural causes, risk
factor approaches to heart disease (and other chronic diseases) are one way that health
policymakers can avoid this problem of ‘bigness’. Such approaches have the double
advantage of seeming suitably scientific, but, at the same time, rendering problems of
chronic disease more manageable in policy terms, because, in terms of both treatment
and prevention, they allow the targeting of individual risk factors, often by
manageable methods such as drug treatments or behaviour change.

These processes of defining diseases and their causes demonstrate ways in which
‘medicine individualises disease and writes-out social deprivation and inequality from
the description of illness’ (Bartley, 1985: 292). Relating health inequalities to socio-
economic class or race can operate a form of black-boxing that obscures the
mechanisms creating these negative health effects.

Epidemiological approaches to managing race, class and sex/gender distill the

effects of social and relational ideological, structures and practices organised
around such differences into characteristics of discrete and self-contained
individuals……Epidemiology thereby renders invisible the very social relations of
power structuring material and psychic conditions and life chances that contribute
to the stratification of health and illness (Shim, 2002: 134).

This is a theme to which I will return throughout my thesis, by looking in particular at

how the social has been omitted from explanations of eating habits and patterns.

2.7 Has there been an ‘epidemic’ of coronary heart disease in the UK?

More radically, Bartley (1985) has challenged aspects of the idea of the demographic
transition itself, and specifically the idea that there was an increase in the incidence of
coronary heart disease after the Second World War. Bartley’s re-interpretation of
mortality statistics, using arguments about changing disease classifications, ‘throws
considerable doubt on the idea of one clearly distinguishable and temporally located
“modern epidemic of heart disease”’ (Bartley, 1985: 293, italics in the original).

101
Instead she argues that the health of certain groups within the British population,
notably middle-aged working class men, has ceased to improve in the last 50 years.
Earlier epidemiologists were aware of differential declines in mortality rates but
ascribed this to changing disease patterns:

During the nineteenth century death rates in middle age in England and Wales
were high…but about the turn of the century sanitary reform and the rise in the
standard of living began to show results in this age group. Death rates began to
fall, both in men and women, and they continued to fall until the early 1920’s.
Then rather abruptly, there was a change. Mortality in women kept its
downward course; but the decline in male mortality slowed, and for the last
twenty years has hovered around 22 per 1,000. A hundred years ago the death
rate among middle aged men was about 15 per cent higher than in women,
after the first world war it was about 33 per cent higher, now the male rate is
twice the female.….[gains from advances in medical science have been
counteracted by other changes]…The most important is that two diseases,
affecting men in middle age far more than women, and highly lethal, have
emerged from obscurity to become exceedingly common: “coronary
thrombosis” or ischaemic heart disease, and cancer of the bronchus (Morris,
1964: 1-2).

Bartley argues that this apparent post-war rise in coronary heart disease was actually
due to a reclassification of forms of disease that were already prevalent among
middle-aged men. From the 1920s onwards a conceptual framework was developed
by British and American doctors that understood hardening of the arteries, or
arteriosclerosis, leading to heart failure as a major cause of death (Rothstein, 2003:
286-94). Bartley tracks the changing terminology used to describe these conditions
and argues that in the 1930s, a large proportion of heart disease was still classified as
‘degenerative’ – that is, not caused by the blockage of coronary arteries. She
combines the mortality rates for ‘non-coronary’(degenerative) and ‘coronary’ heart
disease to show that between 1921 and 1939 rates of heart disease in men 45 to 64
increased by 85%, and then by 12% to 1951 and by 20% to 1971 (Bartley, 1985: 300).
This means that, contrary to narrative of a post-war increase in heart disease, the
increase in the interwar period was greater than that of 1950s onwards, and, in the
early post-war period there was continuing uncertainty as to whether rates for heart
disease were still rising. From her re-reading of contemporary accounts Bartley
argues that there is no ‘consistent evidence to the effect that the coronary arteries of
the nation deteriorated over the period we are examining, let alone that such
deterioration was more prevalent in those being classified at death as cardiac victims’

102
(Bartley, 1985: 302). Her conclusion is that ultimately, ‘we can take as our problem,
not the “epidemic of heart disease”, but rather, the failure of the health of men
(particularly working class men) in later working life to improve appreciably in the
last fifty years’ (Bartley, 1985: 309).27

Such a conclusion runs counter to the conventional narrative of British chronic

disease epidemiology, and especially those presented by public health policy
textbooks and policy reports from the 1960s onwards. This is not, as far as I can
judge, a widely accepted argument, even within Bartley’s home discipline of medical
sociology. However, such an analysis is important because it leads to questions about
the distribution of the improvements in health between different sections of the
population. In arguing that the high post-war rate of heart disease among middle-aged
working class men is not new, but rather is continuous with the ill health they suffered
between the wars, Bartley is arguing, in effect, that the health improvements enjoyed
by the rest of the population since WWII have largely by-passed this group. And if
that is the case, it raises questions about why this group, in particular, should have
missed out on the benefits that others have enjoyed. Bartley does not offer an
alternative explanation of this phenomenon. Any causal explanation would need to
include social factors, since what is at issue is a differential distribution of health
advantages and disadvantages across social groups. But, the risk factor approach, as it
has developed since Framingham, is ill-suited to clarify the role of such factors, since
it focuses on individual biological indicators of risk.

2.8 Conclusion

The idea of an epidemic of heart disease emerged in America and Britain during the
years following World War 2. Faith in modern medicine and biomedical research
accompanied by economic growth, led to significant increases in funding for health
care and for medical research in the US, and to a lesser extent the UK, for the next 25
years. Since increasing levels of heart disease had become a focus of medical and
political concern, some of this new funding was used to conduct research into its
27
Aronowitz states that one of the reasons prospective studies such as Framingham were undertaken
was the realisation that the positive effects of the epidemiological transition had by-passed the middle-
aged (Aronowitz, 1998: 123).

103
causes. Using the new technique of the large scale prospective epidemiological study,
this research programme developed and consolidated the risk factor approach to
understanding heart disease. Such development included an increasing tendency to
regard specific risk factors, for example hypertension and raised blood cholesterol, as
diseases in their own right. However, the risk factor approach to heart disease was
selective in the factors it considered as causes of disease. This meant that chronic
disease epidemiology, and indeed even the assumption that heart disease was a
peculiarly post-war public health problem, tended to obscure the social determinants
of heart disease.

The following chapters will build on these arguments by looking specifically at the
identification and definition of obesity as a risk factor for heart disease, then
increasingly as a health problem in its own right. In so doing, these chapters will
throw additional light on how the risk factor approach was articulated and
disseminated after its initial pioneering formulation in Framingham. And ultimately,
they will also lead to a further consideration of how that approach may have helped to
obscure the social determinants of ill-health.

104
CHAPTER 3: THE DEVELOPMENT OF PUBLIC HEALTH COALITIONS
1950 - 1999

3.1 Introduction

The previous chapter describes the post-war growth of a major research programme
into the causes of coronary heart disease (CHD), a condition which was seen as an
important and growing public health problem. This was a crucial element in the more
general re-focusing of medical attention towards chronic disease. The risk factor
approach to chronic disease that developed out of this research programme meant that
excess body weight became an object of increased medical interest. Excess body
weight was seen as one of a group of newly identified risk factors for heart disease
that also included smoking, hypertension, raised blood cholesterol, stress and physical
inactivity. Its status was uncertain (see section 2.5) but the links between raised blood
cholesterol, poor diet, excess body weight and increased risk of heart disease were
sufficiently convincing to generate further research in these areas.

At the same time British nutrition researchers were switching focus away from under-
nutrition and malnutrition and towards over-nutrition. A coalition of medical
researchers and practitioners interested in the topic of obesity and overweight began
to emerge in the late 1960s as nutrition scientists began to establish links with
diabetes researchers and those interested in the links between diet and heart disease.
At first this seems to have been a small offshoot of larger groups, but the success of
the anti-smoking public health coalition provided a role model, and gradually the
activities of the obesity science coalition expanded beyond their own professional
organisations to include collaborations with government departments and bodies such
as the Royal College of Physicians (RCP). In the course of this expansion scientific
knowledge about increased body weight grew into a new understanding that I label
the individual paradigm of obesity.

3.2 The development of an anti-smoking public health coalition in the UK

Smoking was the first of the newly identified risk factors around which a policy
coalition developed. The framing of smoking as a medical and public health problem

105
began in the late 1950s (Berridge, 1999b, Berridge, 2003b, Berridge, 2005a, Berridge,
2006, Berridge, 2007b, Berridge and Loughlin, 2006). Virginia Berridge (2003b)
provides a chronology of smoking policy in the UK, dividing it into four periods: the
1950s and 60s when the first epidemiological evidence of the link between smoking
and lung cancer was published; the 1970s when the government relied on voluntary
industry regulation in the face of increasing health activism; the 1980s when the
discovery of passive smoking shifted understandings of smoking from an individual
to an environmental issue; and the 1990s when a ‘medicalized public health’ focused
on pharmaceutical treatments for nicotine addiction. In each of these eras, the policy
situation was characterised by complex and overlapping networks of politicians,
industry representatives, voluntary and professional organisations and expert
government advisors.

In the first of these periods, Berridge argues that there was genuine uncertainty about
the status of epidemiological evidence of the association between smoking and lung
cancer. The results of the Doll and Hill study showed an association between smoking
and lung cancer, and the authors argued that this meant that smoking was the most
likely cause of lung cancer (Doll and Hill, 1950). Although this study is currently seen
as a paradigmatic example of risk factor research, contemporary critics argued that
‘correlation should not be taken as proof of causation’ (Berridge, 2006: 1192, see also
Parascandola, 2004). The nature of public health was changing in this period as it re-
oriented toward chronic disease (Berridge, 2006: 1191). In the UK, this conceptual
change occurred at the same time as an institutional change, as public health moved
out of local authority control to become part of the NHS (Lewis, 1986). Such factors
contributed to what Berridge labels a ‘fluid policy situation’, and although acceptance
of the link between smoking and lung cancer was later seen as a crucial episode in
chronic disease epidemiology (Susser, 1985a: 163), it was not until the 1960s that
British government policy began to change: ‘The establishment of this new
epidemiological risk-focused way of explaining disease was a gradual process of what
can be seen as “scientific claims building”. It did not automatically lead to translation
into policy.’(Berridge, 2003b: 65). One of the reasons for this slowness was that this
approach to public health required new ways of addressing different sections of the
public:

106
 Smoking was a difficult issue in terms of policymaking. It did not “fit” with
what was traditionally considered appropriate as public health intervention.
Much public health concern had been for the containment of epidemics of
infectious, not chronic disease. Health advice about individual behaviour
modification, where it was given, had usually been aimed at women and
children rather than at men, yet the latter formed the majority of smokers in
the 1950s (Berridge, 2003b: 68).

The personal smoking habits of ministers and their scientific advisors, the importance
of tobacco revenues to the Treasury and the long term nature of anti-smoking advice
(which would not produce results for thirty to forty years) also led to political delay
and weak policy responses. In the 1950s, smoking was a widely accepted and normal
activity, and this shaped its regulation. Smoking rates began to decline in the 1970s,
from 51 per cent of men and 41 per cent of women in 1974 to 28 per cent of men and
26 per cent of women in 1998, and this has resulted in the ‘increased marginalization
of smoking and its gradual closer association with poorer groups in society, both men
and women’ (Berridge, 2003b: 63). This marginalisation made intervening in personal
habits more acceptable, and the declining importance of tobacco revenues also meant
that measures that would restrict consumption became more politically possible
(Berridge, 1999b: 1188).

The 1962 Royal College of Physicians (RCP) report, Smoking and Health, marked the
development of a public health policy community around smoking as medical
practitioners and researchers began to adopt a new understanding of their role in
providing the public and politicians with information. A policy community was
developing and lobbying for change. This community consisted of representatives of
chest medicine, cancer and epidemiology and was now based in the RCP, whose 1962
report had provided an agenda for action by this community:

This was a new era in the role of scientific argument and medical lobby. For
the first time, science was reaching out to the public and using the full panoply
of marketing and consumer oriented techniques which were then emergent in a
post-rationing society (Berridge, 2006: 1203 - 4).

Smoking became symbolic of a new role for medicine in explaining health risks and
advising individuals on their management (Berridge, 2006: 1202). The 1962 report

107
was an important episode in the relationship between the medical profession and the
government as it

gave public significance to a new type of public health and to different

scientific ways of studying it. The new epidemiology of the 1950s and the new
focus on risk of chronic disease were translated into a wider public and policy
agenda (Berridge, 2007b: 289).

A focus on preventive health and individual responsibility for avoiding health risks –
an approach that has been labelled ‘new public health’(Petersen and Lupton, 1996) –
was becoming increasingly important in public health policy from the late 1960s
onwards. In the 1950s and early 1960s, British public health was ‘“on the cusp”,
moving away from the mass campaign service-focused public health ethos of the
inter-war years towards a new type of “healthism” epitomized by the concern about
smoking’(Berridge, 2003b: 69). This shift was not fully apparent until the 1970s, but
it originated in the 1950s epidemiological “paradigm-shift” exemplified by
widespread acceptance of the link between smoking and lung cancer (Berridge,
2003b: 73).
Although the anti-smoking coalition had formed primarily around this link between
smoking and lung cancer, it soon came also to focus on the role of smoking as risk
factor for heart disease, which had been demonstrated by results from Framingham
(Dawber, 1960). This link had been known of previously (Aronowitz, 1998: 97-8,
Rothstein, 2003: 239) but the formation of anti-smoking collations in the UK (and the
US) gave it special salience. In this manner, these coalitions contributed to the
consolidation of the narrower or mono-causal approach to risk factors described in
chapter 2.

In 1971, after the publication of the second Royal College of Physicians (RCP) report
Smoking and Health Now, voluntary agreements between government and industry on
regulation of advertising and sponsorship and the use of health warnings were drawn
up. But throughout the 1970s, policy positions on smoking continued to diverge. An
anti-smoking policy coalition developed around Action on Smoking and Health
(ASH) and the Health Education Council, which differed from the harm reduction
approach favoured by the tobacco industry. ASH had been formed in 1971, with
funding from the Department of Health, as an external body that would put pressure

108
on the government to act on the issue (Berridge, 2003b: 73). Initially ASH focused on
policy development and educating smokers, however under a new director in 1973,
the organisation adopted a more high-profile campaigning style that used the media to
create public awareness of the case against smoking. It was ‘a new style of activism in
health, drawing on models developed elsewhere in social campaigning’ (Berridge,
2005a: 106). As Berridge points out, this gave ASH a dual status - it operated
simultaneously as an external pressure group and as an insider member of the policy
network – and it was useful to government ministers who wanted to take action on
smoking (ibid.). Until the 1980s, British government policy on smoking was largely
made in co-operation with the tobacco industry, paralleling the model used in
regulating the pharmaceutical industry (Berridge, 1999b: 1187). This situation has
been described as consisting of a ‘producer network’ of tobacco industry and retail
interests with links to government, and an ‘issue network’ of outside public health
professionals and anti-smoking organisations (Berridge, 2003b, Berridge, 2005a).
However, an analysis of the work of ASH, the Independent Scientific Committee on
Smoking and Health (ISCSH) and the Tobacco Products Research Trust (TPRT)
shows that the networks in the area of smoking policy were more complex, and that
public health scientists, particularly those on expert committees, had links with all
three bodies (Berridge, 1999b: 1187, Berridge, 2005a: 117). In this period, links with
industry were still normal, and not damaging to the credibility of advisors and
policymakers.

However, the anti-smoking coalition was vehemently opposed to the harm reduction
or ‘safer smoking’ approach preferred by industry. From the 1970s onwards, it
campaigned for an abstentionist agenda based on public education, increased taxation
and restrictions on smoking in public places (Berridge, 1999b: 1187-8). Such an
approach was given increased legitimacy in 1981 by the publication of a study
showing that the non-smoking wives of heavy smokers had a higher risk of lung
cancer. There had been early evidence of harm to unborn children which led to anti-
smoking campaigns targeting pregnant women, but this was the first evidence of
smoking as harmful to the general population, rather than just the individual. The
existence of ‘innocent victims’ such as women and children ‘widened the debate and
provided a more powerful engine for driving policy’ (Berridge, 1999b: 1190). Again
the evidence for re-defining the understanding of the risks associated with smoking

109
came from epidemiology, but it also relied on the development of biochemical
markers that could be used to judge levels of exposure to tobacco smoke. Berridge
describes the new science of passive smoking as ‘science waiting to happen’ because
‘the policy objectives of the anti-smoking alliance had already begun to shift in the
direction to which the scientific concept implied policy should go’ (Berridge, 1999b:
1191). This new scientific fact led to a breakdown in the relationship between
orthodox science and the tobacco industry and the entry of new actors such as the
British Medical Association into the anti-smoking coalition.

Passive smoking and the policies it helped initiate essentially combined the
individualism of the 1970s public health paradigm with the environmentalism
of the ‘new public health’, with its resonance with the image of nineteenth
century public health. It was environmental individualism, an alliance of
individual and environmental approaches, just as the science itself married
different technical approaches into a reformulation of the original
epidemiological case (Berridge, 1999b: 1192).

As I have outlined above, these processes ultimately led to a new set of relationships
between the medical profession and government.

The RCP report of 1962 was the forerunner of later College reports on
smoking and a host of other health-related subjects, all of which were aimed at
both the government and the public. The “medical voice” developed important
relationships with both government and the public in areas that would not
previously have been considered the province of either. In the 1970s this
insider/outsider relationship for medicine developed further into a host of
expert committees with close relationships within government (Berridge,
2007b: 310).

These relationships between the medical profession, the government and the public
would also become important in the area of diet and health (see below).

3.3 The early years of the obesity coalition

Smoking was the first risk factor for heart disease around which a policy coalition
formed, and in the next twenty years, researchers and practitioners working on the
relationship between diet, nutrition and chronic disease used the successful creation of
the anti-smoking policy community as a model for the development of a similar
community around diet and heart disease.

110
 Smoking was the major issue which marked the redefinition of public health
around lifestyle issues. The “new public health” policy programme focused on
fiscal (taxation) and media strategies (advertising bans and mass media
campaigns), with a new and distinctive role for “health activist” groups with a
strong anti-industry stance, like ASH. This was a model of public health
activism that was replicated in other areas, for example, diet and heart disease
(Berridge, 2003b: 81).

Obesity and overweight were often listed as subsidiary risk factors for heart disease
(see section 2.3 above) and the public health coalition around them appears to have
developed out of a shared interest in the topic between researchers interested in
diabetes, nutrition and heart disease modelled on the anti-smoking coalition, and
partly overlapping with the diet and heart disease coalition. It started off as a much
smaller off-shoot group, but its framing of increasing body weight as a major public
health problem has been so successful that, by the beginning of the twenty first
century, obesity had become a mainstream political issue (see chapter 10).

The late 1960s mark the start of this important period for British biomedical research
into body weight and health. A burst of activity around the topic occurred between the
late 1960s to the late 1970s and gave rise to a series of symposiums and conferences,
a couple of reports and textbooks. The newly formed Obesity Association held its first
symposium in October 1968 (Butterfield, 1969: 1). A 1969 textbook on the subject of
obesity was described as the first ‘comprehensive’ treatment of the topic for more
than a decade, whereas by the third edition the author was stating that ‘more than a
dozen books’ had been published on the subject between 1973 and 1977 (Craddock,
1978: vii). The First International Conference on Obesity was held in 1975, and in the
late 1970s another author referred to ‘several major international conferences’ on the
topic in the previous four years and a ‘recent explosion of literature in the
field’(Garrow, 1978: 1). However, the topic of obesity was framed in a seemingly
contradictory fashion: although the condition was widely accepted to be an important
health problem, the consensus was that very little was known about it and more
research was required:

whilst the indictment against obesity appears to grow in strength each year
enormous gaps in our knowledge of the subject remain. There is great
uncertainty regarding both an appropriate definition and possible methods of

111
 measuring obesity and consequent lack of information on its prevalence in the
United Kingdom. This paucity of fundamental data makes it extremely
difficult to assess the precise role of obesity in disease patterns (OHE, 1969:
3).

This juxtaposition of great uncertainty about the definition, incidence and health
consequences of obesity, with a clear sense that it is an important health problem,
both for individuals and society, was typical of obesity science writing in this period.
Despite its apparent contradictoriness, it allowed excess body weight to be presented
as an important issue in health policy, whilst at the same time calling for more
research funding in this area.

Participants in the early period of obesity science mainly came from two medical
specialities - diabetes and human nutrition. Specialists from other areas, such as heart
disease, also participated but they were heavily outnumbered by these two other
groups (see table 3.1). Those who attended the first symposium organised by the
Obesity Association in 1968 were largely clinicians and researchers from London
universities and teaching hospitals. An interest in the topic brought together those
interested in research into diabetes and its treatment, those studying human nutrition
and pharmacologists investigating treatments for obesity. For many of these
participants, obesity was interesting primarily as it related to other conditions rather
than as a condition in its own right. The significant number of diabetes researchers in
this group shows that it was another important area for the development of knowledge
about obesity, since the link between overweight and type 2 diabetes had been known
since the mid nineteenth century (Schwartz, 1986: 100) and many of these individuals
were also present at a meeting of the British Diabetic Association in 1970 which
included six presentations on the subject of obesity (BDA, 1970). The first President
of the Obesity Association and chair of the symposium was Professor John
Butterfield, whose central research interest was diabetes, including basic
physiological research, epidemiological surveys and diagnostic criteria. He was later a
member of the Committee on Medical Aspects of Food Policy and the World Health
Organization's expert committee on diabetes. As an acknowledged expert in his own
field, his reputation would have added to the legitimacy of this new professional
body.

112
The nutrition researchers present at the OA symposium, such as Professor John
Yudkin and Dr I. Howard Baird, began their careers researching deficiency diseases
and the effects of malnutrition. In the immediate post-war period human nutrition
research focused on protein malnutrition in developing countries. Because nutrition
research in developed countries focused so strongly on animal nutrition, researchers
interested in human nutrition had been forced to switch their interests to malnutrition
in developing countries (James, 2006: 83). The identification of one form of
malnutrition, kwashiorkor, in the mid 1930s, along with the food shortages associated
with World War II, had also led to nutrition researchers shifting into the area of
deficiency diseases in order to identify their causes and, therefore, preventive
measures. According to another researcher in the area, protein malnutrition – ‘the
protein gap’ - was seen as a major public health problem until the late 1960s (Payne,
2010).28

The second report Research into Obesity (henceforth the 1977 DHSS/MRC report)
was produced by group from the Department of Health and Social Security and the
Medical Research Council chaired by Professor John Waterlow of the London School
of Hygiene and Tropical Medicine (LSHTM) who also worked in the area of

Dr Ian Mclean Baird Prof John Butterfield Sir Douglas Black (chair)
West Middlesex Hospital, University of Cambridge, Royal College of Physicians
(human nutrition) (diabetes) (health inequalities)
Prof John Butterfield Dr J Durnin Dr Gordon M Besser
(chair) University of Glasgow St Bart’s Hospital, London
University of Cambridge (physiology) (diabetes)
(diabetes)
Dr Alan Howard Dr John Garrow Dr Charles Brook
University of Cambridge MRC Clinical Research Middlesex Hospital, London
(human nutrition) Centre, Harrow (paediatric endocrinology)
(human nutrition)
Professor Alan Keswick Professor Nick Hales Dr Denis Craddock
West Middlesex Hospital, The Welsh National School of GP/textbook author,
(human nutrition) Medicine (obesity)
(diabetes)

28
This author also referred to the ‘great protein fiasco’ when researchers based at LSTHM participated
in a FAO/WHO consultation in the mid 1970s and successfully argued for a reduction in the accepted
levels of protein requirements (Payne, 2010).

113
Dr June Lloyd Dorothy Hollingsworth Dr John Garrow
University of London/Great British Nutrition Foundation MRC Clinical Research Centre,
Ormond St Hospital (nutrition policy) Harrow
(paediatric endocrinology) (human nutrition)
Dr Derek Miller Dr WPT (Philip) James Dr Derek Hockaday
University of London MRC Dunn Nutrition Unit Radcliffe Infirmary,
(food history) Cambridge Oxford
(human nutrition) (diabetes)
Dr Galwan Pawan Dr June Lloyd Dr WPT (Philip) James
West Middlesex Hospital, University of London/Great MRC Dunn Nutrition Unit
London Ormond St Hospital Cambridge
(human nutrition) (paediatric endocrinology) (human nutrition)
Dr Trevor Silverstone Prof John Waterlow Dr Barry Lewis
St Batholomew’s Hospital, (chair) St Thomas’s Medical School,
(pharmacology) London School of Hygiene London
and Tropical Medicine (heart disease)
(human nutrition)
Dr Steve Szanto Dr J L Mann
University of London University of Oxford
(human nutrition) (diabetes)
Dr Paul Turner Dr Derek Miller
St Batholomew’s Hospital, University of London
London (food history)
(pharmacology)
Professor John Yudkin Dr TRE Pilkington
University of London St George’s Hospital Medical
(human nutrition) School, London
(diabetes)
Dr Trevor Silverstone
St Batholomew’s Hospital,
London
(pharmacology)

Table 3.1 Committee members for three early obesity reports showing their
institutional affiliations and research areas.

malnutrition.29 His research focused on protein turnover, nitrogen balance and

childhood malnutrition as a public health problem. In 1956 he had established the
Tropical Metabolism Research Unit (TMRU) at the University of the West Indies in
Jamaica to study malnourished children, identifying that they were suffering from
kwashiorkor which resulted from infective conditions such as diarrhoea and
developing a treatment programme (Millward, 2010). Dr (later Professor) W.P.T.
(Philip) James who compiled the DHSS/MRC report, and was at that time based at the
MRC Dunn Nutrition Unit in Cambridge, had also worked on malnutrition at the
TMRU in the early 1960s, but by the 1970s had become interested in coronary heart

29
Information about the careers of individuals provided in this section comes from the bibliography of
obesity (Whelan and Silverstone, 1974), PubMed/Medline and The Lancet.

114
disease and obesity. Dr John Garrow was another nutrition researcher who worked at
the TMRU in the 1950s and moved into the study of overweight and obesity in the
1960s. By the late 1960s he was working at the MRC Clinical Research Centre,
Harrow and he ultimately became a Professor of Human Nutrition at the University of
London, as well as consulting at several London teaching hospitals. He wrote on a
wide range of topics around human nutrition and dietetics, including measuring
obesity/body composition/adiposity, energy expenditure and the effects of diet and
exercise, including text books such as Energy balance and obesity in man (Garrow
1978), Treat Obesity Seriously: A Clinical Manual (Garrow, 1981), Obesity and
related diseases (Garrow, 1988) and Human nutrition and dietetics (Garrow and
James, 1993). In the UK he became a prominent advocate of both the energy balance
model and the BMI as a means of measuring and classifying overweight and obesity
(see chapters 4 to 6). Another notable participant was Dr J. Durnin from the
Department of Physiology at Glasgow University whose research areas were physical
activity, energy expenditure, dietary intake and body fat measurements. In 1971
Durnin had been part of an FAO/WHO ad hoc committee of experts on energy and
protein, and in 1987 he took part in a consultation on chronic energy deficiency with
Philip James and John Waterlow. From the outset, therefore, members of the
emerging obesity science coalition already had a well developed professional network
as well as links with UN bodies such as the FAO and the WHO. Despite the shift in
focus towards over-nutrition, there was continuing research and policy interest in
under-nutrition.

As I have outlined, by the late 1960s there appears to have been a major shift in
British nutrition science as researchers shifted their focus towards problems
associated with over-nutrition, such as heart disease and overweight. One explanation
for this shift may be that whilst it was acceptable to study malnutrition in poorer
countries, the growing prosperity of the post-war economic boom made it
unfashionable to study malnutrition in rich industrialised countries. From the early
days of their discipline, nutrition researchers have become involved in the framing of
policy and debates on issues such as wage levels, rationing requirements and the
relationship between poverty and malnutrition (Aronson, 1982, Mayhew, 1988).
Nutrition researchers were often highly entrepreneurial scientists, who sought to use
their expertise in outside arenas. It is possible that as malnutrition became a less

115
fashionable research topic, researchers based in developed countries saw obesity as an
area, with potential policy relevance, which they could claim as their own.

The DHSS/MRC report was a step up the institutional ladder for the obesity coalition.
Their early activities consisted of establishing professional organisations, organising
meetings and publishing collections of research articles and textbooks. However, the
1977 report was the first time they had collaborated with a government department,
and this marks an important step in the acceptance of their arguments that obesity was
a significant public health problem. The increasing profile of obesity research was
further enhanced when, in 1983, a working party of the Royal College of Physicians
chaired by Sir Douglas Black produced the report, Obesity. Lord Black, who was
president of the organisation at the time, had produced the famous 1979 report into
health inequalities, which contained substantial discussions on heart disease and the
role of lifestyle factors in its high rates. Controversy around the newly elected
Conservative government’s attempt to suppress this report had made Sir Douglas
Black one of the most high profile medical experts of the period and his involvement,
as president of the RCP, must have enhanced the visibility of obesity science. But this
recognition appears to have been short-lived as there were no more reports on obesity
until the mid 1990s.

The honorary secretary of the RCP report was Philip James, and another member was
Professor Barry Lewis, a high profile researcher on coronary heart disease, who went
on to be chair of the International Task Force for Prevention of Coronary Heart
Disease at its foundation in 1987. At this time, Philip James was also involved in the
Coronary Prevention Group (CPG). The CPG aimed to operate in the style of a
campaigning pressure group, such as ASH, and had been involved in the production
of a well publicised report on policies for the prevention of heart disease, known as
the Canterbury Report (Rose et al., 1984). Thus there was still overlap between
obesity science and the wider field of coronary heart disease research. However, a
coalition of researchers interested in obesity, hitherto regarded as one of the lesser risk
factors for heart disease, was beginning to establish the field of obesity science as
separate from mainstream heart disease research through strategic alliances with other
groups of researchers and with bodies such as the MRC and the RCP.

116
Policy making in the area of diet and heart disease showed many parallels with that in
the area of smoking, including the problems of dealing with scientific uncertainty and
the deployment of this new medical voice, particularly in the form of advice from
expert committees. The gradual acceptance of the link between high fat diets and
increased rates of heart disease by the Committee on Medical and Nutritional Aspects
of Food (COMA) demonstrates these processes. The committee’s panel of diet and
heart disease moved from complete lack of agreement consensus in 1974 (resulting in
no report) to consensus in the 1984 and 1994 reports which advised individuals to
reduce their average total fat consumption by 10 per cent, as well as reducing their
consumption of sugar, salt and alcohol (Bufton and Berridge, 2000).

3.4 The coalition’s initial model of obesity: the individual paradigm

In this section I will outline the main features of what I label the individual paradigm
of obesity. This paradigm was an emerging set of assumptions and claims about
obesity that circulated, largely unchallenged, within the obesity coalition. It developed
alongside the institutions of early obesity science: the formation of the obesity
coalition and the articulation of the individual paradigm were effectively two halves
of the same coin.

This model of the condition largely defined it and measured it using the tables of ideal
weight produced by the Metropolitan Life Insurance Company (see chapter 2). It
relied on small-scale surveys to estimate the condition’s prevalence, and saw excess
body weight as largely caused by overeating, best treated by weight loss diets and
associated with heart disease, diabetes and osteoarthritis.

3.4.1 Defining and measuring obesity

In this early period of obesity science, defining obesity objectively was often seen as
difficult: ‘Many clinicians have argued that the definition of obesity must remain
subjective’ (OHE, 1969: 4). This led to problems for research: ‘the great paucity of
studies to establish the incidence of obesity probably almost as much reflect the
problem of definition as the early lack of concern with the condition’ (OHE, 1969: 7).

117
This persistent problem of definition was related to understandings of the causes of
obesity, as it was not yet understood as a unitary entity:

Obesity is neither one condition nor one disease. The fat baby, the fat
adolescent girl, the woman who gets fatter after each pregnancy, the
traditional example of the fat business executive – these all have fatness in
common, but it is very doubtful if the aetiology and natural history are the
same in all of them (DHSS/MRC, 1976: 1).

However, despite this uncertainty, in order to frame it as a public health issue, it was
strategic to provide an appropriate numerical measure so that the incidence of obesity
could be measured. Throughout this period, statistical methods were being
increasingly used within medicine (Rothstein, 2003) and numerical indices were
becoming important in the framing of convincing arguments for the existence of a
health problem. However, competing definitions and measures of obesity, some of
which were difficult to convey concisely, made this difficult for early obesity science
writers.

Often obesity was defined using ideal weight tables produced by the American
insurance industry. These were based on US data from 26 large companies collected
over 20 years from nearly 5 million people (OHE, 1969 : 4 - 5). In particular, the
Metropolitan Life Company used relative weight30 to produce a range of categories:
average weight, overweight (greater than 110% of average weight), and obese (greater
than 120% of average weight). These weight ranges were then incorporated into
charts including height and frame sizes and correlated with mortality data to give
ranges of “ideal weights” – for a 5 foot 6 inch woman the desirable weight range was
124 to 156 pounds, depending on whether they were small, medium or large framed
(ibid.). However, the link between ‘above average weight’ and being ‘too heavy’ was
seen as unclear and methods such as under-water weighing and measurements of skin
fold thickness had to be used to determine total body fat (OHE, 1969: 4). To counter
accusation of the narrowness of these ranges, another author argued that 10% excess
weight ‘allows a 76kg (168lb) man to accumulate 7.5kg (16lb) of fat without being
obese!’ (Craddock, 1978: 2). Worrying about the stringency of recommended weight
ranges was not a feature that recurred in these sources.

30
This is weight relative to height.

118
In addition to the widely used Metropolitan Life tables, there were several other
definitions of obesity used in these texts. One author defined obesity as weighing
10% above ‘normal’ or ‘desirable’ weight31, and ‘excessive’ obesity as weighing 20%
above the desirable weight, but mentioned others using 30% as an appropriate figure
(Craddock, 1969: 2 - 3). Another referred to Seltzer’s ‘ponderal index’32 which was
proposed in 1966 but does not seem to have become widely used, but also used tables
from Framingham that framed risks of heart disease in terms of relative weight (Baird,
1969: 17-19). Nearly ten years after the OHE report, a later textbook was still using
an approach derived from insurance company tables:

For practical clinical purposes it is convenient to take the range of “desirable

weight” from life insurance experience, from the lower end of the small frame
to the upper end of the large frame (since frame size is undefined), and accept
that people above this weight are obese (Garrow, 1978: 149).

Garrow also referred to body mass index (BMI)33 as an alternative formulation of an

acceptable weight range, implying that it was known about, but had not yet become
the standard index.

The 1977 DHSS/MRC report is the first to make a distinction between two different
ways of measuring obesity – relative weight and numerical indexes, such as W/H².
The latter express the relationship between weight and height and are more absolute
measures because they do not refer to population averages. A fundamental problem
with relative weight measures using insurance company data was that the selection of
reference standard was arbitrary, since it was based on a particular and
unrepresentative sub-group of the population (DHSS/MRC, 1976: 3 - 4). The authors
did discuss W/H² in this report, but in the context of the Metropolitan Life Company
ideal weights: ‘A cut-off point often used for separating obese from non-obese is a
relative weight 120% of the “desirable” weight. This corresponds to values for W/H²
of 27.5 for men and 27.0 for women of medium frame size and 29.9 for men and 29.5
for women of large frame size’ (DHSS/MRC, 1976: 4). The main definition of obesity

31
However, he admitted that others used a figure of 15% or 20% above desirable weight.
32
This is defined as the ratio of the height in inches divided by the cube root of the weight in pounds –
one of a number of precursors to the BMI (see chapter 4).
33
Body mass index equals weight divided by height squared or W/H².

119
was based on percentage over desirable weights from the Metropolitan Life Insurance
Company tables. Other methods of measurement were being discussed, but there were
several competing alternative measures rather than one agreed alternative to desirable
weights.

3.4.2 Estimating the prevalence of obesity in the population

At this time there were no large scale studies to give an accurate picture of the
prevalence of obesity within the British population, and such health statistics were not
yet collected by government departments. Researchers seeking to frame obesity as a
public health problem were, therefore, forced to refer to the American insurance
company data mentioned above, and extrapolate from the small number of British
surveys that had been conducted. These authors strategically used the available
information to argue that a problem existed, while at the same time pointing out the
inadequacy of this evidence, and thus stressing the need for further research.

To illustrate the incidence of obesity, one author referred to Metropolitan Life data
demonstrating that one in five men and one in four women aged 20 or over was at
least 10% over the average weight, and one in twenty and one in nine at least 20%
over (OHE, 1969: 7). American data were cited because ‘No comparable study has
been undertaken in this country, and we are therefore left with less precise data on
which to base an assessment of the prevalence of obesity’ (ibid.). However, the results
of various smaller studies ‘would seem to indicate a picture similar to that found in
America with about one in five adults considered to be clinically obese’ (OHE, 1969:
10).

A 1969 textbook referred to only three small scale UK studies of the weights of
general practice patients (Craddock, 1969: 3 - 5), but a later edition, although
mentioning a lack of national data, was also able to cite a 1968 study of 11000 BP
workers, and two studies of London patients. Using this new data, the author
concluded that ‘it is likely that up to one half of the women over 30 in Great Britain
are at least 10 per cent overweight and a similar proportion of men’ (Craddock, 1978:
4). An earlier author states that the incidence of obesity in Britain was not accurately
known but, after citing the same studies, went on to argue that ‘at least half the

120
middle-aged persons in this country are overweight’ (Baird, 1969: 15-16). The
DHSS/MRC report referred to the same studies as the other sources, but came to a
different conclusion, arguing that ‘at all ages and in both sexes, there has been a small
increase [in the prevalence of obesity] since the war time surveys of 1943’
(DHSS/MRC, 1976: 9). This conclusion was strangely at odds with the sense of
urgency conveyed by other sections of the report (see section 3.6 below).

3.4.3 The health consequences of obesity

In the 1960s the relationship between obesity and mortality was still contentious, and
the case had to be made for a causal relationship (OHE, 1969: 15). The results of the
1959 Build and Blood Pressure Study34 showed that mortality amongst men 10%
above average weight was 20% higher than that of those of average weight. However,
it was possible that ‘obesity does not lead to the complications which cause death but
is simply a consequence following from the significant disease. If this is so, reduction
in weight would not influence survival’ (OHE, 1969: 17). This author also cited
evidence from the same study to show that weight reduction lowered mortality
suggesting that there is a causal relationship. Such discussions further demonstrate the
heavy reliance on statistics from life insurance companies, despite the acknowledged
problems with their generalisability, since the insured population was generally more
affluent and in better health than the general population (see also Baird, 1969: 16,
DHSS/MRC, 1976: 20).

Until the 1980s, the list of negative health consequences given in these sources was
relatively short and based on understandings of the relationship between obesity, heart
disease (which was still controversial), diabetes and mechanical problems such as
joint pain and breathing difficulties. One such list split the consequences into the
mechanical - effects on joints, obstruction of breathing, hernia and varicose veins -
and the metabolic - diabetes, raised cholesterol, gallstones and atheroma35 and CHD
(OHE, 1969: 15 - 17). Another author listed the health consequences as diabetes,
coronary artery disease, osteoarthritis, hypertension (although this is still not definite)

34
This study pooled data on 4.9 million policy holders from 26 life insurance companies in Canada and
the US, and one author argues that Metropolitan Life used it to produce a new table of reduced
recommended weight ranges in 1960 (Glaesser, 2002: 46).
35
Atheroma is a disease resulting from fatty deposits in the linings of the arteries.

121
and varicose veins (Butterfield, 1969: 4-5). Later on the same author used early
Framingham data to argue that obesity was also associated with an increased risk of
heart disease (although the relationship was not clear and several possible
mechanisms were proposed), angina, impaired pulmonary function and diabetes
(Baird, 1969: 17-21).

In the 1970s, heart disease was becoming seen as an increasingly important health
consequence of obesity. However, the attribution of a causal role for obesity was
described as ‘controversial’ (DHSS/MRC, 1976: 22), due to contradictory data from
two of the large scale epidemiological studies: results from Framingham showed that
reduced weight led to a reduced risk of heart disease, whereas those from the Seven
Countries study appeared to show no effect.36 Other negative consequences
highlighted at this time included hypertension, diabetes, damage to weight bearing
joints, obstetric and post-operative complications and reduced physical fitness
(DHSS/MRC, 1976: 22 - 25). By the late 1970s, authors quoted Framingham
researchers arguing that ‘If everyone were at optimum weight we would have 25 per
cent less coronary heart disease and 35 per cent less congestive heart failure and brain
infarction’ (quoted in Craddock, 1978: 6). Moreover, Craddock’s 1978 discussion of
the health consequence of obesity was more detailed than in 1969 when it relied
heavily on 1923/4 mortality data from Metropolitan Life, and he divided the
‘complications, hazards and disadvantages of obesity’ into three categories –
increased mortality, increased morbidity, and physical discomforts and disadvantages.
The increased mortality derived from increased rates of cardiovascular, respiratory
insufficiency, diabetes, gall stones and liver damage (Craddock, 1978: 6 - 8). The
evidence for these increased risks came from Framingham, Metropolitan Life
Company and three other named studies. Examples of morbidity cited included
hernias, colon cancer, diverticular disease, arterial disease, haemorrhoids, varicose
veins, thrombophlebitis37 and complications of pregnancy (Craddock, 1978: 8).

36
The authors of the DHSS/MRC report argued that these apparent contradictory results were due to
the difference in the two study populations - the Framingham study population was aged from 30 to 59
whereas Seven Countries one was aged 40 to 59. As the effects of obesity on mortality were more
marked among the young, it would be expected that by not studying this age group the Seven Countries
Study would be less likely to show a correlation between obesity and heart disease mortality
(DHSS/MRC, 1976: 23).
37
Diverticular disease is another name for diverticulitis or inflammation of diverticula in the intestinal
tract and thromboplebitis is inflammation of a vein associated with the formation of a blood clot.

122
In his textbook of the same year, Garrow also used data from early large scale
epidemiological studies. He lists increased risk of coronary heart disease as the main
mortality risk and diabetes, hypertension, gall bladder disease, plus endometrial
cancer, osteoarthritis of the knees, risks to mother and child in pregnancy and poor
lung function as associated morbidities using data from American insurance
companies, a Finnish study of 5000 insured men, the Seven Countries study and a
study of 73, 522 members of an American slimming club known as TOPS38 (Garrow,
1978: 146 - 152). As prominent researchers in the field, Garrow and the committee
producing the DHSS/MRC report, would be expected to include recently published
data from prestigious research, such as the large scale American epidemiological
studies. These novel sources of information were beginning to be cited, but they
would not lead to an expanded understanding of the health consequences of obesity
until later reports in the 1990s (see chapter 7).

3.4.4 The causes of obesity

These authors largely framed the causes of obesity in individual terms. There was a
partial acknowledgement of the social factors involved in eating behaviour and body
size: occasionally, these discussions were even framed in terms of social class or
gender (see below). However, the most commonly cited and best described causes
were psychological or behavioural ones. These included a tendency for the obese to
eat in response to external cues such as time of day or anxiety, rather than internal
cues of hunger; their greater preference for sweet foods, and the relationship of
obesity to conditions such as neuroticism and depression. As well as containing an
appealing element of common sense, these explanations also fitted well with
individualised, biomedical accounts of overweight and obesity. They drew on
psychiatry, a discipline with a long standing interest in eating disorder that had
become part of orthodox medical practice and research by the early twentieth century,

38
TOPS stands for Take Pounds Off Sensibly. This study involved analysing questionnaires asking for
information about diagnoses of particular conditions such as diabetes, hypertension, heart disease and
anaemia and relating this information to degree of overweight. The most overweight women were
found to have significantly higher rates of diabetes, high blood pressure and gall bladder disease
(Garrow, 1978: 150 -1).

123
unlike social sciences such as sociology or anthropology which have also analysed
eating practices in considerable depth (see chapter 6).39

In discussions of the causes of obesity, there were further attempts to sort out whether
there was more than one kind of obesity. The majority of the overweight were thought
to suffer from ‘simple obesity’ caused by eating too much and insufficient physical
activity, rather than obesity resulting from psychiatric ‘abnormalities’, genetic factors
or damage to the hypothalamus (demonstrated in rats to lead to obesity) (OHE, 1969:
11). Although Silverstone saw obesity as a normal response to ageing:

Most people tend to put on weight as they grow older (due to a reduction of
energy expenditure without a concomitant reduction in calorie intake), hence
the increase in the prevalence of obesity with age, but against this there are
considerable social pressures acting on individuals in the upper social classes
to take remedial action – that is to diet – whereas such social pressures are
probably much weaker among lower social classes, hence the increased
prevalence of obesity in this group (Silverstone, 1969: 47).

He labelled this ‘maturity-onset obesity’ and separated it from ‘early-onset, neurotic

obesity’ for which the causes included eating to manage anxiety, disturbed body
image and disassociation between eating and physiological hunger (Silverstone, 1969:
48 - 51).40 This categorisation of obesity as an eating disorder is illustrated by the way
textbooks at the time (e.g. Bruch, 1974) routinely dealt with both topics.41 In his
summary of the psychological aspects of obesity Craddock argued that,

The majority of obese people probably have no obvious psychological factors

affecting their tendency to obesity. Many of them are people who eat under the
minor stresses of everyday life instead of, or as well as smoking, drinking
alcohol or tea or biting their nails. They are often hypersensitive individuals
whose reaction to life is passive rather than active, they have a genetic

39
The DHSS/MRC report was unusual in that it stated that information on the ‘social factors’ involved
in the development of obesity was ‘limited’ and argued for more collaborative research into eating
patterns involving nutritionists and social scientists (DHSS/MRC, 1976: 65). In a more standard mode,
Butterfield discusses the use of ‘psychosocial workers’ to help with communication with patients and
work with teenagers and the middle aged to prevent the development of obesity (Butterfield, 1969: 8-
9).
40
He also has a third category of reactive obesity – those who become obese in later life due to severe
stress (Silverstone, 1969: 54) – which was understanding obesity as an eating disorder with an
identifiable psychological or psychiatric cause.
41
The boundary between obesity and eating disorders is fluid and has shifted in the last thirty years,
partly due to different groups of clinical practitioners and biomedical scientists claiming the subject as
their own.

124
 tendency to put on weight easily and they have been conditioned to eating as
reward since early childhood. They therefore eat when under stress, whereas
most people are less hungry and lose weight when under stress owing to
increased adrenalin production which mobilises fat and glucose. Many of them
do not bother to consult their doctors about the problem of overweight for a
variety of reasons. Many of them are not distressed physically by their
overweight and are not worried enough about their looks to consider the
discipline of serious dieting. Many of them appear to overeat instead of getting
depressed or neurotic and this is often a reasonable compromise for them.
Those who express a neurosis by over-eating instead of in other ways usually
have difficulty in losing weight (Craddock, 1969: 100).

This excerpt may reflect his background in general practice, where he would see
fewer seriously overweight patients, rather than hospital medicine. The matter-of-fact
description of the causes of obesity, and many individuals’ acceptance of excess body
weight was not found in other writing of the period, and is not found in contemporary
writing on the topic. But his account of the psychological aspects of obesity, focusing
on patterns of compensatory and compulsive eating, was very standard.

Anticipating later concerns, one author argued that: ‘The problem of inadequate
exercise is probably a growing one. The development of a prosperous industrial
society means that many of the essential physical activities of life have been
removed’(OHE, 1969: 13). Another made the point in more detail:

the increasing prevalence of obesity in Western Societies can be attributed to

under-exercise and over-eating. It is emerging that the former is really more
important than the latter, and that, as a race, we have inherited appetites
commensurate with unmechanised agricultural communities. We have not yet
adapted to the luxury of steady supplies of calories in our markets all year
round. Nor have we mastered the inevitable laziness, in terms of muscular
effort, which follows the exploitation of the world’s energy supplies
(Butterfield, 1969: 3).

This thread of nostalgia for an idealised healthier past remained important in obesity
science writing (see chapter 8). However, Garrow criticised arguments that obesity is
caused by physical inactivity since no studies have been able to find significant
differences between the levels of physical activity of obese and lean subjects (Garrow,
1978: 105). There is a tension in this writing between assertions that can be backed up
by research evidence, and those, like the statements above, which are more general

125
and tap into fears about the effects of modernisation and industrialisation (see
chapters 8 and 9).

A common approach of these authors was to discuss the causes of obesity in terms of
genetic, environmental factors and finally social factors: ‘Obesity is a product of
plenty, but social factors play a big part in determining the relative influence in
differing groupings’ (Craddock, 1969: 22). Under the heading of social factors topics
as varied as social class, national habits, eating habits, smoking and pressures
affecting women were noted in different combinations by different authors. The
‘social factors in obesity’ could also include food consumption, eating patterns,
physical activity (obese people show reduced physical activity, but it cannot be
established whether this is cause or effect), social class, individual attitudes and
psychiatric and behavioural aspects (DHSS/MRC, 1976: 13 - 18). The causes of
obesity were coming to be labelled as complex and multi-factorial, and this rapidly
became conventional wisdom in obesity science

It seems probable that in the great majority of cases the cause of obesity is
multifactorial: its onset is determined by a combination of genetic,
psychological, metabolic and endocrine factors, and we are unable to
disentangle the relative importance of each component. These are the cases
sometimes referred to as “simple” or “idiopathic” obesity because there is no
obvious cause (DHSS/MRC, 1976: 2).

Describing obesity as a complex and multi-factorial condition put it in the same

category as many chronic diseases, notably heart disease, treating it as condition in its
own right, rather than merely a risk factor for other conditions. It also implied that
more research would be required in order to fully understand these causes, and that it
might be a difficult condition to treat. This in turn gave status to the professionals
whose research sought to understand these causes and whose clinical practice
attempted to treat the obese.

3.4.5 Treatments for obesity

Between the 1950s and the 1980s, the standard treatments for obesity were restricted
diet (of various types although the basic elements are fairly stable), exercise (with
reservations about its effectiveness), behavioural therapy and drugs. Surgery was a

126
last resort treatment, reserved for severely overweight patients who have failed to lose
weight by other means. These treatments were largely based on existing, clinical
knowledge about the efficacy of such techniques:

There are virtually no published trials of different therapeutic regimes in

properly matched groups of obese patients studied over a period of long
enough for at least some of the patients to achieve normal weight. The
proportion of people in the community who, by their own endeavours, reduce
weight to normal and then maintain weight by conscious dieting remains
unknown. In hospital practice, however, it is evident that all forms of medical
treatment for gross obesity have a low “cure” rate; long term follow up shows
that many of those who have achieved normal weight by slimming relapse
(DHSS/MRC, 1976: 57).

Not only had many of these techniques not been formally trialled, but even the best
recognised treatment, weight-loss diets, was accepted to be ineffective in many
cases.42 This was not a new or uncommon understanding: ‘Despite the availability of
several useful dietary aids, the long-term treatment of obesity can best be achieved by
a re-education of the patient in his dietary habits’ (Howard, 1969: 109) – that is, not
by short-term restrictive diets. Despite this knowledge a reduced calorie diet remained
the main treatment recommended in all of these texts: the ‘obvious’ and common
sense cure for obesity was to reduce the amount of calories consumed, and a common
recommendation is for a diet of 1000 calories a day (OHE, 1969: 21). Restriction of
food intake to reduce weight was seen to be the method favoured by doctors and the
general public. Presumably, as a GP, Craddock would have been familiar with dietary
treatments, and he outlined a range of options, including low carbohydrate, high fat or
high protein and calorie controlled versions, various type of meal replacements and
substitute foods, and he gave detailed diet sheets for nine different diets that included
daily meal plans detailing appropriate portion sizes for particular food items
(Craddock, 1969). In the short term, weight loss was seen as a matter of the
overweight individual finding sufficient motivation and a palatable diet. Different
macronutrient compositions were possible and Yudkin’s high protein, low fat and low
carbohydrate diet fitted the criteria that a diet should be ‘calorically reduced,
nutritionally adequate, socially acceptable, economically feasible, and potentially

42
This was not an isolated finding (for a discussion of the topic see Ernsberger and Haskew, 1987 and
Campos 2004).

127
permanent’ (Yudkin, 1969: 92). Yudkin argued that a low fat diet was unpalatable and
not satisfying to most people, and so his low-carbohydrate diet was ‘a most effective
way of losing excessive weight and losing it permanently’ (Yudkin, 1969: 94). Other
options discussed included starvation therapy and high protein low calorie diets
(Howard, 1969: 96 - 105).

At first the role of exercise was seen as problematic since increased activity can
increase appetite and caloric requirements:

There is an obvious relationship between reduction of caloric intake and an

increase in calorie expenditure resulting from increased activity. Unfortunately
whilst both activity and resultant heat loss owing to muscular inefficiency
dissipate body energy, they also lead to increase in both caloric requirements
and appetite (OHE, 1969: 22).

Craddock, on the other hand, argued that the role of exercise had been underestimated
‘because of theories as to the amount of exercise needed to “work off” a certain
weight of fat, which have not taken into account the increased metabolic rate
produced by exercise’ (Craddock, 1969: 92). Therefore, he recommended that
exercise should be part of every treatment programme unless medically contra-
indicated. Yudkin also stated that an extra 30 minutes walking a day (with no increase
in food intake) would lead to weight loss of about a pound a month (Yudkin, 1969:
92). He also rejected the argument that increased levels of physical activity led to
increased appetite and therefore food intake. For these reasons he favoured his
patients undertaking regular frequent ‘mild activity’. However he admitted that most
patients rejected this advice and bemoans that ‘Much in our affluent lives is directed
to reducing our activity rather than to increasing it – not only motor cars and
television and washing machines, but plastic surfaces, non-polish floors, cars with
automatic gears and even electric toothbrushes’(ibid.).

The author of the OHE report admitted that ‘most efforts to reduce weight, whether or
not taken under the supervision of a doctor are unsuccessful’ (OHE, 1969: 24). Such
low levels of success often led to the prescription (and abuse) of drugs, such as
amphetamines, about which the BMA was becoming more cautious at this time
(OHE, 1969: 25). Drug treatments at that time fell into two categories – bulking

128
agents (which were not recommended) and appetite suppressant drugs such as
amphetamines, fenfluaramine and diguanides which were used to treat diabetes, but
caused loss of appetite in both diabetics and non-diabetics (Turner, 1969: 112 - 4).
The accepted approach was that drugs should only be prescribed when dieting had
been first tried and that they should be used to improve adherence to a dietary regime,
rather than as a weight loss method on their own. Another author divided the drugs
used for weight reduction into five categories – anorectics, oral hypoglycaemic
agents, metabolic stimulators, laxative and diuretics – and of these recommended only
the anorectic drugs (Craddock, 1969: 70).43

Behavioural therapy was occasionally discussed as another way of ensuring that

patients adhered to restrictive diets (DHSS/MRC, 1976: 60, Garrow, 1978: 172 - 3).
Garrow hoped that ‘a means of permanently rendering obese patients more amenable
to dietary restriction will be found, but no sure method of achieving this objective has
been published so far’ (Garrow, 1978: 173). The best treatment results were thought
to occur when the individual followed a low calorie diet, and was seen and weighed
regularly by the same person (DHSS/MRC, 1976: 57 - 60). Simple surveillance was
the most effective method of ensuring patient compliance.

Surgery was very much seen as a last resort. Intractable obesity that was ‘resistant to
normal methods of treatment’ (Craddock, 1969: 133) should be treated by fasting
(short and long term), intensive out-patient support and a low calorie diet, surgery
(reduction of specific areas or intestinal by-pass), group therapy or hypnotherapy.
In the mid 1970s surgery was discussed under the heading of ‘extreme measures’
which included in-patient treatment by starvation and surgical therapies such as
jejunum-ileal by-passes or dental splints44 (DHSS/MRC, 1976: 61 - 62).

Thus, there was a growing body of clinical knowledge about appropriate treatments
for overweight and obese patients that included weight loss diets and drugs as well as
additional treatments such as monitored fasts and surgery for extreme cases. There
43
The 1969 edition contains a great deal of detail about the various drugs available – including the
chemical structure, mode of action and possible side effects. This is almost entirely omitted from the
1978 edition, possibly due to change fashions in obesity treatment.
44
A jejunum-ileal by-pass was a surgical procedure in which all but 30cm to 45cm of the small bowel
were detached in order to significantly decrease food absorption, and dental splints is another name for
jaw wiring which means that the patient can only consume liquids.

129
were debates about the role of exercise and the most effective and palatable diet, but,
in spite of the low rates of success, there was already a large measure of agreement
about the ‘rational treatment of obesity’ (Craddock, 1969: vii).

3.4.6 Costs to healthcare systems

There was virtually no consideration of the economic costs of obesity in early obesity
science. Only the author of the OHE report discussed the issue, as this was in keeping
with the organisation’s remit to examine the economic issues around healthcare. It is
also a pharmaceutical industry funded organisation and pharmaceutical treatments for
overweight and obesity have always been very profitable. In contrast to later accounts
of the expense of treating obesity, the author of the OHE reports described the costs of
obesity to the NHS as relatively insignificant, as they derived from visits to GPs,
some out-patient appointments and pharmaceutical treatments. Obesity was suggested
to result in a small proportion of GPs work – 1.5% of consultations in the UK – and
the total costs of obesity treatments were calculated as £3.5 million, although the
author did acknowledge that the retail market for over the counter weight-loss
products and dietary aids was significantly larger (OHE, 1969 : 26).

Medical textbooks and reports outlining research findings did not include discussions
of the costs of healthcare at this time, which may be why many of my sources did not
discuss these issues. Health economics was also a relatively underdeveloped
discipline (Ashmore et al., 1989: 5 - 6). But, from the mid 1960s there was concern
about the increasing costs of the NHS, particularly the rising costs of drugs (Webster,
2002: 46-7). This concern increased after the UK economy slumped following the oil
crisis of 1973, and again after the industrial unrest of the Winter of Discontent in
1978/9 (Berridge, 1999a: 55). Discussions of the economics of particular conditions,
and the development of new treatments, therefore, took place in a context of severe
economic constraints. However, as many of the individuals in this coalition were
employed in the relatively well-funded hospital sector of the NHS, these
considerations may have seemed less important to them.

In 1983, the wider clinical medical community claimed ownership of the problem of
obesity when the Royal College of Physicians produced a report on the topic. This

130
report was broadly similar in content to the other sources discussed above and can be
considered as the culmination of the individual paradigm of obesity. The report’s
authors discussed the BMI, but defined overweight and obesity as greater than 120%
of the upper limit in 1960 Metropolitan Life tables. They considered research into a
range of endocrinological and metabolic causes of excess body weight, but concluded
that in the majority of cases such changes were secondary to weight gain. Their
account was firmly based in Garrow’s work on the energy balance model (see chapter
4), and considered the physiological mechanisms underlying food intake, the role of
declining levels of physical activity and (in much less detail) ‘social factors’ such as
parental pressure, food availability and customary eating patterns. For the report’s
authors the management of obesity involved unsupervised slimming, dietary treatment
under medical supervision, behavioural therapy (the results of which were sometimes
‘disappointing’) and appetite suppressing drugs.

However, there were three important areas of difference between this report and
earlier publications. Firstly, the RCP report was the first to use larger scale British
data on body weight. In the past ‘Selected studies have been undertaken in different
parts of the country but the criteria for assessing overweight and obesity have varied’
(RCP, 1983: 20). In this report, the authors could cite prevalences derived from a
1981 OPCS survey which was designed to be representative of the adult British
population, and which used BMI cut-off points to define overweight and obesity.
According to this new data 6% of men and 8% of women were obese and 34% of men
and 24% of women were overweight (RCP, 1983: 21). Apart from a 15% increase in
numbers of overweight men, these figures were very close to the Metropolitan Life
figures quoted at the beginning of section 3.4.1.

Secondly, the RCP report used data from both Framingham and the American Cancer
Study (ACS) (ACS, 2010)45 to discuss the health consequences of obesity. Data from

45
This was a prospective mortality study of nearly a million adult men and women who were enrolled
in 1959 and 1960, and followed up until 1972. Participants came from 25 US States, having been
recruited in household units by ACS volunteers. On enrolment ‘each participant completed a four-page
baseline questionnaire providing information on height, weight, demographic characteristics, personal
and family history of cancer and other disease, menstrual and reproductive history (women),
occupation, diet, alcohol and tobacco, and physical activity’ (ACS, 2010). Supplemental questionnaires
were sent to participants in 1961, 1963, 1965, and 1972, and there were annual follow-ups to check
whether participants were still alive. If they died death certificates were used to establish cause of
death.

131
Framingham and other American studies had been used before, but the ACS study
was a new resource and it led to an expansion in the number of conditions associated
with excess body weight. Specifically, it led to several cancers being included in the
list of negative health consequences of overweight and obesity (see chapter 7). And
finally, the report’s authors discussed the importance of preventing obesity: ‘The high
prevalence of overweight in the community means that public health measures as well
as individual health education and medical advice are needed’ (RCP, 1983: 52).
Suggested interventions included breeding animals for less fatty meat, doorstep
delivery of low fat milk, the manufacture and promotion of low fat and low sugar
convenience foods, increasing consumption of cereals fruit and vegetables, financial
measures to decrease consumption of alcohol, provision of advice on behavioural
change, and standard nutritional labelling of food. In addition, catering organisations
should provide more varied and healthier menus and doctors should provide
information on appropriate diets to parents of overweight babies and children (RCP,
1983: 50 - 51). Although firmly based on an individual behavioural change model this
was a call for wide-ranging government action in the areas of food production and
health that overlapped in several places with proposals in the Canterbury Report on
preventing heart disease (see section 3.2). It does not, however, seem to have been
influential.

This report was the last in the early series of reports: the next report specifically on
the topic of obesity was not published until 1994. It is not clear why the topic of
obesity was less of a priority for the British government in the late 1980s and early
1990s. A partial explanation may be that one of the key individuals in this field –
Professor Philip James – began to work on WHO publications rather than UK
government ones. After 1983 he worked on a technical report on prevention and
control of cardiovascular disease (WHO, 1986) wrote a report on healthy eating for
the WHO European Office (James et al., 1988) and was involved in producing the
technical report on nutrition and the prevention of chronic disease (WHO, 1990).
Following the political controversy around the report of National Advisory Council
on Nutritional Education (NACNE, 1983), he may have found the subject of nutrition

132
and health less politically charged in WHO than in the UK (Cannon, 1987)46. Despite
Professor James’ administrative energy and engagement with the topic of obesity, an
explanation that depends on the absence of one individual is not entirely satisfactory
and further research is required to explain this finding.

3.5 The obesity coalition goes mainstream

In this section I will describe how the public health coalition around obesity
successfully developed its activities in the late 1980s and throughout the 1990s. The
first two sections draw on my Master’s research and uses surveys of writing about the
obesity epidemic and contemporary obesity organisations to develop an initial
mapping of the policy arena of obesity and overweight. The final section gives more
details about the documents analysed in chapters 4 to 6.

3.5.1 Mapping the usage of the term ‘obesity epidemic’

Between June 2006 and January 2007 I conducted retrospective searches of Medline,
Web of Knowledge and other scientific journal databases for the term ‘obesity
epidemic’ (Fletcher, 2007). I was investigating the usage and spread of the term
‘obesity epidemic’ as it seemed to mark a new understanding of the increasing
incidence of an important public health problem. Combining Medline and Web of
Knowledge gave the widest range of results, and further searches produced increasing
duplication of results. The search results from these two databases were, therefore,
combined and the results cleaned up by removing duplicate and non-relevant results
to give the results in table 3.2 below. These results show a sharp increase in the use of
this term in the scientific press from 1999 onwards.

YEAR NUMBER OF ARTICLES

1996 2
1997 3
1998 5
1999 10

46
Chaired by WPT James, NACNE produced a report in 1983 that was alleged to have been suppressed
by the Conservative government at the time.

133
 2000 12
2001 26
2002 34
2003 48
2004 87
2005 101
2006 114

Table 3.2

Before 1996 the term ‘obesity epidemic’ was used only occasionally and often
interchangeably with ‘epidemic obesity’. As table 3.1 shows, usage of the term in
scientific journals increased slowly between 1996 and 2000, and then after 2000
increased very rapidly. In parallel to this increase in the scientific literature, a search
of LexisNexis showed that there was an increasing usage of the term ‘obesity
epidemic’ in the popular press. I tracked this increase by searching the content of two
British daily newspapers, the Daily Mail and the Guardian for the same period: the
former is politically centre right and socially conservative the latter politically centre
left and socially liberal, but both regularly contain detailed coverage of health issues.
The results of that search were very similar to those in the table above and showing a
pattern of increasing usage of the term after 2000. In this period, there were also a
number American authors publishing popular science books about nutrition science
and the public health aspects of increasing rates of excess bodyweight (Fumento,
1997, Pool, 2001, Critser, 2003, Shell, 2003, Brownell and Horgen, 2004, Oliver,
2006). Some of this writing was reproduced in the British popular press, expanding
the size of the coverage and, occasionally, the terms of the debate, by unpacking the
wider moral and political implications of these technical debates. This popular science
coverage fed off the growing scientific and policy discourse about the health
consequences of rising rates of excess body weight, but also provided a rare
opportunity for critics of the obesity epidemic discourse to provide sustained counter-
arguments, for example against the linkage made between excess bodyweight and ill-
health or the use of BMI-based definitions of overweight and obesity.

Contrary to my expectations, and popular discourse, there was no controversy

amongst the various professional groups, who might be thought to have a stake in this
discourse, about the rapidly increasing usage of the term ‘obesity epidemic’. For
example, I thought that epidemiologists might object to this new application of the

134
term to a non-infectious risk factor for chronic disease, rather than an infectious
disease like influenza. In fact, I would argue that the rapid spread of this usage
demonstrates the existence of a previously developed consensus (or at least lack of
controversy) around this usage.

The most heavily cited articles containing the term ‘obesity epidemic’ came from
journals such as the Journal of the American Medical Association, Science, the New
England Medical Journal and The Lancet. Articles in these leading journals were
often written by individuals who contributed to reports, such as the WHO Technical
Report 894 Obesity: Preventing and Managing the Global Epidemic, who were thus
simultaneously developing scientific knowledge and public health policy. In this
period British and American research into the health effects of excess body weight
was characterised by ‘scientific entrepreneurs’, high profile scientists who used their
status to publicise a particular problem and ‘promote it both within and beyond the
parameters of science’ (Hannigan, 1995: 153). These scientific entrepreneurs were
publicising their ideas by writing in high profile general science journals, and their
articles were widely cited both by other researchers and by journalists from the
scientific and popular press.

However, the journals that contained most number of articles using the term were
different from these high prestige general science ones: they were journals such as the
International Journal of Obesity and Related Metabolic Disorders, Obesity Research,
The Journal of the American Dietetic Association. These are the specialist journals in
which researchers developing a detailed body of knowledge write. It also did not seem
that this usage spread from a particular group of specialist journals: the pattern I found
seemed to be more one of sporadic uses of the term ‘obesity epidemic’ that became
increasingly frequent after the publication of the WHO Technical Report 894 in 2000.
This term derived much of its initial legitimacy from the activities of expert
committees developing public health policy and this legitimacy was then re-
appropriated by several different communities of research scientists who used it in
article titles and abstracts as a way of highlighting the importance of their ongoing
research in the areas of diet, nutrition and chronic disease.

135
3.5.2 Contemporary British obesity organisations

The most active British obesity organisations between 1996 and 2006 were
professional medical groups such as the Association for the Study of Obesity (ASO).
ASO was set up in 1967 and describes itself as ‘the UK's foremost organisation
dedicated to the understanding and treatment of obesity’ (ASO, 2011). It is an
organisation for medical professionals working in the area of diet, nutrition and body
weight. High profile committee members include Dr Susan Jebb of the MRC Human
Nutrition Unit in Cambridge and Dr Peter Kopelman, a diabetes specialist based at St
George’s Hospital London. ASO organised the First International Congress on
Obesity (ICO), held in London in 1974, and also publishes the International Journal
of Obesity (from 1977). The impetus for the1974 London conference and the journal
both came out of the work done to prepare for the first Fogarty Center International
Conference on Obesity in 1973 in Bethseda. The second ICO and the second Fogarty
Center International Conference on Obesity were both held in Washington DC in
1977. The ICO is still held every three years, the last one being in held in 2010 in
Stockholm.

In the 1980s obesity science became more established and the organisation expanded
geographically and diversified its activities. The North American Association for the
Study of Obesity (NAASO) was founded in 1982, the International Association for
the Study of Obesity (IASO) in 1985 and the European Association for the Study of
Obesity (EASO) in 1986. IASO now acts as an umbrella organisation, currently for 52
national obesity associations. It also publishes a range of journals including the
Obesity Newsletter (from 1986), Obesity Reviews (from 1988) and the International
Journal of Pediatric Obesity (from 2005). The annual (or sometimes biennial)
European Congress on Obesity also began in 1988 with a meeting in Stockholm.

In 2002, IASO merged with the International Obesity TaskForce (IOTF), founded in
1995 by Philip James as a ‘policy and advocacy “thinktank”’. According to official
IASO history ‘the IOTF prepared the first scientific research report on the global
epidemic of obesity, which served the basis of the first WHO expert consultation on
obesity held in Geneva in 1997’ (IASO, 2011). As part of IASO, the IOTF now
functions as ‘a global network of expertise, a research-led think tank and advocacy

136
arm of the International Association for the Study of Obesity’ (IOTF, 2011).
Individual members of the IOTF, such as Philip James and other Rowett Research
Institute researchers, were centrally involved in the production of the WHO Technical
Report 894, Obesity: Preventing and Managing the Global Epidemic. The IOTF has
subgroups working in the areas of childhood obesity, obesity prevention, obesity
management and the economic costs of obesity. Its main activities seem to consist of
producing research reports and organising conferences in conjunction with other
bodies such as the World Heart Organisation. 47

Other obesity organisations were formed by or aimed at other groups of healthcare

practitioners, members of the public and patients. In the UK, the National Obesity
Forum (NOF) was set up in 2000 by a group of GPs ‘to raise awareness of the
growing health impact that being overweight or obese was having on patients and the
National Health Service (NHS)’ (NOF, 2011). It is an umbrella campaigning group
that tries to enrol both healthcare professionals and members of the public in a
campaign against increasing excess bodyweight. The NOF has a long list of partner
organisations that includes ASO, drug companies, food companies and diet industry
companies. It does not seem to contain any particularly high profile committee
members. The All Party Parliamentary Group on Obesity, a subdivision of the NOF,
was set up in 2001. It was chaired by Dr Howard Stoate MP and Vernon Coaker MP,
neither of whom are particularly well known outside parliament, and produced
thirteen reports on various aspects of obesity, but seems to have been inactive since
2008, despite a re-launch in 2006. Despite their attempts to become involved in the
growing obesity epidemic discourse, neither of these organisations became as
influential as the elite practitioner groups, such as ASO, IASO and the IOTF.

By 2005, there were a couple of patient organisations operating in this area. The most
high profile was The Obesity Awareness and Solutions Trust (TOAST) set up in 1998
with the stated aim of enabling people whose lives were affected by obesity to
influence both policy and treatments. However, its activities were limited and it went

47
As well as the conferences organised by organisations such as IASO/IOTF and ASO there are also
regular obesity conferences organised by commercial companies, such as Childhood Obesity or the
Annual Obesity Europe Conference . These are hard to document since they do not last online: once the
conference is over its details disappear, and advertising starts for next year’s event.

137
bust in 2006 amid accusations of fraud. Another patient group, the British Obesity
Surgery Patient Association (BOSPA) was set up in 2003 and is still active. Neither of
these groups managed to become high profile participants in the debates around
obesity taking place in the British scientific and popular press between 2000 and
2005.

Other private organisation who became more opportunistically involved in the

discourse about obesity and the obesity epidemic included the International Life
Sciences Institute (ILSI), a food-industry-funded body that has identified
overweight/obesity as one of four Global Science Issues (www.ilsi.org). Another was
the RAND Corporation whose senior economist wrote regularly on the health risks of
obesity and the links between obesity and chronic ill-health (Sturm, 2002, Sturm,
2008, Sturm, 2007, Sturm et al., 2007, Sturm and Cohen, 2009). This increasing level
of activity demonstrates how the obesity epidemic discourse was becoming widely
distributed and a resource to which new, non-medical groups could contribute.
However, this expanding group of contributors did not diminish the authority of
professional medical groups, such as ASO, IASO and the IOTF: they remained the
most legitimate creators of knowledge about the individual and social consequences
of increasing rates of excess body weight.

3.5.3 Principal sources for chapters 4, 5 and 6

In these chapters I consider nine further publications: four reports on the topic of
obesity and three textbooks by key researchers. The first textbook is a special edition
of an endocrinological periodical edited by Philip James, with contributions from
academic medical researchers from the UK, North America and Europe. The North
American contributors included two high profile American researchers, George Bray
(see below) and F. Xavier Pi-Sunyer (see section 6.3)48, and other, less well known
researchers from institutions such as Harvard School of Public Health. The British
contributors were from teaching hospitals such as St Georges in London and
Ninewells in Dundee, and the Dunn Nutrition Laboratory in Cambridge, and the

48
Both these individuals are endocrinologists specialising in diabetes and obesity At the time of
publication, Dr Bray taught at the School of Medicine in the University of South California and
Professor Pi-Sunyer was associate professor of medicine at the College of Physicians and Surgeons in
Columbia University New York.

138
European contributors were from universities in Lausanne, Geneva and Paris. This is
similar in composition to the earlier symposia described in section 3.3, apart from the
wider geographic spread of participants.49

The second textbook is an edited collection resulting from an international meeting on

body weight control that took place in Montreux, Switzerland in April 1985.
Participants who have already been referred above included George Bray, John
Garrow, Philip James, Derek Miller and Trevor Silverstone. Other high profile
participants included Albert Stunkard, professor of psychiatry at the University of
Pennsylvannia; Audrey Eyton, founder of Slimming Magazine and author of The F-
Plan Diet; Theodore B VanItallie, professor of medicine at Columbia University and
co-director of its Obesity Research Center; Geoffrey Rose, professor of epidemiology
at the London School of Hygiene and Tropical Medicine, and Per Bjorntorp, professor
of medicine at Gotenberg University, Sweden. Other participants came from
universities and hospitals in France, the Netherlands, Germany and Switzerland and
seven of the thirty two participants were from Finland, mostly from the National
Public Health Institute (NPHI) in Helsinki. The Finnish group included Audilikki
Nilesen co-principal investigator of the North Karelia project, Pekka Puska, professor
and director of the epidemiology department of the NPHI and Jakko Tuomilehto, head
of the MONICA data centre, also at the NPHI. The presence of such a large and
distinguished Finnish group shows the ongoing links between research into heart
disease and obesity science (see chapter 2)50. One of the editors, Arnold Bender, was
an emeritus professor of nutrition at Queen Elizabeth College and, therefore, a former
colleague of John Yudkin (see section 3.3). The final textbook was written by John
Garrow (see section 3.3) and was one of a series he produced during between 1978
and 2000.

The first of the reports is a 1994 Office of Health Economics report titled Obesity
written by Richard West. I have been unable to find out anything about this author
apart from that he worked for the OHE. The second report, Obesity: Reversing the

49
This presumably partly reflects the reduced cost of requesting the contribution of an article rather
than attendance at an event.
50
Pekka Puska, George Rose and Philip James had previously worked together on a WHO technical
committee responsible for compiling technical report 732, Community prevention and control of
cardiovascular diseases (see section 7.2).

139
Increasing Problem of Obesity in England, published in 1995, was part of the Health
of the Nation programme and came out of a symposium on obesity held in February
1994 by two Department of Health bodies, the Nutrition and Physical Activity Task
Forces. Neither of these bodies appears to be currently active. The chairman of both
taskforces was Philip James and, as well as John Garrow and Peter Kopelman,
members included civil servants (especially the secretariat which was mostly drawn
from the Department of Health), academics from psychiatry, physical education and
nutrition departments and representatives of various organisations. These included
industry groups such as the National Dairy Council, Slimming Magazine Clubs, and
the British Retail Consortium, professional groups such as the Association for the
Study of Obesity, the Coronary Prevention Group, the British Nutrition Foundation
and the British Dietetic Association and quangos like the Health Education Authority
and the Sports Council. The political momentum necessary for the organisation of the
symposium that produced this report ultimately seems to have derived from the target
set out in the 1992 DoH Health of the Nation White Paper for the reduction of obesity
by 25% for men and 33% for women as part of a wider set of targets to reduce the
levels of mortality and morbidity cause by CVD (DoH, 1992: 55). In the late 1990s or
early 2000s, when it became apparent that obesity rates were rising rather than falling,
these targets appear to have been quietly abandoned.

The Scottish Intercollegiate Guidelines Network publication Obesity in Scotland:

Integrating Prevention with Weight Management was published a year later in 1996.
This document was published by the RCP in Edinburgh, and demonstrates the shift
towards evidence based medicine since the authors stipulated the type of evidence
used for each of their recommendations and graded it using a system developed by the
US agency for Health Care Policy and Research (SIGN, 1996). As a way of
acknowledging previous policy concern about obesity, the report’s introduction
referred to the 1977 DHSS/MRC report, the 1983 RCP report, as well as targets
included in the 1992 Health of the Nation White Paper and the 1993 Scottish Diet
report (Scottish Office, 1993). This later reference appears to be mostly rhetorical
since the Scottish Diet report focused on health eating and did not discuss the subject
of obesity in any detail.

140
The final report was produced by the British Nutrition Foundation (BNF). The BNF
describes its function in the following terms:

The Foundation promotes the nutritional wellbeing of society through the

impartial interpretation and effective dissemination of scientifically based
knowledge and advice on the relationship between diet, physical activity and
health. It works in partnership with academic and research institutes, the food
industry, educators and government. (BNF, 2004a)

It is open only to corporate members and seems to be a food industry funded

organisation that works to publicise nutritional research by means of organising
conferences, seminars and lectures, developing educational material, funding
research, producing reports on particular topics and providing nutritional information
to journalists and broadcasters (ibid.). Judging by its publications, the organisation has
existed since the late 1990s and to date it has produced eight taskforce reports on
topics such as cardio-vascular disease, trans-fatty acids, oral health and healthy
ageing. These publications ‘are written by internationally recognised experts’ to
‘provide a comprehensive and authoritative review of a particular area of nutrition
science’ (BNF, 2004b). The remit of the task force that produced the BNF’s report on
obesity was to review ‘the present state of knowledge of the causes, consequences,
prevention and treatment of obesity’ (BNF, 1999: xi) in order to make policy
recommendations and identify future areas of research. This is very similar to the
remit for the expert committee that produced the WHO technical report on obesity
(see chapter 8). The chair of the taskforce was John Garrow, who was now apparently
retired. Other taskforce members include Dr Ann Fehily from H J Heinz Company
Ltd; Dr David Mela from Unilever Research; Peter Kopelman (see above); Professor
Mike Stock, a physiologist from St George’s Medical School, London; Professor Jane
Wardle, a health psychologist from University College London and Professor Andrew
Prentice, a nutritionist from the MRC International Nutrition Group at the LSHTM.
Observers came from the DoH, MAFF and the Department of Dietetics and Nutrition
at Leeds General Hospital. Named contributors to the report included Dr Tim Cole a
biostatistician from the Department of Epidemiology in the Institute of Child Health,
London; Gail Goldberg from the MRC Dunn Clinical Nutrition Centre in Cambridge;
Dr Susan Jebb from the MRC Human Nutrition Research Unit, Cambridge; Dr James

141
Stubbs, a behavioural researcher from the Rowett Research Institute in Aberdeen51,
and two members of the faculty of Agriculture and Biological Sciences at the
University of Newcastle. This is a much wider academic network than existed for
other British obesity reports, both geographically and institutionally, since it went
beyond London hospital based biomedical researchers to include representatives from
other institutions, including three different MRC funded nutrition research centres and
a department of agricultural science.

This body of material demonstrates the increasing success of British obesity science
as reports on obesity were now being produced by increasingly influential bodies –
including taskforces convened specially by the Department of Health – and involving
greater numbers of participants from a wider range of organisations.

Diagram 3.1 the policy coalitions that developed around each of the
identified risk factors for coronary heart disease in British chronic disease
epidemiology, 1970-2000

51
At this time Philip James was Director of the Rowett Research Institute which conducts research into
human and animal nutrition. It was founded in 1913 and its first director was Dr John Boyd Orr (later
Lord Boyd Orr).

142
3.6 Conclusion

We are unanimous in our belief that obesity is a hazard to health and a

detriment to well-being. It is common enough to constitute one of the most
important medical and public health problems of our time, whether we judge
importance by a shorter expectation of life, increased morbidity or cost to the
community in terms of both money and anxiety (DHSS/MRC, 1976: 1).

The authors of the DHSS/MRC report produced no evidence for this dramatic
assertion and it contradicted their evidence of a small average increase body weight
having taken place since the 1940s (see section 3.4.1). As a pharmaceutical industry
funded body the OHE had an obvious economic interest in disseminating knowledge
about “new” health problems. But the financial interests of a government department
such as the DHSS would have gone in the opposite direction, since a widespread and
hard to treat new health condition could lead to significant increases in health
expenditure. They legitimated their choice of topic by referring to ‘widespread
anxiety among the general public’ on the subject of obesity (DHSS/MRC, 1976: ix).
Such reports are exercises in framing arguments as well as fact gathering (Hilgartner,
2000), and so it should not be assumed that such public anxiety existed, but it was a
useful justification for a technical piece of writing on medical research findings.

In this chapter, I have outlined the formation of the obesity coalition, modelled on the
anti-smoking coalition. I have briefly described the articulation by that coalition of a
set of largely unchallenged claims about the nature and public health significance of
obesity that I identify as the “individual paradigm” and that serve to justify further
research into obesity, especially its causes and prevalence. I have illustrated how a
limited amount of American evidence, on changing average body weights and the link
between obesity and heart disease, was picked up by an emerging British public
health coalition and made the basis of expansionist claims-making about obesity.
Obesity was seen as an individual condition, of limited prevalence in the population

143
which resulted from eating too much due to lack of self-control, had a limited range of
health effects that were largely related to heart disease, diabetes and the mechanical
effects of heavier body size, was treated by diet, and which cost healthcare services
relatively little (compared to the sums spent on commercial dietary services and
productions).

In the next few chapters I will go on to document further how key aspects of these
claims were elaborated and substantiated through further definitional work and
research. This will develop my account of the individual paradigm as I move from
looking largely at the articulation of a set of claims about obesity, to also include
fundamental practices such as the measurement of obesity and the assessment of
different treatments.

144
CHAPTER 4: THE DEVELOPMENT OF THE BODY MASS INDEX 1972-
1995

4.1 Introduction

In the previous chapter, I referred to the competing definitions and measures of

overweight used in early obesity science. Often these were based on Metropolitan
Life Insurance Company statistical tables of ideal weights. These were developed
from height and weight tables (classified by frame size) correlated with mortality
statistics and percentages of average weight to define categories of excess weight
(overweight was 10% above average weight, and obesity as 20% above, for example).
In this chapter I will discuss the problems of defining obesity and overweight in
greater depth before describing the development a new index – the body mass index
(BMI) - that combined a measurements and definition of obesity and overweight and
rapidly became the standard measure throughout American and British obesity
science. Eventually the BMI would become a global standard when it was (partially)
adopted by the World Health Organization (WHO) in the mid 1990s.

4.2 The measurement of body fat

The basic definition of obesity is an excess of body fat (for example, see Garrow,
1988: 1). This is relatively easy to identify qualitatively – standards for judging
bodyweight vary across time and place, but can often be applied without any medical
training. But the precise percentage of body fat is very difficult to measure accurately.
Moreover, the relationship between body fat (often understood as the body’s energy
stores) and body weight is not straightforward:

Body weight is easy to measure to one hundredth of one per cent with a simple
beam balance, but the weight of the body bears no simple relationship to the
size of the energy stores, nor is a change in energy stores necessarily reflected
in a change in weight (Garrow, 1988: 25).

Despite this complex relationship, all but the most detailed and laboratory based
approaches to measuring body fat use relative body weight as a proxy measure for fat,
because of the difficulty of measuring exact body composition, which makes it

145
impractical for use in ordinary clinical practice. Body weight and height are simple to
measure and routinely collected in primary health care, and so an index based on
these measures is more economical to collect at the population level.52

A small number of laboratory-based studies have sought to calculate the precise

proportion of body fat through the dissection of dead human bodies. Nonetheless,
such studies still rely on the assumption that the fat and non-fat components of the
body can be separated unproblematically, and that ‘the body consists of fat and fat-
free tissue of constant composition’ (Garrow, 1988: 29), even though research into the
physiological effects of weight loss, amongst other processes, has shown this not to be
the case (see below). Despite this, Garrow (1978: 115) produced a table of the
chemical composition of the human body by compiling the results of 1950s British
and American chemical analyses of a very small sample of dead bodies.

Our understanding of the composition of the human body is based on chemical

analyses of 6 cadavers which were performed between 1945 and 1956.
Mitchell et al (1945) analysed the body of a 35-year-old white male who died
suddenly of a heart attack. Widdowson et al (1951) analysed 2 adults and 1
child; the adults were a man of 25 who died of uraemia and a woman of 42
who drowned herself. Forbes et al (1953) analysed a man of 46 who died of a
fractured skull, and Forbes et al (1956) reported two more analyses: one a
Negro (sic) male with bacterial endocarditis who died aged 48, and the other a
man of 60 who was found dead, presumably of a heart attack (Garrow, 1988:
26).

Even within this small set of results, the parameters varied widely: the proportion (in
g/kg) of water varied from 674 to 775 and that of protein varied from 165 to 238,
apparently without being related to age. The figures for potassium also varied but
within a smaller range (see Garrow, 1988: 27). These variations are important because
such data has been used in physiological research to estimate both body composition
and the effects of weight changes on it. When the practicalities of such an analysis are
considered53, it is perhaps not surprising that so little data exists, but it leads to the
conclusion that this area of research in human physiology relies on a surprisingly

52
A similar case has been made for the use of waist circumference and waist to hip ratios (WHR) (see
section 3.2 below).
53
Such an analysis involves dissecting the cadaver and reducing it to its constituent parts by scraping
the bones clean and extracting the fat by boiling or dissolving it in ether.

146
small and selective set of crucial data. As part of his wider discussion of metrology,
O’Connell (1993) outlines a larger history of the measurement of body composition
that includes chemical analyses of 3, 6 and 25 cadavers performed between 1902 and
1950, hydrostatic weighing by submersion in water, and the development of portable
measurement instruments used in health clubs and weight-loss centres. These
developments in measurement technologies mean that

A body must still be dissected to realize its composition directly, but body
composition has been translated into more easily measured quantities that have
been accepted as representations of it (O'Connell, 1993: 130-1).

Present-day methods of indirectly estimating the relative proportions of lean body

mass (also often known as fat free mass or FFM) and fat mass are technically
complex and involve measuring body density, its water content or its potassium
content, or using CAT and MRI scanners or dual-energy X-ray scanning (BNF, 1999:
17 - 19). All of these methods are both time-consuming and expensive, and these
authors therefore recommended more practical clinically applicable methods such as
skinfold measurement, weight and height and combined with waist, hip and thigh
measurement, bio-electrical impedance measurement and near infra-red reactance54
(BNF, 1999: 19). A study of 106 adult women showed that ‘92% of the variation in
body weight is explained by the variation in body fat’ (BNF, 1999: 18)55 which was
judged to be a high enough percentage for body weight to be used as a reliable
estimate of body fat (with a few well known exceptions, see section 4.6). Studying the
effects of weight loss diets required knowledge of their effects on body composition –
diets that result in weight loss mostly due to water loss or loss of lean tissue were
widely recognised to be both dangerous and ineffective (see for example Garrow,
1988). However,

54
Skinfold measurement relies on the fact that most body fat is stored immediately under the skin, and
the amount of this fat is measured using calipers. However, despite the apparent simplicity of this
method, there has been a significant amount of research investigating the necessary number of
measurements and the best sites on the body to select in order to minimise potential measurement
errors and discrepancies (Garrow, 1988: 37 – 39). Bioelectrical impedance uses a measure of the
body’s conductivity to assess the relative proportions of water and fat, and near infra-red reactance
show the amount of fat present in subcutaneous tissue which is combined with anthropometric data to
give an estimate of body composition.
55
Other sources of such variation include the necessary assumptions made about the density, water
content or potassium content of lean body mass. For example, fluid retention leads to a lowered
potassium level which in turn means that the FFM is under-estimated and fat over-estimated (BNF,
1999: 18).

147
 it is more difficult to measure change in body composition with treatment. The
same amount of weight loss (and hence decrease in BMI), brought about by
two different treatments may signify different proportions of loss of fat, lean
tissue and water. At present, there are no methods that will reliably measure
loss of fat in an individual with an accuracy of better than 1kg (BNF, 1999:
21).

Given that the approximate weight of fat in an average human body has been given
earlier as twelve kilograms (BNF, 1999: 17), one kilogram is a significant margin of
measurement error, although if it was expressed as percentage of fat (which is used
for other measurement methods, such as density and skinfolds), it will be much larger
for a thin person than for a fat person (Garrow, 1988: 29). In spite of such error
margins, measurements of body fat were significant because of their use in estimating
body composition and, therefore, in judgements about the healthiness (or not) of
individual physiques.

4.3 Ancel Keys and George Bray redefine W/H²

As I described in chapter 3, early obesity science writing used the Metropolitan Life
ideal weight tables. But by the 1990s, these tables were replaced by the BMI which
has become the standard method of both measuring and defining obesity. In
comparison to the previous diversity of definitions that I outlined in chapter 3, it is
striking how, after the mid 1980s, reports on obesity and textbooks use BMI to define
obesity and classify obesity in a highly standardised fashion.

W/H²56 was first developed by the Belgian statistician and astronomer, Adolphe
Quetelet for use in comparing populations. Quetelet was important in the history of
human statistics because he was the first person to argue that human physical
characteristics such as weight and height follow a normal or Gaussian distribution
(Desrosieres, 1998, Hacking, 1990). However, as Quetelet’s interest was in
comparing populations, he did not use W/H² to assess individuals (Desrosieres, 1998:
73 - 77).

56
W/H² (weight divided by height squared) is the mathematical formula used to calculate BMI.

148
In 1972 Ancel Keys was one of the authors of an article arguing that W/H² was the
most useful of the available indices of relative weight for indicating the relative
proportion of body fat in individuals, and suggested that it be re-named the body mass
index or BMI (Keys et al., 1972). Keys was the lead investigator of the Seven
Countries Study (see section 2.5) and although it is not mentioned in the article the
data used in the 1972 article came from this study. The authors compared the
usefulness of W/H² to that of W/H, the ponderal index (W/H³), and percentage above
average weight, and argued that W/H² was superior on two counts. First, W/H² was
less sensitive than W/H to variations in population height, provides a better measure
of overweight rather than simply of stature. Secondly, W/H² was found to have a high
correlation with skinfold thickness and body density (which were used to estimate
body fat). BMI was thus found to be, ‘if not fully satisfactory, at least as good as any
other relative weight indicator as an indicator of relative obesity’ (Keys et al., 1972:
339). While this was not a particularly enthusiastic endorsement, Keys noted that
W/H² had one additional advantage as an indicator of obesity: it provided both a very
clear index of change in the weight of an individual, and an effective means of
comparing the obesity of individuals of different heights but the same weight:

The important difference between the properties of the ponderal index and
those of the ratios W/H, W/H² and W/H³ is apparent when calculations are
made with increasing weights at constant height. A given increase in weight
with constant height will produce exactly the same percentage increase in the
values of all the ratios W/H, W/H² and W/H³ but a much smaller increase in
the ponderal index. Consider a man 1.70m who weighs 60 kg and then gains
15kg. He gains 25% in weight; his value of W/H changes in the same
proportion, his body mass index W/H² changes from 20.76 to 25.95, i.e. it
increases by 25% also. But his ponderal index changes only from 2.3029 to
2.4807, an increase of only 7.7 per cent. Now consider two persons of the
same weight of 60kg, one is 1.70m tall, the other 1.45m in height. The
ponderal index of the shorter person is (2.700)(2.303) = 117.2 per cent that of
the taller person, while the percentage comparisons using W/H, W/H² and
W/H³ are 117.2, 137.4 and 161.1 respectively. Of the various indices
considered, the ponderal index is the least sensitive to differences in weight
(Keys et al., 1972: 340).

This was, therefore, a relatively sensitive and discriminating measure of relative body
fat that could be usefully applied in situations of increasing average body weights
both to describe population changes and to compare the relative fatness of different
individuals.

149
A classification for body weights based on BMI values was proposed by George Bray
at the first international conference on obesity in 1979, organised by the NIH and held
in Bethesda, Maryland. The results of this conference were published under the title
Obesity in America (Bray, 1979a) and the introductory overview chapter contains a
nomogram57 which includes BMI classification system for men and women that is
summarised in table 4.1

women men
obese BMI > 30 BMI > 30
overweight BMI > 23.5 BMI > 25
acceptable BMI > 18.5 BMI > 20

Table 4.1 (adapted from Bray, 1979b: 6)

This classification was not discussed in the text of the introductory chapter but a
subsequent chapter did refer to the difference between relative weight definitions of
obesity and measures that combine height and weight

Various indexes involving height and weight have also been tested. However,
they can never provide anything more than an index of overweight, since they
falsely suggest that a muscular football lineman is obese and they fail to
characterize a patient with atrophic muscle mass and increased body fat. The
so-called body mass index (weight/height²) … has the highest correlation with
independent measures of body fat; but in some series this may be as low as 0.6
or less (Sims, 1979: 24).

Evidently, the contributors to the 1979 conference were aware that their chosen index
of obesity was imperfect, but a numerical index such as BMI had a variety of useful
functions (see chapter 6), and so they made a pragmatic choice from the available
options.

4.4 The British adoption of BMI: Garrow’s definitions

57
A nomogram is a diagram used for calculating a variable. The simplest version is a chart consisting
of three parallel lines which are scaled. The scales represent three related quantities and the third
quantity is derived by joining the two known quantities with a straight line, and reading off the value
where it intersects the third line.

150
Subsequent versions of the BMI-based definition of overweight and obesity with
different cut-off points were developed by John Garrow in his series of textbooks.
In his 1978 textbook, Garrow referred to BMI in a discussion of ideal weight range
linking them to mortality rates, but he did not explicitly define obesity or overweight
using BMI cut-off points. Table 4.2 shows the classification that first appears in his
1981 textbook along with a graphic representation in both metric and imperial units.
This material appeared unaltered in his 1988 textbook.

Grade III W/H² > 40

Grade II W/H² 30-40
Grade I W/H² 25-29.9
Grade 0 W/H² 20-24.9

Table 4.2 (adapted from Garrow, 1981: 2)

Grade 0 obesity was previously referred to as normal weight, and as such this re-
labelling was an extension of the range of obesity science – now it was not only
concerned with the moderately and seriously heavy (grades 1 to 3), but also with the
normal weight. This basic system remained intact for at least a decade. Later
publications, such as the 1999 British Nutrition Foundation report, with which
Garrow was associated, continued to use this classification system citing his earlier
textbook.

Garrow was also one of the few authors in this period who discussed frame sizes, but
only to argue against their use:

Life insurance tables may distinguish between large and small frame sizes, but
all attempts to refine the estimation of fatness from weight and height by
adding various body diameters have failed … There is no advantage in
specifying different weight ranges for people of large, medium or small frame
if there is no usefully accurate way in which frame size can be measured
(Garrow, 1988: 3).

This was a means of distancing his research from the earlier and very influential work
of the insurance industry. As a means of increasing the credibility of obesity science,
he was demonstrating the superiority of biomedical categories over actuarial ones.

151
Because of Garrow’s advocacy, most of the British publications discussed below cite
him as the source of this classification scheme. For example, Philip James used
Garrow’s classification when writing in 1984 (James, 1984: 636), but when writing in
a later collection refers to cut-offs based on the American NHANES data (see below)
and percentages of ideal weights taken from the Metropolitan Life tables (James,
1988: 90 - 94). Presumably, he used the definitions embedded in the different sets of
data and this switch reflects the use of American data for an American readership.
This shows that in the 1980s different classification systems were still circulating.
James is unusual in that he explicitly addressed the effect of wider social context on
defining obesity, by arguing that an overweight individual in a country like India
(where average weights of adults were lower than those of the UK and there was the
possibility of periodic famine), might live longer than a person of similar weight in
the UK:

It should be recognized, therefore that any definition of overweight and

obesity is an operational one which requires an understanding of the morbidity
and mortality associated with varying degrees of weight and/or adiposity in
each country. If this were established for each society one could then judge
which weights to specify as inappropriate (James, 1988: 90).

However, James went on to argue that such a socially embedded definition is a

‘council of perfection’, and that a simple definition covering all developed countries
is required, such as obesity as >120% of the optimal weight for height ranges in the
1959 Metropolitan Life tables: ‘Obesity is therefore, considered in anthropometric
terms and is defined on an actuarial basis’ (ibid.). This meant that although he
included BMI ranges at the bottom of his tables of guideline weights, he did not use
them as the primary definition of overweight and obesity.

4.5 The British adoption of BMI: other authors use Garrow’s definitions

Garrow’s grades of obesity

Ungraded BMI < 20 Underweight
Grade 0 BMI 20-24.9 Desirable weight
Grade 1 BMI 25-29.9 Overweight
Grade 2 BMI 30-40 Obese
Grade 3 BMI > 40 Severely Obese

152
Table 4.3 (adapted from West, 1994: 7)

The 1994 Office of Health Economics report on obesity, gave a definition of obesity
based on Garrow’s 1981 classification (see table 4.3 above) with added labels, and a
couple of examples to show how the BMI was calculated. However, the author
qualified this definition by arguing that

In practice the grades of obesity should not be used as rigidly as the

classifications may suggest since body frame and build should also be
considered, thus, the categories in reality will have an element of overlap and
flexibility. It has been suggested that a different BMI scale be used for
women, with BMI 18-23 regarded as desirable, BMI 23-28 considered
overweight and BMI over 28 judged to be obese (Bray 1979) … A clear case
can be made for the interpretation of obesity changing with increasing age.
Among Finnish men over 80 years the highest five-year survival was among
those with a BMI of over 30 (West, 1994: 7).

This quote shows the complexity of merging Bray and Garrow’s classification
schemes with data from epidemiological research, as well as debates about whether
the same limits should apply to men and women, or to different age groups. Later on
similar arguments were made about different definitions for different ethnic groups
(see below and chapter 8).

By the mid 1990s overweight and obesity was defined and measured solely in terms
of a simplified version of the BMI classification in Garrow’s 1988 textbook (DoH,
1995: 3, BNF, 1999: 6-7). There was little discussion of the relationship between
body fat and obesity or of alternatives to BMI, which was described as a ‘relatively
simple index of body fatness’ that is useful because ‘in general [it] is relatively
height-independent, ie short and tall people of similar proportions but very different
weight have similar BMIs’ (DoH, 1995: 3). This consensus was reflected in the
partial adoption of BMI-based definitions of obesity and overweight by the WHO (see
chapter 7). Moreover, by the late 1990s the consensus around the use of BMI was
well enough established that its history was being outlined in reports:

There is international consensus that tables showing weight-for-height can

conveniently be replaced by a single index. The Belgian astronomer Quetelet
observed in 1869 that, among adults of normal body build, weight was
proportional to the square of height; in other words weight in kilograms (kg)

153
 divided by the square of height in metres (m)² was constant. Keys et al. (1972)
made a similar observation and named the relationship Body Mass Index
(BMI) (BNF, 1999: 4).

The BMI was seen to derive prestige both from its origins in Enlightenment science
and its re-invention by modern epidemiological research.

It is worth noting that BMI cut-off points had an explicit age-related normativity built
into them. They were based on the average weights of young adults, and the fact that
ageing often led to weight gain was not taken to mean that standards should be
adjusted to reflect this. Two American obesity researchers explained this in their
discussion of the relationship between BMI cut-off points and population average
weights. Their data58 used population-based definitions of overweight and obesity, so
that overweight was defined as a body mass index of higher than that of the 85th
percentile for the surveyed population of men and non-pregnant women aged 20 to
29, and obesity was defined as a body mass higher than that of the 95th percentile:

The 20-29-year-old group was used as a reference population because young

adults are relatively lean and the increase in body weight which ordinarily
occurs in men and women during ageing is almost entirely due to fat
accumulation (VanItalllie and Woteki, 1987: 40).

Translating these percentile based cut-offs into equivalent BMIs meant that for men
overweight was a BMI of greater than 27.8 and severe overweight was greater than
31.1. For women the equivalent cut-off points were 27.3 for overweight and 32.3 for
severe overweight (ibid.). These alternative cut-off points were used in US
government publications until the late 1990s (see section 4.7 below).

British authors also discussed whether recommended BMIs should increase with age.
The BMI associated with lowest mortality increases from 21.4 among men aged 20-
29 with successive decades from 21.6, 22.9, 25.8 to 26.6 for those aged 60-69 (for
women the comparable figures are 19.5 aged 20-29 then 23.4, 23.2, 25.2 and 27.3 at
age 60-69). Despite these data, Garrow argued against allowing for an increase in the
desirable weight ranges at later ages:
58
This data comes from the American NHANES II (National Health and Nutrition Examination
Survey) which took place between 1976 and 1980. See chapter 7 for more discussion of the use of data
from this series of surveys.

154
 A striking feature about the curves relating QI59 to mortality ratios is that they
become very much flatter in older age groups … In the younger age groups
there is a sharp decrease in mortality ratios around or somewhat below average
weight, with a high mortality ratio in the most overweight individuals, but in
the older age groups relative weight has less and less influence on mortality
ratios (Garrow, 1988: 3).

In his judgement, the practical problems of excess weight in old age out-weighed the
decreased effect on mortality:

It cannot therefore be concluded that there is no disadvantage to an old person

being overweight, since exercise tolerance and mobility may be greatly
impaired by excess weight in an elderly person with degenerative disease of
weight-bearing joints. In practice, therefore, the classification given above of
grades of obesity serves quite well, at least over the range 20-65 years (ibid.).

Later authors continued to make the point that ideal body weights are in this range of
BMI 20 to 25, irrespective of gender or age. The British Nutrition Foundation report
(produced by a committee chaired by Garrow) stated that, despite the fact that the
BMI at which mortality is lowest increased from 20 at aged 20 to 25 at aged 50, this
did not mean that there was a health benefit from weight gain by adults: ‘In fact there
is good evidence that, for an individual, minimum mortality is associated with a
constant weight between the ages of 20 and fifty years’ (BNF, 1999: 4). This links to
the importance of weight stability which was beginning to be discussed at this time
(see section 5.6).

From the mid 1990s authors also began to discuss another aspect of obesity – body
shape – and another measurement – the waist to hip ratio (WHR). Often this was done
using the image of apple- and pear-shaped individuals. A French endocrinologist,
Jean Vague, appears to have been one of the earliest researchers to argue for the link
between body shape and health, and to classify his patients accordingly (Vague,

59
QI stands for Quetelet Index. Garrow tried to popularise this name in preference to BMI ‘out of
respect for our European anthropometric forefathers’ (Garrow, 1988: 2). His attempt failed as he was
the only person who used it.

155
1968), but this topic was not regularly discussed in English language writing for
another 30 years.60

It is thought that people’s shape as well as their weight is an important factor

with regards to potential hazards to health. A measure of “central” fat
distribution can be obtained by calculating the waist/hip ratio (WHR), by
dividing the waist measurement by the hip measurement. Patients with a high
WHR are the “wrong” shape since they are “apples” rather than “pears”. A
waist/hip ratio of above one in men and 0.85 in women has been identified as
a meaningful cut-off point associated with increased health risks (Bray, 1993).
Many women are pear-shaped, with fat on the hips and legs, whilst an “apple”
shape with fat around the middle appears to be the harmful form of obesity,
leading to diabetes and heart disease … Those with excess abdominal fat and a
BMI of over 27 may be a greater health risk than those adults with a BMI of
over 30 but who have their fat peripherally distributed (West, 1994: 7 - 8).

The BMI cut-off point of 30 was by now very standard, but this paragraph
demonstrates a further elaboration of the ideal body – it had to be a particular shape as
well as a certain size. And although this quote appears to relax the criteria for good
health using the distinction of body shape, the author stressed that the ‘pear-shaped’
individual was still at risk of the mechanical complications of obesity and, ultimately,
used body shape to argue for more stringent BMI categories. Sorting out how these
two methods of assessing risk relate to each other was difficult:

Recent studies have shown that the distribution of fat on the body is at least as
relevant to the risk of disease as total body fatness (Björntorp 1990). Excess
abdominal fat with a BMI of only 26 or 27 may lead to greater risk than a BMI
of over 30 in a person whose fat is more evenly distributed (eg Filiporsky et al
1993). Women tend to have more subcutaneous fat than men but most of this
tends to be accumulated in the hips and extremities and is not associated with
an increased risk of cardiovascular disease (Krotkiewski et al 1983). In
contrast men tend to accumulate fat around the waist … Men who have thin
limbs but a ‘pot belly’ are at especially high risk. Pear-shaped men who
deposit their fat on their hips have a low waist-hip ratio and are at lower risk.
Women with a “male” pattern of fat distribution (“apples”) have been shown
to have the same risk of CHD and diabetes (Larsson et al, 1992) (DoH, 1995:
4).

60
I think WHR was a measure used in clinical assessments of risk of heart disease for diabetes patients,
in which case these discussions would illustrate the complexities of aligning medical knowledge from
different research areas. These discussions were attempting to integrate clinical data from different
areas of practice with other data that also derived from different (although overlapping) fields of
research. However, more historical research is required in this area.

156
These authors reported that a relationship had also been found between drinking
alcohol and increased abdominal obesity (ibid.). Research was attempting to tease out
the complex relationships between related risk factors for heart disease, social norms
around diet and alcohol consumption, and physiological differences between men and
women.61

Despite defining obesity using BMI, the authors of another report on clinical
guidelines for treatment argued that the waist-hip-ratio – ‘the traditional method’ of
identifying individuals at increased risk – was sufficient, since the allowance for
variation in height introduced by BMI was unnecessary:

New research indicates that measurement of waist circumference alone is

preferable for most purposes and best reflects the intra-abdominal fat mass
without any need to adjust for height. The waist circumference should be
measured at a specific level, viz. half-way between the superior iliac crest and
the rib cage in the midaxillary line (SIGN, 1996: 3).

These instructions show that even such a seemingly straightforward activity as

measuring the waist can be rendered as a medical procedure. This report gave a table
of gender-specific waist measurement and their correlations with increased risks of
CHD that is reproduced as table 4.4 (below).

Increased Risk Substantial Risk

Men 94 cm ( 37 inches) 102 cm ( 40 inches)
Women 80 cm ( 32inches) 88cm ( 35 inches)

Table 4.4 (SIGN, 1996: 4)

A twelve year long Swedish study had also found that a high WHR was a more
important risk factor than BMI, but the causality was still seen to be uncertain: ‘it is
not yet known to what extent the health risk is a direct effect of the visceral fat, or if
the disease risk and visceral fat are both indicators of other risk factors’ (BNF, 1999:

61
The protective effects against heart disease of hormones such as oestrogen for women of
reproductive age have been researched in depth, but I am unaware of historical or sociological writing
on this topic.

157
16). Moreover, problems of universal applicability with these cut-off points were
similar to those for the BMI. After stressing that women with the same waist
measurements are at the same risk as men, one author introduced an important
exception by arguing that ‘adults from the Indian subcontinent are particularly prone
to abdominal fat deposition on weight gain, are very susceptible to glucose
intolerance and diabetes, and are more prone to coronary heart disease than
Caucasians’ (SIGN, 1996: 4). This was part of an ongoing attempt to sort out the
differences between population groups that formed an important element of later
reports (see chapter 8).

In the writing of this period, BMI-based definitions of obesity and overweight became
standard. However, due to the overlap between obesity science and different areas of
clinical practice, other measurements of cardio-vascular risk, such as WHR and waist
circumference continued to be discussed. These two indices were both considered as
clinically significant and, despite the dominance of BMI, both were the subject of
ongoing discussions as authors analysed new research findings.

4.6 The relationship between BMI and mortality

The BMI cut-off points that were used to define overweight and obesity were derived
from plots of mortality rates against BMI which often had a bathtub or a J-shaped
curve where BMI 30 or 31 marked a point of significantly increased risk as the curve
became much steeper.62 In the 1970s and 1980s the source of this evidence was still
American research, often the large epidemiological studies discussed in chapter 2. For
example, one discussion of the relationship between BMI and mortality used data
from the Build and Blood Pressure studies of 1959 and 1979 and the American
Cancer Study of 1979 (see chapter 7). The authors described the limitations of the
Build and Blood Pressure Study (BBPS) data63, but argued that the American Cancer

62
Hacking argues that the cut-off point of 25 between acceptable weight and overweight derives from
increases in other risk factors that occur at this point, and he regards it as an illegitimate distinction:
‘the obesity label for BMI over 30 does, in crude but useful terms, indicate an objective point of
increased mortality. But the overweight label for BMI does no such thing’ (Hacking, 2007: 20 italics in
the original). He cites textbooks referring to this issue, but my sources do not discuss it.
63
The reasons given were that because it comprised of those who bought health insurance it was an
unrepresentative sample of the American population – more male, affluent and healthy than the average
– and also it relied on self-reported data which was deemed unreliable, especially for weight.

158
Study showed the same curvilinear relationship between mortality and relative weight
or BMI that was shown in the older BBPS study (Hautvast and Deurenberg, 1987).
Figures 4.1 and 4.2 show these graphs.

Figure 4.1 The relationship between mortality rate (%) and relative weight as
found by the American Cancer Study (Hautvast and Deurenberg, 1987: 66)

Figure 4.2 The mortality rate (%) in relation to the body mass index (kg/m²) in
different age categories (Hautvast and Deurenberg, 1987: 67)

159
In the 1988 edition of his textbook, Garrow referred to a Consensus Conference
organised by the National Institutes of Health in 1985 to provide expert evidence for
his claim that obesity is associated with high blood pressure, high blood cholesterol,
diabetes and cancer. He argued that

Medical experts have been curiously slow to realise that obesity impairs health
and reduces longevity through an effect on related diseases. Since the first
actuarial investigation of mortality in 1903, life insurance companies have
been aware that overweight people tended to die young, and hence were less
profitable to insure (Garrow, 1988: 9 -10).

Garrow thought that this lag was due to a number of factors: the effects of obesity on
mortality take many years to become apparent; the development of conditions like
diabetes and cancer leads to weight loss, and so obscures the effect of obesity; and the
adverse health effects of smoking confound the correlation between obesity and
mortality, since ‘smokers tend to be lighter than non-smokers and to die younger, so
when studied in a population mixed with non-smokers they distort the true
relationship of weight to mortality’ (Garrow, 1988: 11). In fact, the relationship
between smoking and reduced body weight was widely accepted, and Garrow had to
argue against the idea that ideal weight ranges for smokers should be different than
for those of non-smokers:

It has been suggested than the desirable weight-for-height for cigarette

smokers differ from that of non-smokers, but there is no reasonable basis for
this. Smokers tend to be lighter than non-smokers, and have a greater risk of
early death, but when the relation of weight-for-height to mortality is
examined for smokers and non-smokers separately the curves are virtually
identical in shape for both men and women, with the smokers showing a
higher mortality at each weight (Garrow, 1988: 4).

This was part of the continuing analysis of what factors were relevant in the
specification of ‘ideal’ body weights and what could be legitimately ignored. This
issue of confounding – sorting out the true causes of disease due to the presence of
distorting factors – became widely discussed in this period.

In the 1990s American data were still being used in the framing of such arguments.
The author of one report also cited studies based on the American Cancer Study data
(see section 3.4) to argue that all-cause mortality in both men and women ‘showed a

160
gradual rise from a BMI of 25 to an almost 2.5-fold higher risk at a BMI of 40’.
Furthermore:

Severe obesity (BMI over 40) is associated with a 12 fold increase in mortality
in persons aged 25 to 35 years as compared to those with a BMI 20-25 and
BMI 30-40 with a two fold mortality increase (West, 1994: 14).

But there were also British data including a seven year study of 18 400 civil servant
which showed that for men ‘at age 45 BMI over 30 carries about three times the
mortality risk of BMI 20-25’ (ibid.). Again, the author compared the health
consequences of smoking with those of overweight and obesity, by describing how
the presence of smokers affected analysis of the relationship between mortality and
body weight:

A non-smoker, for example with a BMI 20-25 would have to increase his or
her weight to BMI over 30 in order to experience the same mortality risk as a
person with BMI 20-25 who smokes 20 or more cigarettes a day (West, 1994:
13).

Later on, to reinforce this message, he argued that for those with a BMI of over 40,
their obesity ‘is a greater threat to health than that of being within the desirable weight
range but smoking 20 cigarettes a day’ (West, 1994: 19). In this argument the health
consequences of obesity and smoking were seen as directly comparable. This
comparison was presumably made partly as a result of the growing success of the
anti-smoking public health coalition (see section 3.3).

Discussion of the relationship between body weight, smoking and mortality continued
in these reports, with evidence being used to argue for stringent standards of ideal
body weight:

Epidemiological studies which do not take account of smoking behaviour may

not reflect the risk associated with being overweight because overweight non-
smokers survive better than thin smokers. Smokers die early and account for
much of the increased mortality of the thinnest people in the population. This
has led to inaccurate suggestions that weight gain in middle age is conducive
to better health (DoH, 1995: 4).

161
The argument about how smoking confounded the relationship between body weight
and health was being used here to argue against data demonstrating that survival rates
increased with a moderate increase in body weight in later life (see section 4.8).

Direct comparisons of the effects of obesity and smoking continued to be a routine

element of these discussions. Another report argued that ‘the mortality risk of a
normal-weight adult smoker exceeds that of non-smokers with a BMI of 30-35’ but
continued: ‘both smokers and non-smokers, considered separately, show the lowest
mortality rates in the 18.5-24 range of normal BMI. Long-term follow-up of non-
smoking adults suggests an optimum BMI of 20 or less’ (SIGN, 1996: 10). The
comparison was made only for illustrative purposes, rather than as part of a treatment
argument, since in the same paragraph the authors argued that stopping smoking was
a priority even if it resulted in weight gain.

A final example of these discussions began with an illustration of the J-shaped curve
of plot of BMI age mortality ratio showing the BMI at which average mortality is
lowest. Figure 4.3 shows this graph.

Figure 4.3 (BNF, 1999: 5)

162
Describing this plot, Garrow argued that:

Between the ages of 20 and 50 years, [the BMI at which mortality is lowest]
increases linearly from a BMI of 20 to a BMI of 25, which in a person 1.73m
tall, implies an increase in body weight from 60-75kg. This does not mean, as
has been suggested, that there is a health benefit from a gain of 15kg during
adult life…..In fact, there is good evidence that, for an individual, minimum
mortality is associated with a constant weight between the ages of 20 and 50
years (BNF, 1999: 5).

There were seen to be a series of confounding factors in operation – some very

overweight individuals would be dead by 50, chronic degenerative diseases take years
to develop and so the life expectancy of younger individuals was more likely to be
limited, and conditions such as cancers and chronic infections caused weight loss and
increased the mortality rates for lower BMIs. Moreover, data from the Nurse Health
Study64 showed a different relationship since, when smokers and ex-smokers were
taken out, the J-shape of the curve disappeared: ‘Minimum mortality risk is present in
the lowest BMI groups and there is a significant increase in risk at and above a BMI
of 27’ (ibid.). Therefore, ‘the J-shaped curve, with a nadir (lowest point of the curve)
of a BMI of about 25, is caused by deaths among smokers or among those with pre-
existing disease’ (ibid.). The shaping of the standard curves was defined as an artefact
resulting from confounding factors, rather than an expression of a ‘true’ relationship.

The parallels that were drawn with smoking performed two important functions in
these discussions. Firstly, repeated comparison with an increasingly stigmatised and
marginalised habit that was framed as the result of individual choice (see section 3.4)
invited the inference that weight gain was also voluntary and resulted from individual
ignorance and poor food choice. Secondly, the logic of confounding factors allowed
researchers to argue that deaths among the underweight were due to smoking or
underlying disease (see chapter 6).

4.7 Prevalence and incidence: estimating the extent of obesity and

overweight

64
This study of 121 700 American registered female nurses began in 1976.

163
By the late 1980s, as well as data from American studies, data from large-scale British
research was also becoming available. One of these was Rosebaum’s 1985 survey of
approximately 10 000 members of the British adult population selected from the
electoral register (Garrow, 1988: 5). Using the classification from Table 3.3, the
results showed that 40% of men and 32% of women are ‘obese to some extent’ (ibid.).
In discussing the effect of age on BMI Garrow argued that,

Among both men and women the proportion in the range QI 30-35 increases
from about 2% at age 16-20 to about 10% at 60-64. Among men there are
relatively few over QI [i.e. BMI] 35, but among women the prevalence
increases with age to about 4% at age 60 years. Grade III obesity is relatively
rare: in this survey there were less than 0.5% of men in this grade at any age
group, but among women age 40 and more years about 1% were grade III. We
can calculate for the whole population about 0.1% of men and 0.3% of women
are in grade III, and these are mainly over 40 years old (Garrow, 1988: 6 - 7).

The relatively low levels of obesity cited for each age group contrast with the figures
given in modern discussions of the issue. Comparative international data on
prevalence rates also existed for the Netherlands, Norway, Australia, Canada and the
US. This shows rates of grade 1 obesity/overweight ranging from 46% for Dutch men
aged 50-64, to 11% for Norwegian women aged 20-24, and those for grade II
obesity/obesity ranging from 24% for Norwegian women aged 60 -64 and 1% for
Norwegian men aged 20-24 (Garrow, 1988: 8). The variation within populations
seems to be as large as that within populations, but Garrow’s overall summary was
that: ‘the prevalence of obesity seems higher in North American than in European
countries, with the UK in an intermediate position’ (Garrow, 1988: 9). Comparison
between populations was of more immediate concern than comparison within
populations.

Other authors discussed American data from the 1976-80 NHANES II survey which
gave the prevalence of overweight among American women as 27.1% compared to a
prevalence of 24.2% among American men (VanItalllie and Woteki, 1987: 47). Using
BMI cut-offs of 27.8 for overweight and 31.1 for severely overweight for men and
27.3 and 32.3 for women, meant that 34 million men and women were defined as
overweight, and of them 12.5 million were severely overweight (VanItalllie and
Woteki, 1987: 40). However, breaking these figures down by age, socioeconomic

164
status and ethnicity, showed that ‘several categories of people have a special
propensity to become obese; namely black women, women below the poverty line and
people with a large frame and/or large musculature’ (VanItalllie and Woteki, 1987:
50). The inclusion of this last group begs the question of what was being measured in
these surveys; as well as being a proxy for fat, weight may also be acting as a proxy
for frame size or musculature, as critics of the BMI argued (see section 4.8 below).

In the late 1990s, the cut-off points used by the US government were changed. Using
the above criteria on the NHANES III 65 data, the prevalence of overweight among the
adult population was 33.3% for men and 36.4% for women. When Bray’s and
Garrow’s now standard cut-off points were applied to this data the prevalence figures
became 59.4% for men and 50.7% for women. As two contemporary commentators
put it

By simply changing the overweight cutoffs, the estimated number of

overweight adults increases from 61.7 million (BMI 27.8 and 27.3) to 97.1
million (BMI 25.0), representing a difference of 35.4 million overweight
adults. This example calls attention to the actual effect that a shift in BMI
criteria can have on determining the population at risk (Kuczmarski and
Flegal, 2000: 1078).

An increase of over 35 million adults is a major shift and, although this quote
appeared in a clinical nutrition article about twentieth-century American weight
classification, it has become part of the wider public discussion about obesity as it is
regularly cited by critics of the obesity epidemic to highlight the arbitrary and
constructed nature of such criteria (Oliver, 2006: 22, Bacon, 2008: 148-51). However,
as this article shows, researchers working in this field were very aware of this point.

Evidence for increases in average body weights started to appear in British

government reports: 1991 data from a DoH report gave rates of obesity as 13% for
men and 15% for women and rates of overweight as 40% for men and 29% for
women, which suggested that over 14 million UK adults were overweight and 6
million obese (West, 1994: 11). The rates found in the UK were seen to be typical
compared to those of other developed countries, including the US: ‘America has

65
NHANES III took place between 1988 and 1994.

165
similar overall rates but a higher proportion of men in the more serious grades of
obesity’ (West, 1994: 13). In fact the highest rates cited in the 1994 OHE report were
from Canada. The Health of the Nation report on obesity argued that there had been a
‘marked increase’ in the prevalence of obesity amongst UK adults in the last 15 years
using figures from four different studies, including one by the OPCS in 1984 (DoH,
1995: 5, see table 3.4 below). The authors also predicted that by the year 2005
average BMIs for both men and women would be 27.5 and obesity rates would be
18% for men and 24% for women which was similar to ‘the pattern observed in the
United States’(ibid.). The situation in the US was beginning to be framed as a warning
of future problems in Britain. This process was aided by the increasing production of
projected trends in average body weights.

Year 1980 1986/7 1991/2 1993

Men
Mean BMI 24.3 25.0 25.7 25.9
Overweight (BMI 25-30) % 33 38 42 44
Obese (BMI > 30) % 6 7 12 13
Total (BMI > 25) % 39 45 54 57
Women
Mean BMI 23.9 24.7 25.4 25.7
Overweight (BMI 25-30) % 24 24 29 32
Obese (BMI > 30) % 8 12 16 16
Total (BMI > 25) % 32 36 45 48

Table 4.5 (DoH, 1995: 5)

The authors of this report also gave projected graphs of the rate of increase in mean
BMI and secular trends in BMI for both men and women between 1980 and 2010 –
the mean BMI for both sexes was given as 28 and the prevalence of obesity for men
was forecast to be 21% and for women 27% by 2010 (DoH, 1995: 6). These show
how growing collection of data was being used to frame obesity as a problem that had
been increasing since the 1980s and would carry on increasing at the same rate, for
the next two decades.

The Scottish Intercollegiate Guidelines Network (SIGN) report on obesity used

different data because it was produced by a Scottish rather than UK body, and quoted
prevalences of 44% of adult men as overweight and 14% as obese, and 32% of adult

166
women as overweight and 17% as obese (SIGN, 1996: 8). Because this was a
document aimed at practitioners, these percentages were then translated into a
‘clinical load’: ‘in a general practice of 10,000 patients about 1,600 men can be
expected to be in need of assessment for their overweight and a further 360 need
management for their obesity. Similarly, about 1,300 women may be overweight and
a further 700 obese’ (ibid.). Prevalence figures were beginning to move into the arena
of health services planning. The authors also quoted rates from regional surveys that
varied widely, from a 2% rate of obesity among 18-23 year old men in Glasgow to a
51% rate of overweight among 45-64 year old women in Dumfries and Galloway.
However, these figures were described as illustrative since comprehensive data from
the Scottish Health Survey did not yet exist (SIGN, 1996: 9).

By the late 1990s, the authors of the British Nutrition Foundation report on obesity
could list 38 studies from 1971 to 1996 that provided data on adult body weight in the
UK. They gave the location of each study and brief details of the participants e.g.
‘10,482 male steelworkers aged 20-64 years’ (BNF, 1999: 23). In this data, BMI was
found to increase with age until 64 for both men and women and then decrease
slightly. In particular, data from the 1996 Health Survey for England showed that
‘The mean BMI of the age group 55-64 years was 27.6 for men and 27.7 for women;
this compared to a BMI of 23.4 in men and 23.5 in women, in the age group 16-24
years’ (BNF, 1999: 24). So, in agreement with earlier data, excess body weight might
have been a problem in the older age group, but was not a problem for young adults.

The existence of a shared definition of obesity was important in the collection and
analysis of this prevalence data. The early documents analysed in chapter 3 contain
frequent references to the problems of comparing data from differing surveys because
they used differing definitions of overweight and obesity. The adoption of BMI and
increasingly standard categories meant that this was less of a problem. However, the
majority of the studies cited by these authors were still relatively small scale, and
apart from the OPCS study of 1985, there appears to be no government collection of
such information in this period. It is not until the 1990s that British and European
researchers had access to large scale epidemiological data on body weights as a result
of studies such as WHO MONICA (see chapters 7 and 8). The prevalence figures
given in these reports do appear to show that rates of overweight and obesity were

167
increasing, but it is by no means obvious that, in the UK at least, they constituted a
significant public health problem.

4.8 Contemporary criticisms of the usefulness of BMI

Contemporary critics from the fields of obesity science or the wider area of chronic
disease research were rare. One of the very few critical articles published by
researchers who could be described as working in the field of obesity science was
written by Dr Paul Ernsberger, a neurologist from Cornell University and Dr Paul
Haskew, a nutritionist specialising in eating disorders from the University of
Connecticut. Their criticisms focused on the question of what the BMI actually
measured, and how well it correlated with health. The overall argument of their article
was that medical understandings of obesity and overweight represent ‘an unbalanced
view of adiposity and health’ since elevated risk factors associated with above
average weight do not translate into high mortality rates and, ultimately, ‘the net
adverse effects of adiposity are relatively modest and may be partly attributable to the
hazards of treatment ’ (Ernsberger and Haskew, 1987: 59). In particular, these authors
questioned what was measured by BMI:

Obesity, as defined by increased weight for height, is not a unitary

phenomenon. In some cases only adipose tissue is increased, while in others
increased lean tissue mass may make a major contribution to overweight. Both
BMI and overweight are highly correlated with measures of body fat (r = 0.7
for data from the NHANES survey).However, BMI is almost as highly
correlated with lean body mass (r = 0.6), or with bony chest breadth (r = 0.5).
The latter is a major determinant of the mesomorphic or muscular build. A
variety of evidence indicates that excess lean body mass or muscle is more
hazardous than excess fat (Ernsberger and Haskew, 1987: 86).

Their argument undermined the correlation between BMI and body fat, and, more
radically, argued that excess lean body mass might also be a cause of hypertension
and other cardiovascular diseases (Ernsberger and Haskew, 1987: 86 - 93). Body
shape was also important in their discussion and they concluded that ‘mesomorphy,
upper body obesity, and excess lean body mass may be interrelated conditions that
signal high-risk obesity’ and ‘individuals biologically disposed towards adiposity,
such as the rounded endomorphs, may tolerate fatness better than naturally muscular
mesomorphs’ (Ernsberger and Haskew, 1987: 91). The evidence they cited for such

168
arguments was that the Pima Indians have their lowest mortality rates at BMIs of
between 35 and 40 and the Maori population, ‘a stocky Polynesian people’ who show
no consistent relationship between BMI and mortality (ibid.) – a further unpicking of
ideas about a single relationship between obesity, body composition and mortality.

These authors argued for adaptation of the BMI cut-off points, using the J-shaped
curves of bodyweight plotted against mortality as evidence that obesity should be
defined as consisting of a BMI greater than 30 since this is the point at which ‘excess’
body weight starts to result in greater risk of illness and death. They quoted another
researcher’s conclusion that: ‘The obese appear to be a greater risk for some forms of
death yet at lesser risk for others’ (Ernsberger and Haskew, 1987: 37). Ernsberger and
Haskew accepted the cut-off point at BMI 30 as meaningful – it did demonstrate a
greater risk of some conditions – but argued that this understanding should be
balanced by an awareness of a reduced risk for other conditions (see below).

A related element of their argument was a re-framing of existing epidemiological

evidence on the relationship between mortality and weight. Data from the Seven
Countries study showed the lowest mortality rates in men who were ‘somewhat over
the average in relative weight or fatness’ (quoted in Ernsberger and Haskew, 1987: 5).
A follow-up of the study had confirmed this finding, and it was also supported by
findings from the ‘world’s largest epidemiological study to date’ (ibid.) which
collected data on 1.8 million Norwegians for 10 years. The results of the Norwegian
study showed that

Even women who weigh more than double actuarial standards, the so-called
“morbidly obese” have a better chance of surviving to retirement age than the
leanest women (BMI of 18 and below) … Current medical concepts of
“desirable weight” are thus unrelated to epidemiologic estimates of risk, but
may in fact be more closely related to current standards of attractiveness
(Ernsberger and Haskew, 1987: 6).

Ernsberger and Haskew also argued that the results of the Pooling Project (a synthesis
of data from Framingham and four other studies) showed a similar pattern of the
underweight group having the highest mortality rates and the ‘fattest’ group having a

169
slightly lower rate than that of the ‘desirable weight’66 group. They drew support for
this approach from the writing of Ancel Keys himself: ‘Commenting on this research
and similar findings of his own, Keys concluded that “underweight is a greater hazard
than overweight”’ (ibid.). The hypothesis of earlier writers, that the high mortality
rates of the underweight were due to an unknown underlying illness, was also, they
argued, not supported by the findings of several studies which showed that the
majority of individuals with unexplained weight loss died within two years, and
excluding them did not remove the excess mortality. In fact ‘exclusion of deaths in
the first five to ten years of follow-up had no effect on the increased mortality in thin
persons in several studies’ (Ernsberger and Haskew, 1987: 9).

Another confounding factor that Ernsberger and Haskew discussed was smoking:

One study links the increased mortality in thin persons to the lower body
weight of smokers…Yet moderate adiposity has been shown to protect both
smokers and nonsmokers alike in seven separate controlled studies …
Adjustment of mortality data in the Whitehall study for smoking behaviour
had no perceptible effect on the U-shaped relationship between mortality and
weight … A study of the Pima Indians, among whom smoking is rare, found
lowest mortality in persons 45 to 90 per cent above actuarial standards
(Ernsberger and Haskew, 1987: 9).

As well as arguing that mortality is lower in smokers with higher body weights, they
also argued that the Framingham data showed that it was only ‘light-to-moderate
smokers who are thinner than average, heavy smokers’ weight is similar to that of
nonsmokers’ (Ernsberger and Haskew, 1987: 10).67 The complex relationship
between smoking, bodyweight, social class - affluent smokers tended to be heavier
than average compared to poor and middle class smokers who tended to be lighter
(ibid.) – meant that it was difficult to analyse the precise nature of the relationships
between risk factors and health outcomes. Furthermore they argued that ‘Fatness may
have an ameliorating influence when it coexists with major risk factors’ (Ernsberger
and Haskew, 1987: 25), having cited studies showing lower mortality rates in fat
individuals with cancer (controlling for the confounding effects of weight loss by

66
This was defined using insurance company categories.
67
They suggested that this was because heavy smokers also tended to consume significant amount of
alcohol.

170
excluding those with pre-existing disease), hypertension and cardiovascular disease.68
Overall, evidence of the negative health consequences of obesity was not seen as
convincing by these authors, since removing studies that analysed mortality due only
to cardiovascular disease69

leaves only the Framingham study as a prospective evaluation of the role of

fatness in all-causes mortality in which the effect of obesity appeared to
become stronger with increasing follow-up. Five additional studies covering
twenty years or more do not support such a trend … Thus of the six
prospective studies of body weight and all-causes mortality covering twenty
years or more, two show a negative relationship … three show no consistent
relationship … and one study has yielded conflicting reports (Ernsberger and
Haskew, 1987: 10 - 11).

The author of a review of 16 studies analysing the relationship between bodyweight

and mortality, Reubin Andres, was also quoted arguing that these studies did not show
that obesity led to increased mortality and that ‘not only does advice on the subject of
obesity need reappraisal but research into the benefits of moderate obesity would be
worthwhile’ (quoted in Ernsberger and Haskew, 1987: 11).

Ernsberger and Haskew raised a number of serious questions about the validity of
BMI both as a proxy for obesity and as an indicator of cardiovascular risk, but they
were almost alone in raising such concerns. Other researchers within the obesity and
chronic disease communities seem to have been remarkably unperturbed by their
objections, and even those of a respected pioneer such as Ancel Keys. The use of BMI
had become firmly entrenched within the practice of these communities, and technical
criticisms such as those raised by Ernsberger and Haskew were insufficient, on their
own, to challenge such practice.

4.9 Conclusion

Between the 1970s and the 1990s, BMI became the standard means of assessing
obesity both in populations and in individuals. This occurred despite occasional

68
The second set of data is from the Framingham study and showed that ‘77 per cent of lean men
developing cardiovascular disease died from it, while only 39 per cent of the extremely fat men died
from cardiovascular disease’ (Ernsberger and Haskew, 1987: 26).
69
‘Several other studies have shown increased cardiovascular mortality but unchanged or decreased
overall mortality among heavy persons’ (Ernsberger and Haskew, 1987: 10).

171
acknowledgements throughout this period that there were important technical
shortcomings with using BMI as a proxy for obesity and as a measure of health risk.
The reason why BMI gained such acceptance despite these shortcomings was
primarily due to the utility of BMI for certain purposes. As Garrow had argued, BMI
provided a useful index for observing obesity both in populations and in individuals
since it was relatively independent of height and sensitive to changes in body weight.
The utility of BMI as a measure not just of obesity but also of health risk subsequently
received a strong endorsement when other researchers validated it with reference to
other, already accepted data, especially the Metropolitan Life data. Garrow
acknowledged that BMI was imperfect as an index of obesity and risk, but held that it
was the most practicable such index based on readily available measurements and
population data.

Garrow’s influential position within the UK obesity coalition meant that his advocacy
of BMI for this purpose was widely heeded, and a growing number of subsequent
research projects began to measure BMI, either using existing data sets or through the
relatively easy collection of new height and weight data. Moreover, this
standardisation of practice around the use of BMI - and its independence of other
population variables such as height - facilitated comparison between different studies.
Increasingly, then, BMI became embedded in practice for practical and pragmatic
reasons, both in terms of the ease of collecting the relevant data and the desirability
that new studies should be comparable with previous ones. At the same time, further
research made clear the complexity of the various factors besides BMI that could
influence risk – but, at least for the time being, this was not sufficient reason to
abandon use of the BMI as a key piece of information in the analysis of health risks
and chronic illness.

172
CHAPTER 5: THE CAUSES AND TREATMENT OF OBESITY AND
OVERWEIGHT

5.1 Introduction

Since the beginning of obesity science the main theoretical underpinning of attempts
to understand the development of overweight and obesity has been the energy balance
model. This model interpreted the human body as a homeostatic system where excess
body weight resulted from a chronic energy imbalance brought about by persistent
over-consumption and/or under-exertion. As well as explaining the causes of excess
body weight, the energy balance model also provided an explanation of the standard
treatments – to lose weight the individual had to create an ‘energy deficit’ by eating
less, being more active or increasing their metabolism with stimulants. By the late
1990s, the acknowledged difficulties of treatment had led to a new approach of
‘modest weight loss’ which stressed the health benefits of losing 5 or 10kg, rather
than aiming for a return to ‘ideal’ body weight.

The energy balance model was first outlined in the early period of obesity science,
and was carried over into later writing largely intact. This chapter will begin with an
outline of earlier approaches from the late 1960s until the mid 1980s, before sections
5.3 and 5.4 describe later accounts of the causes of overweight and obesity from the
mid 1980s to the late 1990s.

5.2 Obesity as energy imbalance: theoretical understanding of body

weight mechanisms in early obesity science

The concept of ‘energy balance’ was central to the most important model of the
mechanisms of body weight maintenance in obesity science. This model interpreted
the human body as a system attempting to achieve balance, where excess body weight
was seen to result from a chronic energy imbalance brought about by persistent over-

173
consumption and/or under-exertion. It was usually justified with reference to the first
law of thermodynamics70 and represented in the form of a diagram such as this:

Total Body Calories = Calories IN – Calories OUT

(adapted from Miller, 1969: 59)

Ideas of energy balance were crucial in the understanding of how individuals gained
weight, and also provided the explanation for particular treatments such as restricted
calorie diets and increased activity levels. In early sources it was often invoked to
counter the argument that there were different types of obesity with markedly
different causes. For example one author argued

There is, as yet, no convincing evidence that obesity is primarily a metabolic

disorder of adipose tissue considered in isolation: it is ultimately a disorder of
energy balance of the whole animal (sic) (Hanley, 1969: 34).

In this period, authors were still discussing the definition of obesity – what forms of
excess body weight could be considered instances of the same phenomenon (see
section 3.2), or which causes were relevant to their approach and which could be
ignored as too esoteric or the territory of other specialists.

Once accepted, one of the early uses of the concept of energy balance was to
differentiate between “dynamic” and “static” obesity: ‘In the first there is an energy
imbalance with progressive accumulation of fat, whereas in the second energy balance
has been restored, but the subject remains obese’ (DHSS/MRC, 1976: 30, see also
Hanley, 1969: 36). The idea of static obesity related to the concept of set point (see
below) and made sense in a context where average body weights were thought to be
rising relatively slowly. An important and related aspect of earlier versions of energy
balance theory was the idea that body weight was relatively stable. This was based on
the discovery of mechanisms controlling appetite and satiety, and an ‘appetite
regulating centre’ in the hypothalamus of both humans and animals, which if
damaged resulted in ‘voracious appetite’.

70
The principle of conservation of energy states that energy can be transformed but it can neither be
created nor destroyed.

174
 The mechanism for weight control is remarkably efficient in most people
living normal, active lives, as the intake of merely 1 per cent of excess calories
each day will lead to a weight increase of approximately 0.9 kg (2lb) in a year
and 25.4 kg (56lb) in 30 years. The reason why weight is normally kept within
such narrow limits is not fully understood and it may be due generally to an
efficient mechanism for burning excess fuel rather than to controlled food
intake (Craddock, 1969: 18).

However, large scale epidemiological study data showed that, in many individuals,
body weight was not this finely controlled – in the Framingham study, individuals
varied by an average of 5kg. Later accounts did not abandon energy balance models,
but incorporated the existence of short-term fluctuations into a more sophisticated
account of weight regulation: ‘although daily variation in energy turnover may be
considerable, there is, over a slightly longer time, a control system which matches
intake and output’ (DHSS/MRC, 1976: 28). It was now seen to occur over a longer
time period.

Factors that were seen to influence caloric intake included body weight, climate,
damage to the hypothalamus, stomach capacity, blood glucose levels, skin
temperature, body fat, exercise and food palatability. On the last point an early author
anticipated issues that were repeatedly raised in later reports:

In times of food shortage man will eat almost anything simply to satisfy his
calories needs, but in an affluent society his palate is tickled by such a wide
range of flavour and texture in attractive dishes and convenience foods it is not
surprising that he eats more than a bare minimum (Hanley, 1969: 65).

These authors did not appear to have a detailed understanding of the mechanisms of
human appetite and most argued that a lot more research was necessary: ‘There has
been much research on the factors influencing appetite in animals but little work has
been undertaken in man, and the degree to which appetite adjusts to energy needs has
not been fully explored’. Moreover ‘There is still no consensus on the degree to which
appetite in man is important in maintaining energy balance’(DHSS/MRC, 1976: 32
and 33). In these arguments, laboratory research was framed as a means of increasing
knowledge about the mechanisms of human appetite and improving clinical and
public health practice.

175
A large element of the research took the form of psychological studies investigating
differences in eating behaviour between the obese and the non-obese. However, this
research had also failed to prove that the obese ate more than ‘normal’ individuals,
although they had been found to be ‘insensitive to “internal cues” which normally
affect satiety or hunger71: ‘a normal weight subject seems less influenced by his
external circumstances and responds to some internal signs or signals which tend to
reflect the state of energy balance’ (DHSS/MRC, 1976: 34). This was part of a wider
programme of research into whether the obese were different from non-obese
individuals in significant ways: due to excess food consumption or pre-existing
metabolic, endocrinological or other physiological differences. All of the report and
textbook authors addressed some aspects of these areas of research. Craddock
discussed the metabolic effects of exercise, overfeeding (‘most are reversible’), cold
and metabolic differences between the obese and the normal weight, concluding that:

In the majority of people most metabolic differences between obese and

normal people are ones of degree only and are due to adaptation to abnormal
intake of food at some time (Craddock, 1969: 35).

Overall the consensus was that, despite the lack of available evidence, the differences
in energy metabolism between obese and lean subjects do not lead to the development
of obesity (DHSS/MRC, 1976: 47). For endocrine changes the conclusions were
broadly similar: endocrine changes in obesity were discussed in terms of adrenal,
thyroid and pituitary function and the role of insulin. Studies of overfed volunteers
showed that ‘many hormonal changes are secondary to the subjects’ increased weight
and are not necessarily an innate characteristic of obesity’ (DHSS/MRC, 1976: 53).
Changes in metabolism and hormone levels were seen as the result of excess weight
rather than the cause of it.

An alternative approach, arguing that childhood obesity led to an increased number of

fat cells – ‘hypercellularity’ – which made individuals prone to obesity later in life,
was also discussed but described as unproven (DHSS/MRC, 1976: 48). This was
similar to Craddock’s distinction between the ‘metabolic’ obese and the ‘normal’
obese:

71
Internal cue are feelings of hunger, such as a growling stomach, rather than habit or external cues
such as time of day.

176
 The clinical findings that some obese patients can lose weight easily (‘easy
losers’) while others genuinely find it difficult (‘poor losers’) is confirmed by
the two distinct groupings found, even in children, as regards the serum insulin
level and the differing response of the serum free fatty acid level to adrenaline
or glucose. It is likely that most of these ‘metabolic’ obese have been
overweight from childhood. Some obese patients therefore not only lay down
fat more efficiently than others, but also mobilise it less efficiently (Craddock,
1978: 48-9).

Such an approach contradicted the others outlined above, but in this period there were
several possible models co-existing, as researchers sought to further analyse a
complex and poorly understood phenomenon, using research from several different
kinds of research. Ideas of excess numbers of fat cells also recurred in a later
distinction made between hyperplastic and hypertrophic obesities:

Early stages of fat storage involve expansion of existing adipocytes

(hypertrophy). Later stages involve the recruitment of new adipocytes
(hyperplasia). Current evidence suggests that hyperplasia is difficult to reverse
once it has occurred, thus emphasising the need for preventative strategies
(BNF, 1999: 38).

Not only has hypercellularity now become part of the arguments for preventative
strategies, but because these processes were not seen to occur only in childhood, such
strategies have become applicable to a much wider section of the population (see
chapter 8).

John Garrow’s research was firmly rooted in the energy balance model – the title of
his first textbook was Energy Balance and Obesity in Man (Garrow, 1978) – the first
edition of which included an introduction of 38 pages to outline the concept of energy
balance and

persuade all these people [physicians, dieticians, biochemists, physiologists,

psychologists, physiotherapists] to think in terms of energy balance, and not to
suppose that there was “a cause” of obesity, like overeating, inactivity, or
some genetic disposition (Garrow, 1978: 1).

Garrow’s textbook contained detailed discussions of the factors affecting energy

intake, energy output and energy stores (Garrow, 1978). In the chapter on energy

177
intake he outlined the role of the hypothalamus, intestinal factors such as gut
hormones and endocrine factors such as insulin, referring to Bray’s72 review of the
topic (Garrow, 1978: 53 - 6). However, a large proportion of this section was taken
up by a description of the social causes of obesity:

There is no shortage of factors which can be shown to influence food intake in

one model system or another. To understand energy balance in man we need
to know how the control system is integrated, and why the human race tends
more often than most species to err in the direction of a positive balance with
consequent obesity (Garrow, 1978: 56).

He labelled this ‘the big head problem’ - the fact that human intelligence leads to
complex eating behaviour – but also discussed the social factors influencing taste, the
effects of drugs and illness, childhood influences and set point theory.

Set point theory was the idea that individuals have a standard weight to which they
are metabolically programmed to return (Garrow, 1978: 63 - 67). Garrow profoundly
disagreed with the concept of a set point since short-term control of food intake in
individuals was too erratic for a set point to exist, and moreover, the evidence of long
term weight stability had been over-stated (Garrow, 1978: 64 - 7). Whilst he accepted
that metabolic adjustment to both over and under feeding could occur, he argued this
was largely limited to weight loss and gain, and so not evidence for a set point:

nor is it necessary to postulate one to explain the observed stability of body

weight in most individuals. Conscious control of food intake and energy
expenditure is the most probable explanation for the observed oscillations in
weight about a preferred level (Garrow, 1978: 144).

Garrow’s ultimate conclusion on the subject of food intake was that:

In man the control of food intake is very complex, and the primitive
hypothalamus reflexes are buried under so many layers of conditioning,
cognitive and social factors that they are barely discernible. In the short term
the physiological control of food intake is very poor, and under laboratory test
conditions he is easily misled by false cognitive clues. Body weight in men is
not constant within 2kg, as has been stated, but usually fluctuates about 10kg

72
George Bray was a prolific and influential American researcher in the areas of body composition,
energy expenditure and the metabolic consequences of obesity. He developed the classification of
obesity by BMI discussed in section 3.4.

178
 during the adult lifespan …The factors determining these short-term (i.e. year
to year) oscillations are unknown, and virtually uninvestigated except on an
anecdotal basis (Garrow, 1978: 76).

Whilst ostensibly stressing the social factors that affect eating behaviour (which often
is how it is framed when quoted), Garrow’s focus on the metabolic processes
underlying energy balance meant that his conclusion also functioned as an argument
for further physiological research into the control of food intake and body weight.

5.3 Physiological explanations of the causes of obesity in later sources

A 1984 collection of articles edited by Philip James summarised contemporary areas

of biomedical research relevant to overweight and obesity. It included: two chapters
on obesity in animals (laboratory mice) that covered dietary and genetic factors; one
on the role of catecholamines in the regulation of thermogenesis73; one on brown
adipose tissue activity in animals and man; one on the neural and endocrine control of
energy balance, again in animals and man; one on the regulation of food intake in
obese individuals; and two on the endocrinological aspects of obesity, one of which
focused on the role of thyroid hormones (James, 1984). In addition to improving
understandings of the underlying bodily processes, such research was still conducted
with the aim of understanding in what ways the obese were different from the non-
obese. For example, in the chapter on catecholamines, the authors described the
existence of individuals with a ‘thrifty’ metabolism who do not burn excess energy as
heat and are therefore liable to obesity (Landsberg and Young, 1984: 493), while one
of the chapters on endocrinology stated that ‘there is little to indicate that obesity
depends on some underlying endocrine abnormality but much evidence to suggest that
the endocrine changes are secondary to the abnormal nutritional state of the patient’
(Jung, 1984: 609).

Two chapters in a 1987 collection of articles also discussed understandings of obesity

rooted in energy balance models. The first outlined current understandings of energy
metabolism in humans, discussing the three components of energy expenditure – basal

73
Catecholamines are hormones such as adrenaline and noradrenaline that are produced by the adrenal
glands and thermogenesis is the process of heat generation which occurs due to physical activity, diet
and some drugs, one being caffeine.

179
metabolic rate (BMR)74, thermogenesis and physical activity. The author focused on
whether obese individuals expend less energy than lean individuals:

Obesity arises when energy intake is chronically greater than energy

expenditure, which leads to a positive energy balance. It is, however, difficult
to establish whether this imbalance results from an excessive input or a
defective output of energy. In some obese individuals hyperphagia is the
obvious cause of excessive body weight, whereas in others an inability to
adapt energy expenditure to a variable intake may play a role favouring energy
gain (Jequier, 1987: 21).

As well as the practical ramifications of such issues, the author argued that the need to
reconcile low levels of food consumption reported by some obese individuals with the
basic theory of energy balance meant that more research into the regulation of energy
intake and expenditure was needed. The second chapter, meanwhile, discussed the
mechanisms of weight regulation primarily in terms of different aspects of the ‘set
point’ of body weight (see section 5.2), including temporal variation and possible
biochemical mechanisms, using findings derived from animal (mostly rat) and human
research (Apfelbaum, 1987).

Garrow’s 1988 discussion of the aetiology of obesity began with a section on the
inviolability of the laws of thermodynamics, which meant that

The search for aetiological factors in obesity can therefore be narrowed down
to factors affecting energy input or output. Do obese people have a higher
energy intake or a lower energy output than lean people? (Garrow, 1988: 102).

Referring to research into human subjects, he reviewed evidence concerning the

resting metabolic rate of individuals of different weight, and their energy expenditure,
to argue that, on average, obese individuals have higher energy expenditures than do
lean, therefore they must have a higher energy intake in order to maintain their excess
bodyweight (Garrow, 1988: 108). It had been thought previously that the obese had
lower metabolic rates than the lean – that was why they put on weight – and so this
was an important point. He also discussed research into the aetiology of human
obesity using rats and mice which focused on the role of regulation of food intake by

74 Basal metabolic rate (BMR) is defined as the amount of daily energy expended while at rest in a
neutrally temperate environment when the digestive system is inactive (which requires about twelve
hours of fasting in humans).

180
centres in the hypothalamus, the adipocyte75 hypothesis that the number of fat cells in
an adult does not increase (see above), and the role of brown adipose tissue in
thermogenesis. However, rodent models could often be misleading due to metabolic
and/or physiological differences between rodents and humans (Garrow, 1988: 109 -
110). Much knowledge about the regulation of food intake and body weight still
derived from rodent experiments, but integrating this research with data from human
experiments could be problematic because human diets were more varied than those
of laboratory rats, and human populations are more genetically diverse than those of
laboratory rats (Garrow, 1988: 111).

Garrow’s book also contained another discussion of the concept of ‘set points’
explaining that ‘individuals are said to tend to revert easily towards their set point, but
to oppose movement away from their set point by metabolic rate adaptations’
(Garrow, 1988: 114). However, Garrow argued that the data used by proponents of
this theory did not support such a hypothesis, rather it supported the idea of a
‘“buffer” control system’ which ‘tends to oppose and minimise any imposed weight
change’ (Garrow, 1988: 115). According to this logic, weight change in any direction
is guarded against, rather than specifically weight loss.

Finally, Garrow also discussed the genetics of obesity, mentioning the ‘thrifty gene’
hypothesis, but arguing that ‘it is a mistake to argue that characteristics that are shared
by a family are necessarily genetically determined’ (Garrow, 1988: 118). Using data
from several large scale studies, including one on 540 Danish adoptees and another on
metabolic rates of Pima Indians,76 Garrow argued that there is a genetic component to
obesity, but not so large that weight loss cannot be achieved:

The problem is that inheritance may be genetic or cultural, and it is difficult to

distinguish these possibilities. For example the total inheritance of percentage
body fat, and of distribution of fat, may be 0.55 and 0.61 respectively, but
some of this inheritance is due to a genetic-environmental interaction, and the
true genetic inheritance is probably about 0.22 and 0.28 respectively (Garrow,
1988: 121).

75
Adipocyte is another name for fat cells.
76
See chapter 7 for further discussion of research into the Pima Indians.

181
Garrow’s use of figures of percentage inheritance provided an additional aura of
scientific authenticity to his discussion of the thrifty gene hypothesis. However the
source of these figures was not provided. Other authors argued that Albert Stunkard’s
study of the weight of Danish adopted children and their parents showed a significant
relationship between the weight of the biological mother and the adopted child, but
that this relationship was strongest for thinness rather than fatness. Such research was
often seen to demonstrate that ‘The genes which lead to a tendency to weight gain are
thought to be insufficient on their own to cause obesity and can be overcome by
manipulating diet and lifestyle’ (West, 1994: 16). Such studies lent themselves to
diverging interpretations, as other authors framed the results of human twin and
adoption studies very differently. In a discussion of the heritability of obesity, another
author argued that ‘Genetic influences appear to account for 50-70% of the difference
in BMI in later life in both monozygotic and dizygotic twins brought up apart whereas
the childhood environment had little or no influence’ (BNF, 1999: 39). In contrast to
earlier authors, this author also argued that adoption studies showed ‘a strong
relationship between the BMI of the biological parents and the adoptee for the whole
range of body fatness… No relationship is shown, however, between the adoptive
parents and the adoptee’ (BNF, 1999: 40). However, this ‘hard line’ approach was
softened by the author’s stress on the complex aetiology of obesity. This then allowed
them to argue more plausibly that the rapidly increasing prevalence of obesity in the
UK over the last decade ‘in a genetically stable population’ confirmed the importance
of environmental causes, and the necessity for further research into gene-environment
interactions (BNF, 1999: 44).

By the 1990s, energy balance approaches were regularly drawn upon in government
policy literature. The 1995 Health of the Nation Report contained a brief discussion of
the concept of energy balance and the factors that were seen to affect it. Although
Garrow’s work was not directly cited - the studies cited came from Philip James and
his collaborators (DoH, 1995: 12) - Garrow was a member of the taskforce
responsible for the report, and the approach outlined was firmly based in the energy
balance model. Unsurprisingly, given Garrow’s chairmanship, the 1999 British
Nutrition Foundation (BNF) report contained an extended discussion of the aetiology
of obesity that was also firmly based on ideas of the role of energy balance (BNF,
1999). The results of further research into human metabolism meant that the main

182
components of energy expenditure could now be quantified - for sedentary
individuals the approximate percentages were given as basal metabolic rate 65%,
thermogenesis 10% and physical activity 25% - and the authors of the BNF report
argued that the overweight and obese have higher basal metabolic rates and total
energy expenditures because of their increased weight (BNF, 1999: 68). This was
emphasised to refute earlier arguments that individuals gain weight due to lower
metabolisms.

Physiological research had also taken place into the metabolic pathways of different
macronutrients, such as carbohydrate, fat, protein and alcohol, that gave a picture of
complex and overlapping metabolic pathways for the different ‘fuels’ burnt by the
body. The tendency to store fat in the form of excess bodyweight was a consequence
of the lack of an exact match between fat intake and its use, combined the fact that fat
was the least preferred of the body’s fuels (BNF, 1999: 71). The report’s authors saw
findings on the physiology of appetite regulation as supporting this approach.
According to their interpretation of the research, human energy balance was regulated
mostly through changes in food intake, and physiological regulation was ‘often deeply
embedded within social and environmental influences, which can readily lead to
dysregulation’ (BNF, 1999: 80). One example of this was the phenomenon of ‘high
fat hyperphagia’ or ‘passive over-consumption’ where individuals were found to eat
roughly the same amounts of food, even if the fat content was significantly increased,
leading to energy imbalance. This pattern ‘is readily reproducible in experimental
settings and helps to explain the increasing prevalence of obesity’ (ibid.) (see section
5.4 below).

An important new line of argument developed by obesity science authors in the 1990s
was that a relatively small energy imbalance, if present over a long period of time,
could explain increasing rates of obesity.

The increase in average BMI and the prevalence of obesity in the UK,
although marked, can be explained by a persistent energy imbalance,
consistently positive, but of a very modest degree. Assuming the average
height has not changed, the increase in average BMI since 1980 would be
equivalent to an average increase in weight in men and women of about 4kg
between 1980 and 1992. The energy content of excess weight amounts to
about 7000 kcal/kg (29MJ/kg). Thus, if an adult puts on 4kg, a net gain of

183
 about 28 000 kcal will have occurred. Assuming there is a smooth, even gain
in body energy and fat, this is equivalent to an excess of 6.5 kcal/day over 12
years. A 4kg weight gain is unusually large but represents a discrepancy
between average intake and output of only about 4% per day over the year
(DoH, 1995: 13)77.

This way of understanding the long term processes of weight gain became an
important element of what I have labelled environmental understandings of obesity,
since it demonstrated how individual and population weight gain can occur largely
unnoticed (see chapter 8). However, the fact that consuming as an excess of as little as
a 6.5 kcal a day was thought to lead to significant weight gain was important, since it
implied that weight gain was very easy to achieve with relatively small levels of
‘over-consumption’, and led to the question of why a greater proportion of the
population was not overweight or obese. The authors of the 1996 Scottish
Intercollegiate Guidelines Network (SIGN) report discussed processes of energy
imbalance and weight gain in very similar terms:

Body weight is regulated by powerful physiological signals which change

appetite and satiety to a far greater extent than is recognized by patients.
However, only a small (i.e. 2%) persistent discrepancy between daily intake
and energy output is required to produce progressive and substantial weight
gain. Metabolic responses to over-eating or semi-starvation play only a modest
role in buffering changes in energy balance. The metabolic rate relates to body
weight of an individual, but there are substantial differences (±20%) between
individuals … A modest but persistent accumulation of only 50-200 kcal daily
leads, over a 4-10 year period, to a slow but progressive weight increase of 2-
20 kg before the metabolic and physical cost of maintaining this additional
weight balances the additional intake. Body weight then stabilizes at this
higher level (SIGN, 1996: 5).

The figures given in this report were different – a much lower percentage discrepancy
was required than in the 1995 report, but a higher excess amount of calories was cited
– however, the authors of the Health of the Nation and the SIGN reports both stressed
the important role of small scale, chronic, over-consumption in the development of
overweight and obesity.

Another theoretical approach being incorporated into obesity science from the mid
1990s was that of ‘critical periods’ for the development of obesity. These critical

77
6.5 kcal is the energy value of 100g (or roughly 3 sticks) of celery.

184
periods were listed in one publication as foetal growth, infancy and childhood,
adolescence, reproduction (pregnancy and lactation) and middle age (especially the
menopause for women) (BNF, 1999: 45). The author admitted that the significance of
these critical periods was ‘not universally accepted’ and that evidence of
susceptibility only applied to a ‘small fraction’ of individuals. I question the practical
uses of such a list, since these critical periods could easily cover the majority of a
woman’s life, leading to the question of whether there were any non-critical periods.
However, judging by the numbers of articles cited in this discussion, studies of in
utero determinants of obesity were becoming an important area of obesity science.
Such studies had addressed the effects of foetal over-nutrition (studying both diabetic
and non-diabetic pregnancies), the effects of foetal under-nutrition (its effect on heart
disease rates from both UK epidemiological studies and data from the Dutch famine
of 1944-5), the relationship between foetal and maternal adiposity, and tracked
weight/adiposity from birth until adulthood (BNF, 1999: 45-59 (chap 7)). The last of
these studies was a continuation of the earlier studies of foetal over-nutrition on
mothers with poorly controlled diabetes, which often results in large babies. One
important source of data for this research was studies into high rates of chronic
disease among the Pima Indians of Arizona:

Pima Indians are probably the most intensively studied people with respect to
longitudinal follow-up and development of obesity. However, this population
is one in which the prevalence of both obesity and non-insulin-dependent
diabetes mellitus (NIDDM), with its consequent gestational effects, are among
the highest in the world. The findings may, therefore, be atypical (BNF, 1999:
46).

However, the evidence for the importance of these ‘critical periods’ appeared to be at
best mixed. The author concluded that maternal obesity and gestational weight gain
tended to produce bigger fatter babies, as did poorly controlled diabetes, but it was not
yet known whether these fatter babies were more likely to be obese as adults (ibid.).
However, such research on specific population groups such as the Pima was an
important source of data for later reports (see chapter 8).

The accounts discussed in this section demonstrate the growth of abstract

understandings of the causes of overweight and obesity based in basic metabolic,
endocrinological and physiological research underpinned by the energy balance

185
theory of body weight regulation. They also show the effects of earlier increases in
funding for laboratory research in these areas (Quirke and Gaudilliere, 2008) and the
many different kinds of research that were being drawn upon in the development of
obesity science.

5.4 Other explanations of the causes of obesity

This section considers another collection of explanations of the causes of overweight

and obesity, also discussed in the 1980s and 1990s, which derive less from laboratory
research into human and animal physiology, and more from other research traditions
including psychology. These causes were often considered under broad headings,
such as the role of physical inactivity, psychological factors (sometimes this included
food choice but sometimes it was addressed under a separate heading) and socio-
economic factors. The last of these categories often functioned as large and poorly-
defined catch-all collection of information, often ‘social’ in nature that was deemed
relevant to the topic.

The role of inactivity in the development of obesity, despite appearing quite obvious,
was difficult to resolve, partly due to lack of data:

It is very difficult to obtain data to show that inactivity really causes obesity,
largely because habitual physical activity is so difficult to measure …
Certainly very obese people tend to be inactive, because they have very low
exercise tolerance, and cannot manage much exertion. On the other hand the
exercise which obese people do undertake costs them more energy than the
same activity undertaken by a thin person. Furthermore many non-obese
people are also very inactive (Garrow, 1988: 121).

According to Garrow, studies had found that there was no difference in the physical
activity patterns of the majority of the obese and the non-obese, since both groups had
sedentary lifestyles, and so after describing the existing evidence as ‘sketchy’, he
concluded that inactivity may be a ‘minor contributing factor’ (ibid.) in the aetiology
of obesity. However, research into this topic carried on, despite these early
disappointing results.

186
In later writing, it is notable how often the link between physical inactivity and the
development of obesity was taken for granted, rather than demonstrated by the results
of specific studies. For example, the authors of the 1995 Health of the Nation report
quoted NHANES-1 (see section 4.5) data showing an inverse link between low
recreational physical activity and weight gain, and three other studies showing that
overweight children spend more time watching television (DoH, 1995: 16). This small
amount of data did not appear to show any evidence of cause and effect, and did not
rule out the possibility that increased television watching was a result of overweight
rather than a cause of it. Nonetheless, the authors’ conclusion was that

It seems reasonable to conclude that not only is the overall population likely to
get fatter as leisure activity declines but that individuals who are particularly
inactive are much more likely to gain weight in both childhood and adult life
(DoH, 1995: 17).

In the absence of any clear epidemiological or clinical research findings, this

statement was presumably based on everyday common sense ideas about the link
between declining levels of physical activity and increased weight. Researchers also
sought to counter the concern that increased levels of physical activity would lead to
increased appetite. Thus the BNF report of 1999 declared that there was ‘no evidence
to suggest that exercise results in more food than would be expected in order to
compensate for the extra expenditure’ and that ‘the evidence is stronger that exercise
may work to improve feeding regulation to better match energy expenditure’ (BNF,
1999: 131). This demonstrates the persistence of earlier worries about the counter-
productive effects of increased physical activity.

More complex thinking about the possible social causes of obesity was notably
absent. Garrow’s 1988 textbook on obesity was typical in this regard, devoting only a
small proportion of his chapter on the causes of obesity to the social, cultural,
psychological and cognitive factors that affected food intake. This section included a
brief discussion of Jeffrey Sobal and Albert Stunkard’s 1989 article78 outlining the
inverse relationship between obesity and social class found in developed countries

78
The Sobal and Stunkard article reviewed 144 studies of the relationship between the incidence of
obesity and socio-economic class, and was an important source and that had been regularly cited in
obesity science writing since its publication (see also section 1.4).

187
which contrasted with the direct relationship found in less affluent countries such as
India and Germany (sic), but Garrow concluded

There is no obvious explanation for the increased prevalence of obesity in the

poorer sections of many communities, nor is it clear if the poverty causes the
obesity, or the obesity causes the poverty (Garrow, 1988: 126).

He also referred to Suzy Orbach’s (1982) argument that obesity and other eating
disorders in women resulted from a rejection of the norms of feminine behaviour
around appearance, sexual relationships and domesticity. Somewhat missing the
point, Garrow suggested that ‘obesity also occurs in men, presumably for unfeminist
reasons, so some obesity in women must occur for unfeminist reasons also’ (Garrow,
1988: 128). This was, however, a rare reference to feminist criticisms of obesity
science.

The author of the 1994 OHE report also used Sobal and Stunkard’s article in his
discussion of socio-economic and psychological factors in the development of
obesity. Sobal and Stunkard had concluded that there was a strong inverse
relationship between socio-economic class and obesity for women in affluent
countries; a weaker inverse relationship for children and men in affluent countries;
and a strong positive relationship between socio-economic class and obesity for
women, children and men in developing countries (Sobal and Stunkard, 1989: 269).
They thought that there were three ways of explaining this inverse relationship in
developed societies – obesity could affect socio-economic status, socio-economic
status could affect obesity, or they could both be affected by a third factor, such as
heredity (Sobal and Stunkard, 1989). The causality was so uncertain, that the author
of the 1994 report did not feel able to draw a definite conclusion:

The fact that lower socioeconomic status tends to be associated with a higher
prevalence of obesity in developed countries could be taken to give more
credence to an environmental rather than inherited explanation for the cause of
obesity although both factors may be important and the direction of causation
may be reversed (West, 1994: 17).

Discussions of psychological factors were still seen as an important factor in the

development of obesity. In his 1988 textbook, Garrow discussed evidence that obese

188
individuals eat in response to social and emotional pressures, but noted that the non-
obese do so also. He also mentioned studies of the link between obesity and binge-
eating, but suggested that binge-eating could be a consequence of dietary restriction
rather than a cause of obesity in its own right. Ultimately, he argued that research has
not been able to identify any psychological disorder that was a defining feature of the
obese (Garrow, 1988: 129). He also argued that the existence of weighing devices
such as bathroom scales, and the sale of diet books, showed that the majority of
people take action to restore or maintain their weight within a ‘desirable’ range:

If this is so then cognitive factors explain why many people maintain energy
balance rather accurately over long periods, although they fluctuate in weight
by many kilograms in the short term (Garrow, 1988: 132).

His conclusion was that many people are overweight or obese because they ‘do not
know, or care’ and so need education by professionals in the realities of their situation
(ibid.).

Psychological factors remained an important explanation of the cause of obesity and

overweight. Other authors in the 1990s discussed the concept of ‘reactive eating’,
eating as a response to psychological distress (West, 1994: 18), or ‘externality theory’
which suggested that obese individuals were more responsive than average to food
cues (food accessibility or attractiveness) and, therefore, eat more when food is
readily available. This last theory had become less fashionable by the mid 1990s, but
new evidence on the link between ‘externality’ in eating behaviour and BMI
subsequently led to calls for more research in this area (BNF, 1999: 84 - 5). The
concept of ‘dietary restraint’- the idea that there is a link between dieting and
abnormalities in the eating behaviour of the obese - was also discussed during the
1990s, but the British Nutrition Foundation report of 1999 noted that there was
‘controversy’ about the direction of cause and effect, and thus about whether restraint
led to eating disorders or was a method for maintaining a healthy body weight (BNF,
1999: 85 - 6). In a similar fashion, the report noted that conditions such as binge
eating disorder, emotional eating, food addiction and night eating syndrome tended to
be defined as psychological disturbances associated with obesity rather than as the
causes of it. However, the report concluded that:

189
 In general it has been difficult to separate cause from effect in the relation
between obesity and emotional disorders. A current view is that personality
and emotional factors play only a minor role in the aetiology of obesity, but
may be important in relation to responses to treatment (BNF, 1999: 90).

Insofar as the report discussed the behavioural and social aspects of eating, it referred
only to the externality theory and to ideas about the relationship between the sensory
characteristics of food (‘palatability’) and the food preferences of the obese and the
non-obese. The report argued that obesity and dietary restraint were understood to be
linked to increased responsiveness to particular foods, and so ‘Palatable, high fat,
energy-dense foods may, therefore, provoke behavioural problems of weight control’
(BNF, 1999: 114). The report’s author, thus, reduced eating to an issue of individual
choice that was most appropriately studied and treated by the discipline of
psychology.

At the same time, the composition of the diet was also seen as an important cause of
overeating. Whilst acknowledging the ‘complex interplay of physiological, social and
psychological factors influencing appetite, satiety and over-eating’ (DoH, 1995: 14),
the reports of the 1990s tended to identify the fat content of the diet as a key factor,
since a high fat diet was thought to encourage increased energy consumption. The
results of one study suggested that ‘it is easier to overeat fat than carbohydrate and
that there are fewer compensatory mechanisms when fat is used to boost energy
intakes’ (DoH, 1995: 14). Evidence for the role of artificial sweeteners in the
reduction of energy intake was conflicting, and for alcohol the evidence on the
relationship between intake and BMI was inconclusive (DoH, 1995: 15). Such
research was part of a wider investigation in the different metabolic pathways of
different foodstuffs (see section 5.3).

In the late 1990s discussions on dietary factors considered epidemiological evidence

for secular trends in energy intake, such as the UK National Food Study. This study
showed that, since the 1970s, there had been a decline in the overall per capita energy
intake, and a shift in composition of the diet as the proportion of carbohydrate
declined whilst that of fat had increased significantly (BNF, 1999: 92 - 3). Amongst
other data sources, the author of the BNF report describes results of studies such as
Seven Countries and MONICA (see section 2.5) as providing ‘modest, but

190
inconclusive, evidence that high-fat/low carbohydrate diets favour the development of
obesity’ (BNF, 1999: 100). This was one way in which the enormous quantity of data
generated by such research was incorporated into this body of knowledge (see also
section 7.5).

Overall, discussions still stressed the complex causes of excess body weight, regularly
quoting Garrow’s summary: ‘the aetiology of obesity in man has genetic, social,
cultural and psychological components in different proportions in different people’
(West, 1994: 18). Similar kinds of statements remained common throughout the
1990s:

Recent increases in the prevalence of obesity show the importance of

environmental factors in determining the onset of this condition. Many factors
interact to induce weight gain, including behavioural, physiological, genetic,
medical, therapeutic and psychological processes. To identify a single factor
as the cause of obesity in a patient oversimplifies a complex process (SIGN,
1996: 5).

However, it is notable that many of the reports published at this time also tended to
bracket out wider, what I have labelled ‘environmental’, causes of obesity and
overweight, and give greater weight to individual or personal factors. Thus the SIGN
report’s discussion of particular factors conducive to weight gain focused on
individual factors including pregnancy, medical and therapeutic causes (endocrine
disorders and drug treatments) and genetic disposition (SIGN, 1996: 6-7). It was only
in accounts of the behavioural causes of excess body weight that environmental
factors were referred to, often obliquely, when the authors’ described the role of
physical inactivity, snacking, energy dense and highly processed diets and alcohol
consumption (SIGN, 1996: 6). But, these practices were described as if they were
largely the product of individual choices, rather than socially stratified differences in,
for example, access to particular types of food or leisure activities.

This focus on individual behaviour was retained in the policy recommendations that
were made in some reports published in the 1990s. For example, the Department of
Health’s Nutrition and Physical Activity Taskforces recommended a reduction in the
average percentage of food energy derived from fat and an increase in average levels
of physical activity. When these guidelines were translated into specific

191
recommendations they became uncontroversial suggestions to improve consumer
knowledge through information and education, make low fat products more widely
available and develop voluntary national guidelines for caterers (DoH, 1995: 19 - 20).
The recommendations on physical activity became even more modest suggestions to
encourage individuals who take no exercise to aim for one period of 30 minutes of
moderate activity a week, and to increase both the number of individuals who take
moderate activity 5 times a week and those who vigorously exercise (DoH, 1995: 20).
The only other significant suggestion was ‘a mass media campaign to improve
awareness amongst professionals and the public of the benefits of physical activity –
particularly moderate activity’ (ibid.). During this period the approach of increasing
physical energy expenditure in order to lose weight was no longer routinely described
in terms of ‘exercise’ or ‘physical exercise’. These terms carried connotations of
drilling, training and the need for specialist equipment and facilities. In the 1990s,
such behaviour was re-labelled ‘physical activity’, and its non-specialist, cumulative
and mundane character was stressed. Activities such as walking, cycling, gardening
and housework, that fitted this new definition, were also given an important role in
preventative strategies (see chapter 8).

The relevant section of the BNF report began with an acknowledgement of the
complexity of the causes of obesity:

Obesity is a complex syndrome with multifactorial origins. Its aetiology can

range from the purely molecular (e.g. Prader-Willi and other obesity
syndromes) to the purely social (e.g. Sumo wrestlers). Most cases probably
cluster towards the middle of this spectrum and can best be described as the
result of an adverse ‘obesogenic’ environment working on a susceptible
genotype (BNF, 1999: 37).

This shows the developing idea of environmental causes of obesity, and was the first
reference to the concept of the ‘obesogenic’ environment in these documents. Wider
environmental factors such as food industry marketing policies, government policies,
the structure of food retailing, and changes in food composition (which refers to the
use of artificial sweeteners and lower fat products) were discussed in this report. A
very careful framing of food preference in the first half of the chapter (arguing that
palatability has no effect on long term food choice) then allowed the authors to
conclude that:

192
 There is no strong evidence to suggest that the marketing policies of the food
industry directly impact an individual’s predisposition to obesity … The main
impact of the EU’s Common Agricultural Policy on the UK diet has been via
altering the relative price of major food product groups and has been broadly
positive from a nutritional point of view (BNF, 1999: 114).

Presumably, the involvement of several representatives of major food manufacturing

companies and government departments like MAFF had some influence on the
formulation of these conclusions.

From the mid 1990s, obesity science authors start to provide more general
explanations of the causes of obesity and overweight that began to include arguments
for possible environmental causes, such as increasing consumption of processed foods
and increasingly sedentary lifestyles. However, the policy recommendations, that
authors also provided, largely still addressed individual behaviour, rather than wider
‘environmental’ causes.

5.5 Conventional treatments for obesity and overweight

As I described above, Garrow’s reformulation and promotion of BMI-based

definitions of overweight and obesity meant they had become the standard definitions
used within obesity science (see section 4.3). The embedding of these BMI-based
categories into clinical practice in turn contributed to their consolidation. For
example, Philip James used Garrow’s standard definition of obesity as a method of
categorising potential patients, as shown by table 5.1 (below).

Grade of obesity Body mass index Management strategy

0 20 – 24.9 Patient ‘normal’ assess basis for anxiety (sic)
I 25 – 29.9 Community based group therapy
II 30 – 40 Individual attention, detailed management
III > 40 Special measures e.g. surgery

Table 5.1 (James, 1984: 636)

James used this classification to outline what he saw as appropriate treatments for
different groups of patients. Proposed treatment for patients with grade 1 obesity

193
involved advice to follow a low fat, high carbohydrate diet, attend a group such as
Weight Watchers and have their weight monitored;79 treatment for patients with grade
2 obesity involved the use of very low calorie diets based on liquid formulas,
behaviour modification focusing on understanding the individual’s eating habits and
anorectic drugs such as fenfluramine; treatment options for patients with grade 3
obesity included those for grade 2 obesity plus jaw-wiring and abdominal surgery
such as by-pass surgery or stomach stapling (James, 1984: 644 - 57). Unsurprisingly,
given the time of writing, this list was not substantially different from the treatments
recommended during the 1960s and ’70s, as outlined in chapter 3. Garrow also
categorised patients on the basis of BMI into grades of obesity and gave a breakdown
of appropriate treatment strategies that I have reproduced in table 5.2.

Treatment strategy Grade III Grade II Grade I Grade 0

Diet
Starvation NO NO NO NO
very low calorie Possible NO NO NO
conventional YES (1)* YES (1) YES (1) NO
Milk YES (2) YES (2) Possible NO
jaw wiring/cord YES (3) Possible NO NO
exclusion surgery Possible NO NO NO
Drugs
Anorectic Possible Possible NO NO
thermogenic NO NO NO NO
Physical training NO Possible YES YES
Reassurance NO NO Possible YES

* these numbers refer to preferred treatment choices within each grade of obesity
Table 5.2 (Garrow, 1988: 185)

These treatment options were similar to the ones outlined by James. Garrow’s basic
argument was that the more overweight a patient, the greater the number of treatment
options that should be considered: ‘everything should be done to encourage the Grade
III obese person to tackle the problem sooner rather than later, since with time the
health risks will become greater and weight loss will become more difficult to
achieve’ (Garrow, 1988: 192). He ruled out starvation and thermogenic drugs for all

79
This programme was extended by a further refinement based on the presence of a series of additional
risk factors such as smoking, high blood pressure, high blood cholesterol, family history of CHD and
other CVDs and central distribution of fat in patients with grade 1 obesity that led to more specific and
tailored dietary and activity changes (James, 1984: 643).

194
of his patients, and described exercise as ‘impractical and possibly dangerous’ for this
group of patients. Treatment for this group, therefore, should rely heavily on different
kinds of dietary regimes. For grade 2 patients, Garrow saw specialist techniques such
as jaw-wiring as a last resort, whereas he now recommended physical exercise. He
stressed that this class of obesity should still be taken seriously:

It is the job of the therapist to guide and encourage the patient, but with grade
II or III obesity it is very rarely ethical to console the patient by saying that
weight loss is not very important, so not to worry too much (Garrow, 1988:
202).

For Garrow, the selection of treatment options for grade 1 obese patients presented
different problems than those of the other grades. Firstly, the large number of
potential patients meant that treatment could not be offered on a one-to-one basis.
Secondly some GPs resisted treating obesity as part of general practice due to the
ineffectiveness of the available methods of treatment (see section 5.6). Garrow did not
accept this argument, suggesting that it was easy to monitor a patient’s weight: ‘The
doctor who does not keep an accurate record of the patient’s weight change every 2-4
weeks cannot claim to be treating obesity seriously, and has no right to complain if
the results are poor’ (Garrow, 1988: 208). He thought that, despite their lack of
consistent success, one solution to both these problems was the use of slimming clubs:

a proper integration of non-profitmaking slimming clubs with professional

dieticians as group leaders, backed up by a hospital service for special cases, is
a practical and economical way of providing care for the large numbers of
moderately obese people in the community (Garrow, 1988: 212).

Garrow saw this as more economical for the health service and, in terms of lost work
hours (although not money), the patient. However, these would be slimming clubs run
with the approval of medical and nutrition science, and not solely for profit. Other
authors also framed the treatment of this category of patients as a largely non-medical
activity:

For those who are overweight (BMI 25 – 30) medical treatment is unlikely to
be necessary, but it is important to prevent further weight gain and a
conventional diet may be used in an attempt to reduce the BMI to 20 – 25
(West, 1994: 19).

195
For grade 1 obesity, the main treatment should be dietary, and Garrow gave an
example of a commonly recommended 1000 kcal a day diet in the appendix. It was
based on eating three meals a day and strict limits on the consumption of sugar, fat,
sweetened drinks, alcohol, cereals and baked goods, some fruit and pulse and
‘slimming foods’. Finally, his discussion of the treatment of grade 0 patients – he
quoted a reference to ‘thin fat people’ – clearly refers to people who would now be
described as suffering from disorders such as compulsive eating, bulimia and anorexia
(Garrow, 1988: 217 - 20).80 Although he recognised that the majority of these patients
were female, and that compulsive eating was one common reaction to prolonged
dietary restriction, Garrow explicitly denied that a link existed between weight
reduction and eating disorders (Garrow, 1988: 219). Eating disorders were the
professional concern of psychology not obesity science (see section 5.4).

The list of treatment options proposed by obesity science researchers throughout this
period still included dietary methods, drug treatment and behaviour therapy and self-
help groups. Surgical techniques were discussed more often than in early obesity
science, but still seen as an uncommon option, a last resort for seriously obese patients
(who had tried all other techniques unsuccessfully). There were many different forms
of dietary treatment – including calorie-counted, carbohydrate restricted, fat-restricted
and high-protein ones (Garrow, 1987: 111-4). Dietary restriction was still considered
as ‘the mainstay treatment for all grades of obesity’ and could be supplemented at low
cost by minimal intervention strategies e.g. verbal instructions from a GP on how to
lose weight (West, 1994: 20).

Due to its intended audience of general practioners, the SIGN report focused largely
on the treatment of obesity and overweight within primary healthcare: ‘Given the very
large number of overweight and obese patients in Scotland in need of medical care, it
is inappropriate for obesity to be managed in a hospital setting on a routine basis’
(SIGN, 1996: 16). The treatment methods that the SIGN report considered appropriate
for primary health care included exercise, behavioural advice and weight loss diets.
The report argued that support from the patient’s family, health professional or
community was an important additional component of such approaches (SIGN, 1996:

80
Presumably these were the majority of grade 0 individuals that hospital practitioners, rather than
GPs, would encounter.

196
23). Exercise combined with diet was seen by the authors of this report as more
effective than either method in isolation for promoting fat loss, and therefore schemes
to increase physical activity should be included in all weight management schemes
(SIGN, 1996: 25). They divided the population into four categories depending on their
activity levels (sedentary, irregular moderate, regular moderate and regular vigorous
activity) and distinguished between moderate intensity activities (brisk walking,
climbing stairs, heavy DIY and spring cleaning) and vigorous and intense activities
(squash, brisk hill walking, running and heavy lifting). Their aim was to provide
encourage individuals to increase their levels of physical activity: ‘to convert inactive
children and adults to a pattern of “active living” where they are on their feet for 4
hours daily’ (SIGN, 1996: 67). Research had found that home-based, informal and
unsupervised activities were most successful, and so the author suggested that 30
minutes of brisk walking every day should be promoted for the purposes of weight
maintenance (ibid.).

The weight loss diets recommended in this report reflected orthodox nutritional
advice to increase consumption of unrefined starchy carbohydrates and fruit and
vegetables, and reduce consumption of fat, sugar and salt (SIGN, 1996: 57 - 59). Such
‘moderate deficit diets’ of 1200 or 1300 calories a day produced weight losses of half
of a kilogram a week for 12 to 24 weeks: ‘The average weight loss after one year is
6% of initial weight, but this depends on a support system and follow up to limit
weight regain’ (SIGN, 1996: 24). The report noted that very low-calorie diets (800
calories a day) had been shown to produce rapid weight loss in the short-term, but not
result in long-term weight maintenance (SIGN, 1996: 25). This was part of an
increasing emphasis on the importance of weight stability, as opposed to repeated
attempts at weight loss followed by rapid regaining, which came to be labelled as
‘weight cycling’ in future reports (see chapter 8).81

Specific dietary treatments varied between different reports. The treatments

recommended in the BNF report of 1999 were low-calorie diets (800 – 1500 kcal per
day), very low calories diets (less that 600 kcal per day) and the milk diet (1.5 to 2
litre a day plus nutritional supplements, equivalent to 800kcal) (BNF, 1999: 151-7).

81
Ernsberger and Haskew were the earliest of my sources to discuss this link arguing that ‘Temporary
weight loss may actually promote further accumulation of weight in the long run’ (1987: 38).

197
Dietary treatments had been heavily researched and so the report’s chapter on the
topic contained discussions about the factors affecting compliance with particular
diets and the role of psychological support and behaviour modification, and the
authors addressed the argument that dietary treatments might do more harm than
good. Having refuted arguments that dieting inevitably led to further weight gain or
was linked to the development of eating disorders, the authors argued that the role of
such diets should be shifted from a short-term approach of ‘going on a diet’ to a much
longer-term one of ‘weight management’.

Any treatment programmes for obesity should address, not only the problem
of weight reduction but should also include measures to help with the
maintenance of lowered weight…. The ability of a treatment to maintain long-
term weight reduction is as important as its ability to cause the initial weight
loss (BNF, 1999: 145).

This would involve permanent changes in eating habits in order to lose weight and
maintain this lower weight for the foreseeable future (BNF, 1999: 164). This framing
was linked to a growing stress on the prevention of weight gain that occurred in this
period (see chapter 8). The BNF report’s chapter on physical activity also outlined a
very similar approach to that of the SIGN report where:

Physical activity should be included as an essential part of obesity treatment,

with patients supported in their efforts to steadily build a daily routine
involving moderately intense exercise, such as brisk walking, more incidental
or opportunistic movement and physical work and reduced time in sedentary
pursuits (BNF, 1999: 174).

Again, the report saw this process as long-term: the overweight and obese should be
encouraged to make permanent changes in their pattern of physical activity. The
report also included discussion of the environmental changes that could be made in
order to increase levels of activity among the population – suggestions include ‘traffic
free’ (i.e. car-less) routes to schools and shopping areas; greater access to parks and
sport facilities; rewards for journeys undertaken on foot or by bike rather than by car;
and alternatives to lifts and escalators (BNF, 1999: 175). This was part of the growing
discussion of the environmental causes of overweight and obesity (see section 5.4).

198
By the 1980s, advocates of pharmaceutical treatments for obesity had proposed a
number of different types of potential drug treatments, including biguanides which
were related to diabetic treatments, bulking agents to compensate for lack of dietary
fibre, thyroid hormones to increase metabolic rate, and anorectic drugs (relatives of
amphetamine that suppressed appetite) (Simpson et al., 1987: 118-21). Obesity
scientists were cautious about recommending pharmaceutical interventions, however.
One influential survey of treatment options, for instance, recommended that only the
anorectic drugs, such as diethylpropion, phentermine and fenfluramine, should be
considered as treatments, and noted that most of them had significant side effects.
Given that weight gain was normal after drug therapy, the same survey argued that the
use of anorectics ‘can only be justified if there is a clearly defined, short term need for
weight loss’ (Simpson et al., 1987: 123).82. But by the 1990s, some of these drugs had
been withdrawn from the market due to their side effects and dexfenfluramine was the
most widely used pharmaceutical treatment, although SSRIs such as fluoxetine were
awaiting approval by the US Food and Drug Administration (West, 1994: 24).

In all of the above discussions, drug treatment was considered suitable only for those
patients who had already attempted and failed to lose weight using the more orthodox
methods of diet and exercise. However, the BNF report is more specific about the
criterion:

It may be appropriate to consider drug therapy for those patients with a BMI
of 30 or greater who have failed to lose this amount of weight [10%], or whose
weight is no longer decreasing after at least 3 months of structured dietary
management (BNF, 1999: 182).

Given the low levels of success reported using dietary treatments, this formulation
appears to allow for the prescription of such drugs to the maximum number of obese
patients. The drugs recommended for such treatment fell into three categories –
depending on whether they act on the serotoningenic pathways (fenfluramine and
dexfenfluramine), catecholamine pathways (phentermine) or serotoningenic and
noradrenergic pathways (sibutramine). Drugs previously prescribed, such as diuretics,
amphetamines and thyroxine, were described as not appropriate in the treatment of

82
This usually meant as preparation for elective surgery, but non-medical reasons such as being able to
enlist in the armed services could also count.

199
obesity (BNF, 1999: 183). Drug treatment was described as short-term – initially for
three months and only extended if the patient has achieved a cumulative 10% weight
loss. However, the report warned that

The absence of extensive safety data means that the use of anti-obesity drugs
must be very closely monitored, with patients subject to very regular medical
review (BNF, 1999: 186).

The withdrawal of fen-phen (a combination of fenfluramine and phentermine) in the

1990s after patients developed serious cardiac problems suggests that this injunction
was not followed closely enough by some practitioners (Gard and Wright, 2005: 162).

Meanwhile, because patients found it hard to maintain weight loss using diets or
drugs, making both techniques effective only in the short-term (if at all), some
researchers explored other approaches in the hope of obtaining better results. These
additional approaches included behavioural therapy and self-help groups. Albert
Stunkard used the results of two clinical trials to argue that

behaviour therapy can help obese patients to lose moderate amounts of weight
and to maintain these weight losses for at least one year. The better
maintenance of weight loss makes behaviour therapy appear superior to the
other treatment modality for obesity – pharmacotherapy (Stunkard, 1987:
128).

As a psychologist (who conducted one of these trials) rather than a pharmacologist,

Stunkard would be expected to argue along these lines. In his research, behaviour
therapy consisted of self-monitoring so the patient kept precise records of what they
ate, where, when and why they ate; control of the stimulus preceding eating by
planning what to eat and minimising opportunistic eating; control of the act of eating
using techniques such as chewing thoroughly, leaving food on the plate and not doing
anything else whilst eating; reinforcement of prescribed behaviours with praise from
friends and family and material rewards; cognitive restructuring to counteract
negative monologues; nutrition education; and physical activity (Stunkard, 1987:
139). Whilst it might have been successful, it was also a demanding and time
consuming process.

200
By the 1990s, behaviour therapy was seen by obesity science researchers as an
important technique because it re-educated individuals’ eating habits and allowed the
maintenance of a new lower weight, and also because it was cheaper than other forms
of treatment as it relied on group sessions with a dietician, rather than individual
treatment by a doctor (West, 1994: 29). The tendency to recommend the use of
behavioural techniques was boosted by research findings showing increased success
with long term dietary modification using such techniques (SIGN, 1996: 25). All of
the reports published in the mid to late 1990s period recommended some combination
of the following practices: patient education about food labelling and preparation:
self-monitoring and recording of eating behaviour; and understanding of the
psychological and habitual elements involved in food consumption. Examples of the
last category included eating as a form of self-medication and how to deal with
situations that may give rise to ‘inappropriate’ food choices such as festivals or eating
out (SIGN, 1996: 66). It was recommended that these kinds of techniques were used
in group situations by health professionals such as dieticians, rather than by doctors.

These authors’ accounts of behavioural treatments argued that obese individuals do

not demonstrate an ‘eating style’ different from the non-obese. Despite this
recognition, they still maintained that cognitive behaviour therapy (CBT) - self-
monitoring, stimulus control and modifying self-defeating thought processes - had a
significant role to play in individuals’ attempts to lose weight. The initial success of
such approaches had, however, led to unrealistic expectations concerning their long-
term effectiveness.

There was initially optimism that the weight loss achieved by CBT, although
usually modest, would be long lasting, as it would permanently alter the
patient’s attitude to food. In fact, longer-term studies show that the affect of
CBT alone disappears by five years after therapy (BNF, 1999: 181).

Research then focused on the possibility of using CBT to prevent weight regain,
whilst recommending restrictive diets and exercise to achieve greater initial weight
loss (ibid.). This pattern of treatment ‘hype’ followed by lowered expectations and
finally incorporation into routine weight management practices, was similar to that
which occurred when new pharmaceutical treatments were being promoted.

201
A related approach, which provided another kind of social support in the weight
process, was the self-help group. However, Audrey Eyton, the founder of Weight
Watchers, argued that the label self-help was not entirely appropriate.

Although group therapy under a peer who serves as a role model is an

essential part of the treatment, it appears to work well only when the leader
follows a precisely structured policy. In many cases this policy has been
developed by a highly professional commercial organization using all the
facilities of modern market research, professional medical, nutritional and
behavioural consultation (Eyton, 1987: 140).

We can read Eyton here as staking a claim to professional expertise in the area of
obesity treatment whilst being careful to acknowledge that of other, more prestigious
professional groups, and to counteract the negative image of the weight loss industry,
particularly among medical practitioners, as unscrupulously profiting from people’s
desire to lose weight.

The final form of treatment considered was surgery. In the 1980s, a description of
different types of surgical methods was labelled ‘unusual methods of obesity control’
(Gries, 1987: 147), and, in the early 1990s, techniques were still limited to different
kinds of stomach stapling and jaw wiring (West, 1994: 25 - 27). By the late 1990s,
new methods were being developed but the basic distinction was still between mal-
absorption techniques such as jejuno-ileal or gastric bypasses, and techniques to
restrict intake which include jaw wiring and stomach stapling or banding (BNF, 1999:
187-9). The long term effects of such surgery was also beginning to be studied in the
Swedish Obese Subjects (SOS) study which monitored the health of approximately
2000 individuals who had undergone such surgery, compared to a similar number
who underwent non-surgical treatment (BNF, 1999: 189). The BNF report
summarised the findings of this research:

In severely obese Swedish patients, gastric surgery causes massive weight

loss, maximal about one year after operation, and this greatly improves the
health and quality of life compared with weight-stable obese controls. Weight
loss of this magnitude, attained by non-surgical methods, confers similar
benefits…However, few patients achieve this level of weight loss, even with
the assistance of drugs (BNF, 1999: 190).

202
Because of the complex nature of these operations, the BNF report argued that they
should be performed by experienced surgeons in specialist centres, on patients who
had been fully informed of the risks involved, and, even if successful, ‘lifelong
monitoring is required post-operatively’ (ibid.). Such surgery was acknowledged, in
this report, to have a significant risk of complications, including death, and to leave
the patient at life long risk of side effects, that included involuntary regurgitation and
malnutrition. This was one of the few acknowledgements of the hazards of weight
loss treatments in this literature, although the side effects of drugs, especially those
that had been withdrawn from the market, were occasionally discussed (OHE, 1969:
24-5, RCP, 1983: 48-9). On the other hand, Ernsberger and Haskell discussed this
topic in detail:

Because virtually all weight reduction methods are only temporarily effective,
they can be hazardous indirectly due to the regaining phase. However, many
of the most popular weight-control techniques have direct hazards also
(Ernsberger and Haskew, 1987: 40).

Ernsberger and Haskell gave a chronological list of ‘hazardous treatments for

obesity’, including a detailed description of the negative health effects of each
treatment. This began with an early version of the low carbohydrate diet (from 1862),
included the use of laxatives (from the 1920s), amphetamines (from 1937), total
fasting as an in-patient (from 1959), intestinal bypass (from 1969) and jaw wiring
(from 1974). Their conclusion was that

In the light of the relatively mild and uncertain hazards associated with
adiposity, and the potential contribution of iatrogenic disease to these hazards,
the use of dangerous weight control measures appears to be highly
questionable. We urge that hazardous treatments for obesity…be replaced with
nutritional counselling and exercise training, supported by behaviour
modification (Ernsberger and Haskew, 1987: 48).

Ironically, this critique was preceded by advice to adopt a diet low in saturated fat,
cholesterol, salt, sugar and alcohol, along with increased physical activity to ‘promote
gradual and permanent, albeit modest, loss of weight’ (ibid.), that was very similar to
that given by the practitioners whose treatments they had just heavily criticised.

203
5.6 The difficulties of successful treatment and a new approach of
‘modest’ weight loss

The difficulty of successful treatment had been an ongoing subject of discussion in

obesity science. From early on, it had been acknowledged that all the recommended
treatments had relatively low success rates: ‘it is evident obesity is difficult to treat’
(DHSS/MRC, 1976: 63). Philip James subtitled his 1984 chapter on treatment ‘The
Constraints to Success’ (James, 1984: 635) which suggests he did not have great
confidence in the measures he was about to describe. On the success of dietary
treatment, Garrow stated that:

A jeremiad on the inefficacy of dietary treatment of obesity is often quoted:

‘Most people will not stay in treatment for obesity. Of those who stay in
treatment most will not lose weight, and for those who lose weight most will
regain it’ (Stunkard 1972). There is some truth in this aphorism, especially
when the results of treatment are followed for several years. To some extent
the prophecy of failure is a self-fulfilling one: if both patient and therapist
expect to fail they will … The poor results for dietary treatment of obesity
apply only to outpatient regimens: virtually any amount of weight can be lost
by dietary methods given time and the complete control of the diet in a
metabolic ward (Garrow, 1987: 115).

Garrow’s parallel between unsuccessful attempts to lose weight and in-patient

treatment does not adequately address the complex issues involved in long-term
failure rates of dietary treatments (see chapter 6). Later in the same discussion, he
argued that ‘The vast majority of people who lose weight successfully do so by
following diets they have read or heard about, and never appear in any statistics’
(ibid.); ultimately, he implied, success depended on the motivation of the patient
rather than medical treatment.

The 1995 Health of the Nation report argued that the targets for reducing the
prevalence of obesity can be met in three different ways: by helping the obese to lose
weight, by helping the overweight avoid becoming obese and by preventing the whole
population becoming fatter. The report recommended focusing on the last two since,
as Garrow had argued, obesity treatment in adults was ‘limited in its success’ (DoH,
1995: 19). In this respect, the Health of the Nation report reflected a wider shift
toward prevention on the part of government departments and professional bodies

204
involved in the planning of health care services (see section 3.2), one facet of which
was an acceptance of the limited possibilities of treatment. A similar
acknowledgement occurred in the 1995 WHO report:

Because a large proportion of the adult population in industrialized societies

will be overweight or obese and because weight loss therapy is ineffective
unless closely supervised and followed up, not all overweight or obese
individuals will qualify for intervention. Priorities [sic] should be given to
those at highest risk, with the primary focus on reducing the risk profile rather
than on weight loss per se (WHO, 1995: 329).

Coming as it did from the WHO, this frank acknowledgement of the difficulties of
weight loss therapy builds was framed primarily for an audience of readers with
responsibility for planning health care and developing appropriate policies.

The difficulty of identifying successful treatment options also led to research into
other treatment approaches. In the 1990s, the potential benefits of ‘relatively modest
weight loss’ were first discussed. A 1992 study by Goldstein demonstrated that an
individual with a BMI of 40 who lost 5% of his body weight (reducing his BMI to 38)
would reduce his mortality risk by 12%, although the report’s author was also careful
to say that greater weight loss would lead to increased benefits, so as not to undercut
the argument for achieving ‘desirable weights’ which were still one of the two main
aims of treatment (West, 1994: 19).83 Other reports also focused on the health benefits
of ‘modest’ weight loss, usually defined as between 5 and 10 kilograms:

A 10kg weight loss results in many benefits e.g. >20% fall in mortality, >30%
fall in diabetes related deaths, >40% fall in obesity-related cancer deaths …
[cites decreases in blood pressure, fasting glucose and blood cholesterol
levels] … Modest weight reductions of only 5-10 kg improve back and joint
pain, lung function, breathlessness, and reduce the frequency of sleep apnoea
(SIGN, 1996: 12).

A weight loss of between 5-10% of the initial body weight is associated with
clinically useful improvements in terms of blood pressure, plasma cholesterol,
triacylgyclerols and HDL cholesterol, and a significant improvement in
diabetic control (BNF, 1999: 150).

83
The other was weight stability (West, 1994: 20).

205
The growing importance assigned to weight stability also led to arguments for the
benefits of ‘modest’ weight loss: ‘a small sustained weight loss is preferable to large
weight losses followed by weight regain’ (West, 1994: 31). This argument built on
data about the rarity of significant long-term weight loss from dieting and exercise
alone, and anticipated later discussions of the dangers of repeated attempts at weight
loss followed by rapid regaining which came to be labelled as ‘weight cycling’ in
future reports (see chapter 8).

Combined with this new emphasis on modest weight loss was a more radical re-
definition of the goal of treatment. This involved a shift from a focus on weight loss
as an end in itself, the health benefits of which were assumed but not measured, to one
that focused on demonstrable improvements in health risks.

Overweight patients frequently assume, as do some health personnel, that the

principal goal should be a return to an ideal weight. Yet, in medical terms,
great health gains can be achieved if a patient’s smoking, dietary and exercise
habits are improved to reduce risk factors without weight reduction (SIGN,
1996: 20).

The authors of the 1995 WHO report (see chapter 7) shared this new approach:
‘Weight loss is recommended but weight loss per se should not be the primary target
of intervention’ (WHO, 1995: 329). The reasoning behind this change was that:

Until recently, the response to intervention was usually assessed in terms of

attaining “ideal body weight” or reducing body weight to below a certain BMI
cut-off or “percentage over ideal body weight”. It is no longer clear that such
goals are optimal. For some people they are unrealistic: in most obese
individuals they imply large sustained weight losses, which few are able to
achieve unless they are enrolled in long-term programmes with extensive
follow-up (e.g.> 5 years). Moreover, substantial improvements in risk factor
profiles have been documented in obese individuals who lost only moderate
amounts of weight and would still be classified as overweight or obese…
Finally, there is no evidence that large weight losses either have beneficial
effects or reduce mortality, and the more extreme diets needed to produce
large weight losses may increase the likelihood of relapse (WHO, 1995: 330).

These claims built on the approach of ‘modest weight loss’, but went much further by
suggesting that there was no evidence for benefits to health or improvements in
mortality rates from large weight losses.

206
5.7 Conclusion

In this chapter I have outlined developments in biomedical understandings of the

causes of overweight and obesity as they were represented in largely British policy
documents of the 1990s. Basic metabolic, endocrinological and physiological research
continued to use the concept of energy balance as foundational, while the concept of
set point for body weight was widely discussed at times. There was a new focus
within this policy-oriented literature on small-scale chronic energy imbalances
combined with low levels of physical activity leading to overweight and obesity.
Obesity science accounts of the causes of obesity initially focused on the individual
level, trying to understand the psychology of over-eating. Complementary accounts
considered the social level, where the authors discussed the relationship between body
weight and socio-economic class, dietary composition, food choice and physical
activity. The idea of physical activity was adapted to include activities such as
housework, climbing stairs and DIY which had not generally been included within the
scope of “exercise”. From the mid 1990s, these authors began to increase the range of
causes they considered as the environmental or structural causes of overweight and
obesity began to be discussed and the term ‘obesogenic’ environment was introduced.

A version of the BMI classification discussed in chapter 4 was widely used as the
basis of policy recommendations about appropriate individual treatment for
overweight and obesity. The recommended treatments in primary health care
remained weight loss diets combined with exercise and, if possible, behavioural
therapies. Supplementary pharmaceutical treatments might be offered to patients who
did not initially manage to lose weight. In specialist clinics, it was recommended that
surgery should be offered to the minority of patients whose weight was seen to pose
such health risks that it outweighed the risks of such procedures. However, the stated
aim of treatment was beginning to change towards to a much longer-term approach
that focused on changing eating habits, improving levels of physical activity, and on
the health benefits of ‘modest’ weight loss and subsequent stabilisation, rather than a
return to ‘average’ or ‘ideal’ weights. Such an approach was at least partly justified by
greater acceptance of the difficulties most individuals found in maintaining long-term

207
weight loss. Importantly, accompanying this was a shift from a policy focus on weight
loss per se, to a focus on demonstrable improvements in risk status.

208
CHAPTER 6: ANALYSING OBESITY SCIENCE

6.1 Introduction

In the previous two chapters, I have outlined the development of British obesity
science in the 1980s and 1990s; the way that a range of scientific findings were
incorporated and represented in a series of policy-oriented documents. I have
described how research into the physiology of appetite and digestion was combined
with data generated by large scale epidemiological studies to create an argument
about the increased mortality associated with excess body weight, and with clinical
knowledge to generate recommendations about treatment options. This body of
knowledge about the definition, causes, effects, treatment and, eventually, prevention
of obesity and overweight was a hybrid entity, assembled from heterogeneous
sources, often as part of government-appointed committees of expert medical
professionals. The group of researchers and clinicians who formed a policy coalition
around obesity and overweight in the late 1960s had become increasing successful in
persuading governments to listen to their message. In the process, they created a body
of policy-relevant science that framed excess body weight as a public health problem
using standard measurement and models.

In the judgements about health that are a crucial part of obesity science there was an
assumption that thinness equalled healthiness. Many critics have questioned this
assumption (Ernsberger and Haskew, 1987, Bacon, 2008, Campos, 2004, Gaesser,
2002, Gard and Wright, 2005, Monaghan, 2005, Oliver, 2006). Those who come from
a background in physical activity often point out that this approach privileges body
size and shape over other indicators such as cardio-respiratory fitness, blood sugar
levels or cholesterol levels (Gard and Wright, 2005). Research shows that such
indicators can often be improved more easily than body weight (SIGN, 1996), and
studies into the effects of physical activity on health and mortality have been
conducted since the 1950s (Morris and Heady, 1953, Morris et al., 1966, Morris et al.,
1973). However, research and policy focusing on body size and shape seems to have
become more widely accepted than that focusing on physical activity and fitness,
certainly until the end of the 1990s. This may be because of the widespread
acceptance of dieting for weight loss within British and American popular culture

209
throughout the second half of the twentieth century (Schwartz, 1986, Stearns, 2002).
It may also be connected with the ‘obviousness’ of body weight as a sign of ill-health.
Without denying the complexities of agreeing standards relating health to body
weight (see below), it is possible to suggest that a visual inspection of an individual’s
body seemed to be a much more straightforward procedure than assessing their
physical fitness.

In the preceding chapters I have charted the development of two important aspects of
the framing of body weight as a public health problem within obesity science: the
body mass index and the energy balance model of weight regulation. In the present
chapter, I will consider the important advantages of the BMI for obesity science. In
particular, I will argue that without such a standardised measure, it would be difficult
to talk meaningfully about an epidemic of obesity, whether or not such terminology is
actually justified by the existing epidemiological data.

Then, I will switch focus to analyse the model of the body used in obesity science. In
common with much of biomedicine, obesity science used a mechanistic understanding
of human physiology which sees obesity and overweight as the result of a positive
energy imbalance due to long-term consumption of more calories than are expended.
The simplicity of this model allowed it to be shared by many different research areas,
but also meant that it could not account for the low success rates for all conventional
weight treatments. Explanations of eating and eating behaviour considered by obesity
science were dominated by individualised accounts largely based on psychology or
economics. This focus on the individual left very little room for more social or
structural explanations and therefore gave a very impoverished account of individual
eating behaviour and changing patterns of consumption.

6.2 The advantages of BMI for obesity science

STS researchers have written on the wider social effects of standards and
classification systems (Bowker and Leigh Star, 1999). Critics of obesity science have
also written extensively on the problems of using BMI-based definitions of
overweight and obesity (Ernsberger and Haskew, 1987, Bacon, 2008, Campos, 2004,
Gaesser, 2002, Gard and Wright, 2005, Monaghan, 2005, Oliver, 2006). I am not

210
going to add to these arguments because they are well-rehearsed, and because those
working in the area are aware of the problems with the BMI (WHO, 2000). As I
argued in section 4.3, public health statisticians were also highly aware of the effects
their redefinitions of overweight and obesity had on the numbers of populations
fulfilling these criteria (Kuczmarski and Flegal, 2000: 1078). Instead I want to
examine the advantages of the BMI for obesity science, to explain why, in a relatively
short period of time, it became the field’s standard measurement and definition of the
problem of excess body weight.

In chapter 4, I described how obesity and overweight were initially defined using a
variety of classificatory schemes and indices. The most common of these was the
Metropolitan Life Company’s table of ideal weight for height categorised by frame
sizes (often the 1959 version), or another definition based on a set percentage above
average weight (usually 10 or 20%). Other indices such as W/H, W/H³ (the ponderal
or Rohrer index) and W = H-100 (the Broca index)84 were also used. There seem to
have been disciplinary and regional patterns in these usages: W/H³ was used in
paediatrics, while epidemiological researchers in mainland Europe used the Broca
Index until relatively recently (Oddy et al., 2009). One account also states that the
Broca index was the basis of the categories in American insurance companies’ height
and weight tables (Oddy et al., 2009: 7). Often authors used a definition based on the
Metropolitan Life ideal weights and referred to other indices as they incorporated
results from different studies into their research. These varying definitions and
measurements caused several problems for epidemiology researchers. Firstly, varying
definitions and measures (often in different units) made it hard to compare the results
of different studies. Secondly, using percentage deviation from average body weight
like the ideal weight tables meant that, if relative proportions stayed the same, then
the same number of people would be defined as overweight or obese, even if the
whole population gained significant amounts of weight. Such a relative index can be
useful when the measured attribute is stable in the population, but an absolute
measure (which does not vary with changing prevalence) would be more useful in a

84
W = H-100 is the Broca index, and is a rough means of calculating ideal weight; the + 15% sets the
threshold for obesity as 15% above this ideal weight.

211
context where the prevalence within a population is changing, especially if the
rationale of such an index is to track this change.

As Ian Hacking describes, W/H² was first developed by the Belgian astronomer and
statistician, Adolphe Quetelet in the early nineteenth century (Hacking, 2006c).
Quetelet was interested in anthropometry and he developed W/H² as one way of
illustrating differences between populations; it was part of his concept of the
‘l’homme moyen’ or average man (Hacking, 1990). The association with Quetelet
was well known in obesity science and regularly referred to in textbooks. However,
unlike these conventional accounts, Hacking tracks these changing meanings of ‘the
fraction’ i.e. W/H²:

We should notice how the changes, in the ways that this fraction have been
used, reflect a profound shift in interest. We began quantifying the human
body and its excellences, and now regard it as a threat, an ever-shifting
kaleidoscope of risk factors (Hacking, 2006c:1).

BMI, in its modern form, developed out of 1960s chronic disease epidemiology, but,
according to Hacking, one of its first important uses was in a Norwegian survey
collecting information about the relationship between levels of underweight and
tuberculosis (Hacking, 2006c: 19). Only subsequently did it come to be used primarily
to identify and monitor over-weight. This change of use parallels the shift from
research into under-nutrition to research into over-nutrition described in section 3.3
(also see below). As I described in chapter 4, the name body mass index was
suggested in 1972 by Ancel Keys (Keys et al., 1972), and gradually adopted, despite
an alternative attempt to call it the Quetelet Index (Garrow, 1981: 3). Keys, Garrow
and George Bray, who developed the first version of the cut-off points to define
overweight and obesity, were central figures in the development of obesity science, as
well as chronic disease epidemiology, and their recommendations carried
considerable authority among obesity researchers.

Like many other indices of bodily measurement - such as blood pressure

(Timmermann, 2006) - BMI functioned as both the measurement of a bodily attribute
(weight related to height) and also the definition of a condition (obesity/overweight).
This dual role makes an analysis of its function within obesity science crucial, since it

212
simultaneously measured and defined the modern problem of excess body weight.
However, until the late 1990s, the combination of varying indices of measurement
with varying definitions of overweight or obesity led to a large number of competing
definitions – one article lists 18 different classifications that were used in the US
between 1942 and 2000 (Kuczmarski and Flegal, 2000: 1076) and a wider search
would certainly locate more examples.

The BMI-based definitions of overweight and obesity consist of two elements – the
BMI index which I have just described, and the cut-off points which are used to
define the levels at which body weight was considered to be excessive and a danger to
health. Even in the 1980s and 1990s when the BMI had become widely used, the
specific cut-off points used varied between studies. For example, an NIH consensus
conference in 1985 decided on a cut-off point for obesity as BMI 27.8 for men and
27.3 for women (NIH Panel on the Health Implications of Obesity, 1985),85 whereas
most British research was then using overweight BMI 25 and obesity BMI 30 for
both sexes. This lack of standardisation of BMI cut-off points was part of a much
wider set of disagreements about how to define obesity, which includes not just BMI-
based measures but also percentages above average body weight (hence the NHANES
definitions) or increases in mortality risk (hence the Keys/Bray/Garrow definitions).
The review article mentioned above, which was a retrospective account of these
changes, also demonstrates some the work involved in trying to develop a consensus
in this area, as the authors created an impression of continuity by stressing the
similarities between the NHANES definitions to the older Metropolitan Life
definitions and the newer lower numerical limits (Kuczmarski and Flegal, 2000:
1077)

As Hacking also points out, one advantage of the BMI was that it was based on
readily available data that is relatively cheap to collect, particularly compared to
measuring body fat which is difficult and technically complicated and therefore very
expensive (see section 4.2 ).

85
These were the sex-specific 85th percentile of the BMI distributions in NHANES II for the age group
20 to 29 and were very close to 20% above the 1983 Metropolitan Life medium frame desirable
weights (Kuzmarkski and Flegal, 2000: 1076).

213
 Adiposity, the ratio of body fat to body mass, is the important health indicator,
but is fairly expensive to measure by any current technique – and thus
comparable to a personal DNA readout. But the BMI is very cheap: stand on a
scale, stand under a device that measures height, press two buttons on a
calculator (or use one of the innumerable online BMI calculators), and there
you have your BMI … A national study of adiposity would have been more
informative and cost about a million times more (Hacking, 2006a: 88).

The development of the BMI was an example of epidemiologists using readily

available, and therefore cheap, data to create large data sets. In additions, because
height and weight have been routinely collected in many areas of medical research,
including epidemiology, existing data sets can be easily converted into this new index.
Much of the data from large studies, such as Framingham, appears to be been re-
worked using BMI cut-off points from the 1980s onwards. The adoption of this
simple numerical index allowed epidemiologists and other researchers to directly
compare the results of different research studies, and thus investigate, describe and
quantify trends in changes in population weight between regions, countries and across
time.

The development of classification schemes such as Bray’s and Garrow’s also included
an increasingly large percentage of the population – in one of Garrow’s schemes those
of BMI 18.5 to 25 were no longer simply labelled ‘normal’ or ‘ideal’ weight but
‘grade 0 obesity’ (Garrow, 1988: 3). The distinction between “normal/ideal” and
overweight is one of kind, whereas the distinction between grade 0 and grade 1obesity
is one of degree, and implies a continuum or gradation from low to high risk. As such,
it underwrites the idea that even those not “at risk” could be at risk of becoming at
risk. This was an enormous expansion in the jurisdiction of obesity science, but it
fitted well with the logic of the risk factor approach and the preventive medicine
associated with it, which emphasised individual responsibility for minimising health
risks by lifestyle changes such as healthy eating and increased activity levels
(Petersen and Lupton, 1996).

6.3 BMI and the epidemicity of obesity

As I described in chapter 4, many of the early reports argue from a very limited
evidence base that obesity was an important public health problem. Until the 1980s,

214
the most frequently cited empirical sources were the Metropolitan Life height and
weight tables, the 1959 Build and Blood Pressure Study, and results from the Seven
Countries study and Framingham, both of which were studies of risk factors for heart
disease, rather than of body weight. This meant that, especially for British authors,
there was a limited amount of large-scale data available to use in their arguments.
Until the 1960s and 1970s, governments did not routinely collect statistics about the
physical stature of their populations: ‘In the middle of the twentieth century few
countries could make general statements about trends in weight gain by their
populations.’ (Oddy et al., 2009: 225). The subsequent growth in collection of this
information was a twentieth century continuation of the processes of nineteenth
century government information collection (Desrosieres, 1998, Hacking, 1990), as
well as a product of the increasing use of statistics within medicine (Rothstein, 2003).

Because of this lack of data, British authors made heavy use of American sources86
combined with the limited number of small-scale British studies (which had often
been carried out as part of various occupational health investigations). Whilst I argue
that the claims of obesity science were developed on the basis of limited evidence, I
am not putting forward a conspiratorial understanding of these processes. Claims-
making is a normal aspect of biomedical research and clinical practice. Researchers
and clinicians need to secure funding by demonstrating the importance of the health
issue on which they work, and this involves using the available evidence. Such
claims-making is a routine aspect of the activities of such coalitions and their attempts
to influence the development of public health policy. However, in the case of obesity
science, the timing of these efforts was significant. As I described in section 3.3,
medical concern focused on under-nutrition in the immediate aftermath of the Second
World War. In particular, protein malnutrition was seen as a pressing problem from
the late 1940s and throughout the 1950s. However, the focus of medical concern
seems to have switched, relatively rapidly, to a concern with over-nutrition at some
point in the late 1950s or early 1960s. The timing of this transition varied between
different countries (Oddy et al., 2009: 224-5), but seems to have occurred in many
European countries between the late 1950s and late 1960s. This was a rapid change

86
The Build and Blood Pressure Study, like the 1959 Metropolitan height and weight tables, was still
being cited in the 1980s and 1990s.

215
since the aftermath of the Second World War had meant that, in the late 1940s and
early 1950s, there were still food shortages in many European countries.

Despite the speed of this transition, obesity science had become well-established by
the 1980s. Judging by the increasing institutional scope of their reports, in the 1980s
and 1990s, the arguments of the public health coalition that had formed around
obesity were becoming increasingly accepted. More results were being generated by
large-scale epidemiological studies, and the WHO MONICA project was set up as a
European version of American risk factor research. Results from NHANES II (see
section 3.5), showing a big increase in the prevalence of obesity and overweight, were
released in the mid 1990s (Kuczmarski et al., 1994). These figures were reported in
the popular press and obesity started to be described as an epidemic in the medical
and popular press (Pi-Sunyer quoted in Pringle, 1994). This new framing of obesity
was largely made possible by the gradual adoption and standardising of the BMI
definitions. With an easily handled numerical measure such as BMI, it was much
simpler to construct prevalence rates for overweight and obesity, and, therefore,
monitor secular changes within populations, and the differences between populations,.

One further aspect of defining and measuring obesity and overweight using this
classification was that it facilitated the diagrammatic representation of increasing rates
of overweight and obesity. The spread of infectious diseases had often been
represented by mapping the geographical spread of numerical increases in prevalence
rates. In the late 1990s this method was applied to increasing rates of obesity and
overweight in a famous series of slides produced by the Centers for Disease Control
(CDC).87 This animation shows an increasing number of the American states going
from blue to red over a period of 15 years, as average rates of obesity increased to
levels above 30% (CDC, 2010). It has been argued that this memorable and widely
publicised animation – first produced in 1999, but still readily available online – was
one of the most important factors in establishing the ‘fact’ of obesity as an epidemic
(Oliver, 2006: 40-3). It is a particularly vivid illustration of a more general co-
production (Jasanoff, 2005) of, on the one hand, the BMI as measurement and
definition of obesity, and on the other hand the ‘epidemicity’ of obesity. Without the

87
This uses Behavioral Risk Factor Surveillance System Survey Data which seems to be equivalent to
or incorporate the NHANES data.

216
BMI to provide a simple way of turning those data into a stark representation of
obesity rates, it would have been much harder to argue credibly that obesity was
increasing to the extent it justified the label of an ‘epidemic’.

6.4 Limits of mechanistic models

As I outlined above, contemporary obesity science relies on a mechanistic model of

human physiological processes of appetite and weight regulation. The metaphor of the
human body as a machine has a long history within in physiological research
(Rabinbach, 1990) and has been developed in various ways depending on the specific
organ or processes under investigation. In the context of obesity science, the
mechanistic model takes the form of the energy balance model, and one of the most
influential British accounts was in the work of John Garrow (see sections 4.4 and 4.5).
As I outlined in chapter 5, obesity scientists generally considered that weight stability
occured when the energy consumed by an individual was equivalent to their
expenditure of energy. Obesity and overweight arose from an energy imbalance due
to excess consumption and insufficient energy expenditure. This framing was
powerful because it tallied with common sense and, sometimes, everyday experience.
Authors also argued that it was based on the first law of thermodynamics (Miller,
1969: 58-9, Jequier, 1987: 17),88 thereby invoking the authority of physics and adding
to the difficulty of questioning its accuracy as a model of individual weight
regulation.

In analysing the problems of the energy balance model, I want to follow Gard and
Wright (2005) and others, who argue that it is a necessary but not sufficient account
of the processes of appetite and body weight regulation. Not only does it not consider
the social aspects of these activities (see section 6.5 below) but its simplicity and
standardised form does not allow for the consideration of variation either between
individuals or specific population groups. Apart from the individual variation in
susceptibility to weight gain that is apparent from everyday experience, two examples
of such variability are age differences and sex differences. These are acknowledged to
result in differences in metabolism, and average BMI is widely acknowledged to

88
This states that matter can neither be created nor destroyed.

217
increase from the twenties until old age when weight loss can occur due to muscle
wasting. However, this variability was ignored when it came to making
recommendations about how to combat obesity. When BMI cut-off points were
initially proposed there was discussion about allowing these limits to increase with
age but this was ultimately decided against, partly on the grounds that weight stability
was best for health (Garrow, 1978: 37, RCP, 1983: 5-6). Women were also
understood to be particularly prone to weight gain under the critical periods model
(see section 5.3) and physiologically different from men, both in their levels of body
fat and the health risks that it gave rise to.89 However, none of these differences were
incorporated into the cut-off point classifications of overweight and obesity. The
resulting inflexibility of these standards, combined with the idea that body weight is
largely under individual control, does not allow for either the propensity to gain
weight with pregnancy and increasing age, or the increasing difficulty that older and
sick individuals may find in following weight loss advice.

The prevalence of the energy balance model in obesity science derived from the
significant overlap between laboratory research into the physiology of appetite and
body weight regulation on the one hand, and clinical research into the development of
treatments for overweight and obese patients on the other. Physiological researchers
need models to study complex systems such as human metabolism, and even with
such models the complexity of the systems governing appetite and body weight was
daunting. Obesity science combined this laboratory and clinical research with the
findings of large-scale epidemiological studies. The simplicity of the energy balance
model meant that it provided a common basis for interpreting all these diverse kinds
of results.

As STS scholars have argued (Fujimura, 1992), research practices need to be

standardised in order to be reproducible. The processes of creating simplified models
and common sets of practice lead to the creation of ideal standards – as in the example
of the BMI cut-off points discussed above. In the case of body weight regulation the
studies that formed the basis of these standards were often done on relatively young,

89
The data that was initially used to argue for the use of BMI came from the Seven Countries study
whose subjects were all men. The differences between men and women are referred to in the quotes in
my discussion on body shape in section 3.4, but this is an area that needs more research.

218
fit men (such as students or prisoners), and so their physiological capacities became
seen as standard. The World War II Minnesota Starvation Experiment, supervised by
Ancel Keys, took place in a laboratory setting over six months and involved 36 white
male conscientious objectors aged 25 to 35, many of whom belonged to religious
groups such as Quakers or Mennonites (Kalm and Semba, 2005). Although about
under- rather than over-feeding, this experiment was canonical and was regularly
cited in obesity science writing. Ethan Sim’s experiments in over-feeding Vermont
prisoners were less well-known but also regularly referred to, and described in a
couple of popular accounts of obesity science (Pool, 2001: 70, Shell, 2003: 79-81).
However, patients seeking treatment for weight problems were not usually young, fit
men – they were often middle aged or older, female and with other long-term health
problems. The discrepancy between research subjects and patients is problematic
because patients are unlikely to achieve the weight loss that laboratory research
subjects appear to be capable of, and, because of its extreme simplicity, the energy
balance model does not acknowledge the differences between these groups. The
energy balance model provides a rationale for assuming that findings produced in
young men can be generalised to other groups, but it cannot deal with variability
between or within groups - even amongst the younger male research subjects studied
there was a wide variation in individual’s capacity to gain and lose weight in the
controlled conditions of a laboratory or a prison.

A final problem with this use of the energy balance model was that when it was
transferred into a clinical context it was used with, and served to underwrite, the
assumption that an individual’s body weight was under their control. To an extent this
is obviously the case – cessation of eating usually leads to weight loss – but the extent
of individuals’ control over their weight was typically exaggerated in discussions
about weight loss. There are two facets to this exaggeration. First was a lack of
consideration of the social aspects of eating that I address in section 6.6 (below).
Secondly, insufficient recognition was given to the limits of individuals’ abilities to
undertake levels of physical activity that will have an impact on body weight.
Voluntary activity only accounts for about a quarter of an individual’s overall energy
use (BNF, 1999: 61), although this amount can be increased by intense physical
activity. Although writers such as Garrow recognised that physical activity could not
be used as a treatment for obesity and overweight, they increasingly stressed its role

219
in prevention. Again, young fit men will be more able to increase their energy
expenditure to a point where it will affect their body weight, but older individuals
with other health conditions are unlikely to be physically capable of achieving the
necessary levels of activity. The energy balance model helps to reinforce this
perspective, by directing attention solely to the relationship between energy input and
output, rather than to the social and other complexities surrounding both those
processes.

Such input/output models exclude any consideration of regulatory mechanisms, but

homeostatic models presuppose regulation around a set point. Considering the body as
a homeostatic system hinges on the idea that there is a ‘normal’ state to which the
body attempts to return. The failure of research to produce evidence of a set point (see
chapter 5) suggests that this may not be the case. Karen Throsby and Celia Roberts
contrast such homeostatic models with Ann Fausto-Sterling’s model of the body as an
allostatic system (Throsby and Roberts, 2010). An allostatic model is one that
‘regularly changes activity to meet anticipated demands’ rather than a homeostatic
one that is ‘constantly trying to achieve a “normal” state’ (Throsby and Roberts, 2010:
87). Fausto-Sterling develops this framework in her discussion of the phenomenon of
essential hypertension90 found in black populations in the US. She links the higher
blood pressures of these individuals to the chronic stress of living in a discriminatory
environment, arguing that ‘different life experience activates physiological processes
common to all, but less provoked in some’ (Fausto-Sterling, 2004: 26). In a similar
fashion, ‘normal’ weight can be seen as the outcome of complex interactions between
an external environment, personal medical history, and habitual behaviour. Such an
approach provides a better explanation of the fact that so many people find it so
difficult to lose significant amounts of weight over the long term. It would also allow
health to be understood as cumulative, as the result of ongoing interactions that may
profoundly alter the body’s capacities to respond. This would fit with life course
approaches within the sociology of health inequalities which study patterns of
accumulation of health advantage and disadvantage in individuals’ biographies
(Bartley, 2004).

90
Essential hypertension is defined as consistently raised blood pressure in the absence of an obvious
physiological cause such as a constriction in a major blood vessel.

220
6.5 Confounding and similarity judgements

“Confounding” is currently defined as ‘the error in the estimate of the measure of

association between a specific risk factor and disease outcome which arises when
there are differences in the comparison populations other than the risk factor under
study’ (Bhopal, 2002: 79). The term has been used in this sense in epidemiology since
the late 1950s (Vandenbroucke, 2004: 317). In this section I will consider
confounding sociologically as the result of a series of judgements that are made about
the causes of disease. Confounding is important in epidemiology because it leads to
mistaken inferences about the causes of disease. Epidemiologists endeavour to avoid
this mis-attribution of causality by adopting appropriate research design and
appropriate methods of statistical analysis. However, the description of discussions
about the relationships between bodyweight, social class and smoking outlined in
section 4.5 above, demonstrates that such analyses are the outcomes of complex
judgements about behaviour and disease aetiology that are both biomedical and
sociological.

Although epidemiology involves some laboratory research, it is primarily a field

science in which researchers have limited control over causality. In the case of
coronary heart disease, large-scale longitudinal epidemiological studies were initially
set up to study risk factors for heart disease which were known to be varied and to
operate over long timescales. However, such research gave rise to particularly
difficult problems of confounding because of the number of potential variables
involved and the complex interactions that could take place between them. This may
be one of the reasons why discussions of confounding start to appear in
epidemiological literature in the late 1950s; another was the controversy around the
Doll-Hill study on the association between smoking and lung cancer (Berridge,
2003b, Vandenbroucke, 2004: 318). Decisions about whether confounding has taken
place are essentially based on similarity judgements about whether one population is
sufficiently similar to another in particular respects for the relevant outcomes to be
considered comparable. To put it another way: researchers must consider in what
ways particular populations differ, and whether these differences matter for the
findings of their research. Smoking behaviour, for instance, needs to be considered in
discussions of the relationship between body size and health, not only because it is a

221
cause of heart disease and other chronic diseases, but also because smokers have been
found to be lighter, on average, than non-smokers.

Establishing the presence of confounding factors requires a sophisticated

understanding of the health and behaviour of the populations being studied. For
example, judgements of whether smoking was a confounding factor in the relationship
between body weight and increased mortality had to take into account not only the
fact that smokers tend to be both lighter and unhealthier than non-smokers, but also
the social patterning of smoking behaviour, which has changed significantly in the
last fifty years, and the use of smoking as a weight control technique by certain
groups (usually young women). Because of this complexity such judgements are
fragile and potentially unstable. They can be easily undermined by new information
casting doubt on the similarity judgement: if more affluent smokers are actually
heavier on average (Ernsberger and Haskew, 1987: 10) this alters the relationship
between smoking and body weight on which early judgements of causality and
confounding relied. It also requires a more complex analysis of the relationship
between smoking, health and body weight since the effects of income level have to be
incorporated.

This analysis demonstrates that confounding is important, sociologically, because it

requires complex judgements about the relationship between health and structural
factors such as gender, socio-economic class and age. It shows that researchers
working in such areas were capable of dealing with such ‘social’ factors in a
sophisticated manner, but that they chose to do this in a restricted fashion, in certain
explanations of disease causality only.

6.6 The dominance of individual explanations: a further ‘writing out’ of

the social

Individualism is an important element of post-war public health in developed

countries (see sections 1.2.1 and 2.6). It marks a well-defended boundary between
public health, on the one hand, and politics or social policy on the other. Claims to
expertise in the area of diet and bodyweight have significant social and political
implications, attract a lot of scrutiny and, therefore, require a great of deal boundary-

222
work to maintain the credibility of the people or institutions making them (see chapter
9). Individual behaviour has historically been seen as coming under medical authority,
whereas the attribution of responsibility for social problems has often been more
contested. The development of social medicine in Britain in the 1940s and 1950s was
an attempt was made to shift the boundary between public health and social policy,
but for a variety of reasons it was unsuccessful, and the work of social medicine
became split between public health and medical sociology (see section 1.2.1). The
growth of an individualist perspective in post- war public health can be attributed to
the influx of hospital doctors and laboratory scientists into this field, both of whom
were more inclined to adopt an individualist rather than a sociological or political
view of illness.

Because of this individualism, the only explanations of eating behaviour that are
considered at length in obesity science writing were based on either psychology or
economics. The presence of psychological explanations derived from historical
overlaps between obesity research and the treatment of eating disorders, whereas the
prestige of economics and its closeness to public policy has meant it was regularly
drawn upon in health policy writing. Individualistic accounts of behaviour, based on
these two disciplines, are typical of policy literature (Steward, 2010), but they leave
little room for more social and structural explanations of behaviours such as food
consumption habits. Consequently, insofar as the obesity policy literature of the 1980s
and 1990s makes any mention of social explanations of eating, these are typically
cursory and tokenistic: brief references to outdated anthropological studies of
‘fattening huts’ for instance (WHO, 1995: 319), or to preference for larger body size
in previous centuries or contemporary under-developed societies (West, 1994: 9-10,
see also WHO, 2000: 127-8) that say more about the authors’ attitudes than about
contemporary eating behaviour. Often in the obesity science of that period, the
‘social’ functioned as a residual category, where seemingly relevant but currently
unexplainable information was assembled and briefly outlined.

By contrast, recent sociological accounts of food and eating stress that such behaviour
is strongly related to social norms around family conduct, gender roles, national,
regional and ethnic identity and age (Atkins and Bowler, 2001, Caplan, 1996, Mennell

223
et al., 1992, Murcott, 1998, Warde, 1997). As Delormier and her co-authors argue in
their summary of the field:

The most important limitation of studying eating strictly as a behaviour under

the control of an individual is that it exaggerates the extent to which rational
choice drives what people choose to eat, and underestimates the extent to
which eating is embedded in the flow of day to day life. People’s eating
patterns form in relation to other people, alongside everyday activities that
take place in family groups, work and school. Eating does involve isolated
choice, but it is choice conditioned by the context in which it occurs
(Delormier et al., 2009: 217).

Classic examples of the effect of social norms on eating behaviour include the link
between personal identity and religious practices of food exclusion, or gender norms
around appropriate food for women and the effects of their role as family meal
providers and preparers. The operation of these norms combines with more prosaic
factors such as individual preference, economic resources and time constraints.
Merely considering psychological attitudes to food and issues of food availability and
price does not address this complexity in a nuanced enough fashion to understand
why, despite the pressure of negative health messages about excess body weight and
the necessity to lose weight, many individuals are unable to adapt their diet over the
long term.

The rational choice model used in framing such dietary decision-making does not
allow for the possibility that such pattern of eating may require choices that are not
easy to make in particular social contexts. Elizabeth Shove discusses the same
approach, which she labels the ABC (attitudes, behaviour, choice) model, in an article
on climate change policy, arguing that

the idea that desires and attitudes drive behaviour produces a blind spot at a
particular crucial point, making it impossible to see how the contours and
environmental costs of daily life evolve (Shove, 2010: 1277).

In a similar fashion, rational choice models obscure the structural factors influencing
food choice. To continue the example above, women still do a disproportionate
amount of the work involved in providing and preparing food, but do not necessarily
have the authority to specify what food is eaten and often subordinate their

224
preferences and needs to other those of other family members (Charles and Kerr, cited
in Nettleton, 2006: 51-2). Changes such as eating differently from family and friends,
going against norms about accepting food and drink offered in hospitality and sharing,
and abandoning familiar foods and patterns of eating are much easier to make in a
social setting where the primacy of health-based or other nutritional arguments for
food choice are generally accepted. Such arguments have been acceptable amongst
affluent groups for much of this century, but, despite their increasingly wide social
diffusion, it is not clear that they are equally acceptable among all social groups
(Crotty, 1995, Levenstein, 2003).

Researchers into health inequalities have also considered the effect that ‘the social’ or
‘culture’ has on individual behaviour in order to explain the links between poor
health, behaviour and socio-economic disadvantage (Bartley, 2004). The authors of
The Black Report (see section 3.3) included ‘behavioural/cultural’ explanations of
health inequalities in their typology and tried to go beyond the existing dichotomy
between individual ignorance and irresponsibility and social structure (Townsend and
Davidson, 1982). However, Bartley argues that very little health inequalities research
has studied differences in health behaviour in terms of this anthropological sense of
culture (Bartley, 2004: 69). One of the few studies that has been done, by Mildred
Blaxter (1990), showed that although the patterns of behaviour she studied – smoking,
drinking, diet and exercise – were patterned by class, gender and age, ‘circumstances’
(by which she meant both the external environment and the individual’s psycho-social
environment) had more effect on health than these behaviours. Another important
finding was that

Unhealthy behaviour does not reinforce disadvantage to the same extent as

healthy behaviour increases advantage. This seems to suggest that the prior
effect on health is the general lifestyle associated with economic or
occupational position. Only in the more favourable circumstances is there
‘room’ for the considerable damage or improvement by the adoption of
voluntary health-related habits (Blaxter, 1990: 203).

Such findings show the limitation of explanations based on individual behaviour and
point to wider structural causes of ill-health. They also demonstrate how a stress on
voluntary behaviour change can reinforce existing inequalities since those who are

225
already advantaged will not only have greater capacity to make such changes, but
also, according to Blaxter’s research, to benefit from them.

These areas of research show that it is necessary to go beyond ‘rational choice’ or

ABC models to adequately explain individual eating behaviour and habits. The
‘writing out’ of the social that I have described here parallels that in risk factor
approaches to understanding chronic diseases and has broadly similar effects. It has
meant that the causes of variations in body size cannot be properly understood within
obesity science. Without a sociologically informed approach that considers these
causes fully, it is not be possible to understand the difficulties individuals have in
losing weight over the long term.

6.7 Conclusion

This chapter has analysed two important components of obesity science: the BMI and
the energy balance model. These are both important components of this body of
knowledge and practice but their effects are very different. The BMI was based on
understandings of body weight first developed in the insurance industry and then
imported into chronic disease epidemiology. It has a complex role of simultaneously
defining the problem of excess body weight in populations and individuals and
allowing it to be measured, quantified and framed as an epidemic. The energy balance
model derived from laboratory and clinical research into the physiological
mechanisms underlying body weight regulation. Like risk factor epidemiology it
involved a simplification and narrowing of understandings of human functioning, and
also a writing out of social processes and structures, in this case in eating behaviour
and habits. One problematic element of this model is its over-emphasis on individual
control over body weight. This meant that it could not explain the widely
acknowledged failure of conventional treatments for overweight and obesity (see
section 5.6). Such failure is an example of the non-plasticity of human bodies. It
demonstrates a social and physiological obduracy that results from factors that as well
as age and sex, include individual variation, life history (including the embodiment of
health inequalities), and differences between laboratory subjects and obese patients.
Eating as part of everyday life is very different from eating as part of a research study.

226
CHAPTER 7: THE INTERNATIONAL ARENA: OBESITY AND
OVERWEIGHT IN WHO REPORTS

7. 1 Introduction

In this chapter I discuss another arena in which researchers working on obesity

science were active – the World Health Organisation (WHO). This discussion gives a
brief outline of selected WHO publications in the area of chronic disease, diet and
body weight in order to illustrate a development of ideas that paralleled processes
taking place in the British obesity science reports discussed in the previous chapters.
It shows how researchers participating in the production of these reports, like many
officially recognised ‘expert’ scientists, were part of overlapping networks, both
national and international. From the 1980s, obesity and overweight began to be
discussed as part of an increasing focus on chronic diseases within the WHO, and
from the mid 1990s, the topic began to be the focus of more sustained interest,
culminating in the publication in 2000 of a widely publicised technical report
labelling excess body weight as a global epidemic (WHO, 2000: 185).

I will give a brief history of such publications and the individuals involved in their
production, and then discuss two areas of obesity science where the move into this
wider international arena was associated with significant developments in obesity
science – global estimates of prevalence and wider understanding of the negative
health consequences resulting from it. As I will argue below, framings of obesity and
overweight as a global phenomenon derived directly from its consideration in WHO
publications. The wider understanding of its health consequences resulted from the
continuing incorporation of large-scale epidemiological survey data into obesity
science. This data was combined with developments in physiological research to
explain the profound metabolic and endocrinological changes associated with excess
body weight. This shift is analysed in this chapter because it started in the mid 1990s,
and is best illustrated by discussions in the WHO reports. In order to demonstrate
development in this area, my analysis is preceded by a brief outline of discussions of
the negative health consequences from the earlier British reports (see chapters 4 to 6).

7.2 WHO reports on chronic disease, diet and body weight

227
In the WHO series of technical reports91, publications on the topic of nutrition are
relatively infrequent. Until the 1980s, those reports that dealt with nutritional issues
mostly considered malnutrition, especially protein malnutrition, rather than over-
nutrition. Protein malnutrition was considered in a series of reports produced by a
joint FAO/WHO expert committee on nutrition, which met every two years in the
1950s and every five years after that. Philip James’ career as an expert committee
participant seems to begin in 1985, as a contributor to a related FAO/WHO/United
Nations University report Energy and Protein Requirements (WHO, 1985), when he
was still Assistant Director of the Dunn Nutritional Laboratory in Cambridge. This
report was one of the first WHO publications to examine nutritional requirements
without an explicit focus on malnutrition or deficiencies. It contained international
standards for assessing the prevalence of obesity that would also be cited in later
reports (WHO, 1986: 27). Other participants from the world of British obesity science
also included J Durnin, who was now professor of the physiology department of
Glasgow University, and John Waterlow who was now professor of human nutrition
at London School of Hygiene and Tropical Medicine (LSTHM) (see section 3.3).
Another notable member was Dr Anna Ferro-Luzzi (see below).

Beginning in the 1960s, WHO also published more infrequent technical reports on
heart disease, but until the 1980s and The Prevention of Coronary Heart Disease
(WHO, 1982) (henceforth TR 678), these did not discuss the different risk factors in
any detail. The authors of TR678 discussed the issue of body-weight as one of a list
of ‘life-style’ factors that included diet and blood cholesterol, blood pressure,
smoking, physical activity, alcohol and oral contraceptives, but the section on body-
weight was brief .They argued that the ‘prevention or correction of obesity is of great
importance’ and referred to ‘the mass nature of obesity and the profound

91
The WHO technical report series currently consists of nearly 1000 reports. Approximately six such
reports are produced every year. The first, a report on a session of the Expert Committee on the
Unification of Pharmacopoeias was published in 1950 (WHO, 1950) and the most recent, number 957,
is the Forty-fourth report of the WHO Expert Committee on Specifications for Pharmaceutical
Preparations (WHO, 2010a). Many of the reports form long running series, such as on biological
standardization, drug dependence and food additives, but others appear more subject to the vagaries of
medical or policy fashions. Technical reports are designed to make ‘available the findings of various
international groups of experts that provide WHO with the latest scientific and technical advice on a
broad range of medical and public health subjects’ (WHO, 2010b). This fulfils the organisation’s
responsibility to act as a globally authoritative source of technical medical expertise (Lee, 2009,
Siddiqi, 1995).

228
sedentariness of many modern cultures’ (WHO, 1982: 22, 30), but framed the issue as
one of persuading individuals to follow appropriate guidelines on diet and levels of
physical activity (WHO, 1982: 31). This was compatible with approaches described in
chapter 2 where bodyweight was understood as a relatively unimportant risk factor for
heart disease: there was none of the sense of body-weight as an important issue in its
own right that later documents display (e.g. WHO, 1995, WHO, 2000). Because this
was a report on heart disease, and obesity science had already started to separate from
this community (see section 3.3), none of the researchers discussed in earlier chapters
were involved in the production of this report, although the chairman was Professor
Geoffrey Rose from the Department of Epidemiology at the London School of
Hygiene and Tropical Medicine (LSTHM) (see section 2.5).

Community Prevention and control of cardiovascular diseases (WHO, 1986)

(henceforth TR 732) also focused on heart disease. It appears to have been the next
WHO publication in which Philip James participated. Now the director of the Rowett
Research Institute (see section 3.4), James was again a temporary advisor to the
expert committee producing this report which also included Pekka Puska from the
National Public Health Institute of Helsinki, (see section 2.5) and Geoffrey Rose of
the LSTHM. The authors of this report described how

The emergence of mass coronary heart disease has accompanied the increase
in affluence in industrialized societies; affluence itself is not to blame for
cardiovascular disease, but only certain specific components of the affluent
life-style. Affluence generates powerful social forces which encourage the
adoption of a lifestyle that includes these adverse elements. That association is
not inevitable, however, since coronary heart disease death rates have been
declining in some countries recently without any decline in affluence (WHO,
1986: 9).

This anticipated discussions in later reports (especially TR797 and TR 894) of the
bodily harms associated with increasing affluence. However, teasing out of which
aspects of the affluent lifestyle were harmful to health was proving difficult, as shown
by the subsequent argument that higher mortality rates amongst the less affluent
‘cannot be fully explained by the standard cardiovascular risk factors’ (ibid.). The
authors framed the problem of heart disease and stroke as one of ‘nutrient excess’ and
argued for the ‘avoidance of weight gain and obesity’ by advising individuals to

229
‘increase their physical activity and to reduce the energy-density of their diet by
consuming more low-fat, low-sugar foods (WHO, 1986: 14).92 This was very
conventional advice at the time, as were the targets to reduce average saturated fat
contents of diet, but the authors’ suggestions that agriculture and food production
should be altered in line with health policy to encourage the production and
consumption of low fat dairy products and lean meats were more radical (WHO,
1986: 16-18). This was very close to the approach advocated in the UK government’s
Health of the Nation report on obesity published ten years later (see chapter 8).
Judgements about the relative healthiness of a ‘Mediterranean diet’ (James et al.,
1988: xii) containing significant quantities of fruit, vegetables, fresh fish, whole
grains and olive oil were implied in the authors’ recommendation of a largely plant
based diet. Given the extent to which such recommendations conflicted with food
production and processing commercial interests, the highly conditional and very
technical way such recommendations were framed might be seen as method of
avoiding the controversies that accompanied producing policy in the area of diet and
health.

In the late 1960s, the WHO Europe regional office had developed a research
programme into cardiovascular disease, particularly coronary heart disease (Lee,
1998: 94 - 95, Kaprio, 1991: 35-6), and it began to move into the area of nutrition and
health from the late 1980s (Baggott, 2000: 172). In this period, Philip James co-
authored Healthy Nutrition, a report on preventing nutrition-related diseases for the
WHO European regional office for Europe (James et al., 1988: 201). One of his co-
authors was Dr Anna Ferro-Luzzi, director of the Unit of Human Nutrition at the
National Institute of Nutrition in Rome, who worked for the FAO and the United
Nations University. Baggott cites the publication of Healthy Nutrition as the
beginning of the WHO’s move into this area, describing how it was followed by the
publication of a technical report on nutrition and chronic disease, the International
Declaration on Nutrition in 1992 and a WHO Nutrition Programme in 1993 which
declared the tackling obesity and diet-related disease as a priority, setting targets to
reduce the proportion of fat in average diets (Baggott, 2000: 172).

92
This approach is comparable to the ‘old approach’ which emphasised carbohydrate restriction and is
no longer part of ‘current nutritional thinking’ since it is now considered to ‘enhance rather than reduce
the cardiovascular complications of obesity’ (WHO, 1986: 27).

230
Healthy Nutrition was an overview of the links between diet and conditions such as
CHD, diabetes, cancer, cirrhosis, bone disease and anaemia. The authors argued that
from the mid-1960s ‘evidence was beginning to emerge suggesting that diseases not
normally associated with malnutrition had their origin in nutrition’ - citing the
example of heart disease - but by the 1980s obesity had become a ‘major public health
problem’ (James et al., 1988: 12). However, in this publication, obesity was discussed
both under the heading of conditions that predispose to major health problems and as
a major health problem in its own right, which demonstrates a certain amount of
confusion about its status. The brief discussion used European and American data to
make the same arguments regarding its prevalence and negative health consequences
as earlier British reports (see section 3.4) (James et al., 1988: 55-60). European
epidemiological data also cited used a variety of definitions of obesity including BMI,
the Broca index93 and Metropolitan Life recommended weights, showing that no one
standard definition of obesity was used at the time. Although the causes of obesity
were barely addressed the authors admitted that there were ‘no satisfactory
explanations for the marked differences in the prevalence of obesity in different parts
of Europe’ (James et al., 1988: 83). Since other section of the report documented
significant variations in dietary composition throughout the European region and
advocated a low-fat diet as a way of preventing obesity and other nutrition-related
diseases, this was an important gap in their argument. Suggesting that obesity rates
may be correlated with average fat consumption (ibid.) did not count as an adequate
explanation of these differentials.

Anna Ferro-Luzzi was also a member of the study group for the publication, Diet,
nutrition and the prevention of chronic diseases (WHO, 1990) (henceforth TR 797).
Philip James chaired the study group which included nutrition researchers from North
America, China, Egypt, Nigeria, the USSR and Japan. The aim of the report was

to describe recent changes in dietary and health patterns of countries, define

the relationship between the “affluent diet that typically accompanied

93
In 1871 Dr. Pierre Paul Broca, a French physician, anatomist and anthropologist created the formula
known as Broca's index where weight (in kg) should equal height (in cm) minus 100 (plus or minus
10% for men and 15% for women). According to the authors of Healthy Nutrition, obesity is defined
as 20% or more above this value, (James et al, 1988: 55).

231
 economic development and the subsequent emergence of chronic diseases, and
explore the need for national food and nutrition prevention policies to prevent
or minimize costly health problems (WHO, 1990: 9).

The affluent diet was described as one ‘characterized by an excess of energy-dense

foods rich in fat and free sugars but a deficiency of complex carbohydrate foods’
(WHO, 1990: 10-11). The report’s authors described the relationship between diet and
chronic diseases, such as CHD, hypertension and stroke, cancer and diabetes, as a
‘recently identified problem’ (WHO, 1990: 10). This framing of the ‘affluent diet’ is
an intermediate stage between the labelling of chronic diseases as ‘diseases of
affluence’ (see section 2.3) and later ideas that they are caused by the effects of
modernisation (see section 8.2). In an argument that anticipates the later idea of the
double burden of disease (see section 8.5), the authors stated that in countries with
relatively modest average incomes94 ‘the burden of cardiovascular disease and cancers
is nearly as great as in the very affluent countries with an income nearly three times as
great’ (WHO, 1990: 33). It was the beginning of a global understanding of chronic
diseases affecting populations of developing countries as well as those of
industrialised ones.

TR797 contains many references to obesity, both as a disease in its own right and as a
result of rapid changes taking place in both diet and life conditions of many
populations. The authors argue that obesity was a problem in developing as well as
industrialized countries, mentioning very high rates amongst women in Trinidad
(WHO, 1990: 28). Widely varying rates of obesity amongst different populations
were argued to be the result of ‘environmental factors’ such as diet and levels of
physical activity, rather than solely the result of variations in genetic susceptibility
between populations, leading to an argument for public health measures to modify
‘the population’s environmental circumstances’ (WHO, 1990: 69). The cause of
obesity was understood to be an imbalance between energy intake and expenditure,
but this imbalance was linked to economic development, since as well as
unprecedented dietary changes, processes of industrialisation and urbanisation had led
to reductions in levels of physical activity (WHO, 1990: 72). The overarching
message of this report was of the health benefits of a high-carbohydrate, low-fat,

94
They refer annual GNP per capita in the range of $3000 to $4000.

232
largely plant-based diet (WHO, 1990: 158), which was important in preventing many
types of diet-related chronic disease as well as obesity and overweight. This too was
part of the new understanding of chronic disease as a consequence of economic
development (see section 5.5 and 8.2), which could be experienced by the poor as
well as the rich.

The first report that contains a significant discussion of overweight (the authors
refused to use the term obesity)95 was Physical Status: The Uses and Interpretation of
Anthropometry (WHO, 1995) (henceforth TR854). The expert committee responsible
for compiling TR 854 included Anna Ferro-Luzzi and Jakko Tuomilehto of the
Finnish National Public Health Institute who was part of the North Karelia project and
MONICA (see section 2.5). The rest of the committee comprised epidemiologists and
nutrition researchers from America, Australia, Brazil, Columbia, India, Netherlands
and Nigeria, as well as advisors from the FAO and UNICEF. Philip James was also
among a long list of individuals thanked for their extensive contributions (WHO,
1995: 412). The chapter of TR854 on overweight closely resembles the next relevant
report which was entirely focused on the topic, Obesity: Preventing and Managing
the Global Epidemic (WHO, 2000) (henceforth TR894). The publication of TR894 in
2000 was part of an important increase in activity around this area of health policy
that began in the mid 1990s, and included several British publications (see chapters 4
to 6). In the international context, TR 854 was published in 1995, the consultation
meeting for TR894 took place in 1997, the International Obesity Taskforce (IOTF)
(see below) was set up in 1998 and TR894 was published in 2000. Two technical
reports were produced in a couple of years - a relatively short period of time for a
large bureaucracy like the WHO and, therefore, a significant burst of activity in the
area of body weight and health.

Several researchers from the Rowett Research Institute participated in the production
of TR894, presumably due to the presence on the secretariat of Philip James in his
role as the chair of the IOTF. Two other individual from the Rowett served on the
95
The authors defined normal weight and three grades of overweight using the standard BMI cut-off
points for obesity and overweight with the following caveat: ‘Because BMI does not measure fat mass
or fat percentage and because there are no clearly established cut-off points for fat mass or fat
percentage that can be translated into cut-offs for BMI, the Expert Committee decided to express
differently levels of high BMI in terms of degrees of overweight rather than degrees of obesity (which
would imply knowledge of body composition)’ (WHO, 1995: 312).

233
secretariat, both of whom are also listed as members of the IOTF. The introduction of
TR894 also describes the consultation process as ‘undertaken in close collaboration
with the Rowett Research Institute…and the International Obesity Taskforce (IOTF)
chaired by Professor Philip James’(WHO, 2000: 1). This level of representation, and
such an explicit acknowledgement of the role of the Rowett/IOTF staff, suggests that
Philip James had an important role in the production of this report.

As is appropriate for a WHO consultation, other members of the expert committee for
TR894 came from all regions of the world, including representatives from the US,
Canada and Europe, China, Malaysia and Japan, India and Pakistan, and Nigeria and
South Africa. One of the American members was George Bray (see section 4.4), the
chairman was Professor James Hill of the University of Colorado and another was Dr
William Dietz, a paediatrician from the North-east Medical Centre, Boston. Two other
high profile members were Professors Arne Astrup of the Royal Veterinary and
Agricultural University, Copenhagen and Per Bjorntorp of Gothenberg University,
Sweden. Per Bjorntorp was a clinical researcher interested in metabolic problems
related to type II diabetes, obesity and cardiovascular disease. Arne Astrup was a high
profile nutrition researcher whose research interests included fat metabolism, energy
expenditure and dietary management of obesity, and was also a member of the IOTF.

7.3 Obesity as a global health problem

Previous obesity science writing discussed prevalences of obesity and overweight in

European and North American populations (see section 4.6). However, WHO is a
global institution whose work has largely focused on health problems in developing
countries. Presumably, this meant that it was important to frame obesity as global
phenomenon. From the mid 1990s, obesity did become discussed as a global health
problem. Standard definitions of overweight and obesity combined with an increasing
amount of data permitted analyses of changing prevalence rates that were both global
in scope and acknowledged the differences between populations. The 1995 WHO
technical report on anthropometry (TR854) was the first to contain a detailed
discussion of overweight as a global phenomenon, citing prevalence rate data from
countries in Africa, South America, the Caribbean and Asia, as well as from more
than twenty Pacific and Indian Ocean island populations. The data for these small

234
island populations showed a striking variation of levels of obesity within as well as
between populations: the prevalence rates quoted for the population of Papua New
Guinea varied from 4.7% for Highland men to 64.3% for urban women, and the rates
for Chinese men in Mauritius were 2.1% compared with a rate of 20.7% among
Creole women in Mauritius (WHO, 1995: 314-5). Overall, the global situation was
summarised as follows:

Overweight is a major public health issue. Grade 2 overweight…is relatively

common in most industrialized societies and also in many less modernised
cultures: data compiled recently show that the prevalence among 20-60 years
olds is about 10-20% among whites in the USA and most countries of
Europe…Prevalence is high (20 – 40%) among women in eastern European
and Mediterranean countries and black women in the USA. Even higher
prevalences are observed among American Indians and Hispanic Americans,
and on the Pacific islands…, with probably the highest rates in the world
among Melanesians, Micronesians, and Polynesians… In some African and
Asian countries prevalence is much lower but in countries of South America
and the Caribbean the prevalence of grade 2 overweight may be close to that
in many European countries (WHO, 1995: 313).

This is obesity explicitly understood as a global public health problem and to

reinforce this, the authors referred to 1980s national survey data for the US, Australia,
Finland, the Netherlands, Sweden and the UK. As this data demonstrated that the
prevalence of overweight ‘remained stable or increased’, it is debatable how much it
contributed to their argument, but the authors still managed to incorporate it.

Since then, there seems to be no indication of a decrease in the prevalence of

overweight in these affluent countries despite increased commercial and other
interests in promoting leanness; on the contrary, the prevalence of overweight
may be increasing further (WHO, 1995: 317).

By the late 1990s, the authors of the technical report on obesity (TR894) had
developed the framing of obesity as a global health problem and tried to provide a
more complete, and therefore authoritative, picture of global trends. An important
new aspect of TR894 was the use of prevalence data from the WHO MONICA
project.96 This information was crucial for the authors’ framing of obesity as a global
health problem and MONICA was described as the ‘most comprehensive data on the
96
The WHO MONICA (Multinational Monitoring of Trends and Determinants in Cardiovascular
Disease) Project 1979-2002 was described as the world’s largest and longest study of heart disease and
stroke (Tunstall-Pedoe, 2003) (see also section 2.5).

235
prevalence of obesity worldwide’ (WHO, 2000: 17 , see also Tunstall-Pedoe, 2003).
Despite arguing that ‘the prevalence of overweight and obesity [is] increasing
worldwide at an alarming rate’ (WHO, 2000: 16), the quality of the evidence cited to
illustrate this dramatic claim varied considerably between different global regions.97

In the African region, the authors admitted that the principal focus of research had
been on undernutrition and food security so that increases in obesity rates were
documented for only a few countries and information of current prevalences was
‘fragmentary and limited’. One study conducted in Mauritius showed the same
upward trends as the other WHO regions, and five other small scale studies were used
to demonstrate the current prevalence rates of obesity in African countries, that
ranged from 0.6% in Tanzanian men in 1986-9 to 44% among black South African
women in 1990 (WHO, 2000: 21). There was thought to be a significant difference
between rural adults in developing countries who maintained their weight and urban
dwellers who did not: ‘with improvements in socio-economic status and increasing
changes due to rapid urbanization, the prevalence of obesity amongst some groups of
black women has risen markedly to levels exceeding those in populations in
industrialized countries’ (ibid.)98. In this report, a total of six small scale studies, on
populations who were unlikely to be representative, were cited as evidence for
changes in the average weights of the populations of an entire continent.

For the Eastern Mediterranean region, the authors admitted that good quality
nationally representative secular trend data was not available, but the limited data that
existed ‘indicate that the prevalence of adult obesity in countries in the Region is high
and that women in particular are affected’ (WHO, 2000: 26). For example, a 1990-3
cross-sectional survey of 13177 adult Saudis showed a prevalence of obesity amongst
women several times higher than current rates in developed countries and higher than
that of Saudi men (ibid.). Similarly, ‘Good quality, nationally representative secular

97
The six WHO regions are Africa, the Americas, the Eastern Mediterranean (most of the Middle
Eastern countries plus Afganistan, Dijbouti, Pakistan, Somalia and Sudan), Europe (which includes
Israel, all of the former USSR countries and Turkey), the Western Pacific (which includes China and
Japan) and South East Asia (which includes India and Bangladesh). This division resulted from a
mixture of geographical and political considerations (Siddiqi, 1995) and it operated as an important
structuring device for the epidemiological data considered below.
98
This relates to discussions about the role of industrialisation and modernisation in increasing average
body weights (see chapter 8).

236
trend data for countries in the South-East Asia Region were unavailable’ (WHO,
2000: 23). Two small scale studies of Thai government officials and affluent urban
dwellers showed that in 1985 2.2% of men and 3.0% of women were obese, and in
1991 3.0% of men and 3.8% of women were obese (ibid.). However, these groups
were unlikely to be representative of the wider population of this region.
Furthermore, because these prevalence rates were notably low and South-East Asia
contains a large fraction of the world’s population, the absence of such data was a
significant omission, weakening the report’s argument that obesity should be
considered as a global public health problem.

For the Americas region, secular trend data showed obesity rates increasing in both
developed countries, like the US and Canada, and in developing countries, like Brazil
(WHO, 2000: 21). Figures for the US derived from a series of national health
surveys99 demonstrated that ‘obesity is an escalating problem in the USA’ (WHO,
2000: 22): between 1960 and 1994 the prevalence among men increased from 10.4 to
19.9%, and among women it increased from 15.1% to 24.9%, and these rates of
increase accelerated in the 1980s. At the time of the report’s publication 20% of
American men and 25% of women and 15% of Canadian men and women were
considered obese and black women and other minority groups in the US were
understood to have ‘particularly high rates of obesity’ (WHO, 2000: 23). The current
prevalence in Brazil was given as 6% among men and 13% among women, and
obesity was also seen as a major public health problem in parts of the Caribbean
(ibid.).

Data from the MONICA study was an important source of information for the Europe
Region. However, the report’s authors argued that MONICA populations were not
necessarily representative of host countries, and so the best data on secular trends in
obesity came from national surveys which were available for England, Finland,
Germany, the Netherlands and Sweden. These surveys showed that ‘the prevalence of
obesity has increased by about 10-40% in the majority of European countries in the
past 10 years’ and the most dramatic increases occurred in England where rates had
99
These were NHES and NHANES (see section 4.4) which took place between 1960-2, 1971-4, 1976-
80 and 1988-94. The data derived from these surveys shows slight increases in prevalence between the
first three surveys and much larger increases between last two.

237
doubled in the last ten years (WHO, 2000: 24). Obesity was seen as relatively
common in Europe:

The average prevalence of obesity in European centres participating in the

WHO MONICA study between 1983 and 1986 was about 15% in men and
22% in women, although there was great variability both within and between
countries (WHO, 2000: 25).

The existence of great variability between populations suggests that excess

bodyweight might be a problem of specific populations rather than all European
countries (see section 9.2) but this is not an approach that was discussed here.

In the Western Pacific region data was available showing increasing prevalence of
obesity amongst Australian, Samoan, Japanese and Chinese populations (WHO, 2000:
27). The current prevalence among the white population of Australia and New
Zealand was given as 10-15%, but in the aboriginal populations it was thought to be
either much higher or much lower depending on location (ibid.). The prevalence
figures given for Japan were 2% for men and 3% for women, and for China 1% for
men and 2% for women with studies finding that obesity is commoner in women than
in men and also in urban rather than rural areas (WHO, 2000: 29). This was evidence
of very low rates in another large country, containing a significant proportion of the
global population, which again undermines claims of a global epidemic. In contrast to
these very low rates, very high prevalence rates of obesity had been found in the
Pacific Island populations of Melanesia, Micronesia and Polynesia (WHO, 2000: 30).
The most extreme example given was urban Samoa where a marked increase in the
prevalence of obesity had been observed between 1978 and 1993, and the prevalence
of obesity was given as over 75% in women and approximately 60% in men.
However, there is an important caveat attached to these figures:

Polynesians seem leaner than Caucasians at any given body size so the
prevalence of obesity in Polynesian populations may not be quite as high as is
currently estimated using Caucasian-derived classifications based on BMI
(ibid.).

The emphasis on this point is part of the ongoing debate about the applicability of
standard BMI cut-off points (see section 4.4 and 7.4).

238
Despite the centrality of increasing prevalences to the argument that obesity was a
significant global health problem, the amount of available data for different regions of
the world varied widely and so did the quoted incidences of obesity - from 0.8% for
both men and women in Mali and 1% for women and 2% for men in China, to 15%
for men and 17% for women in England, 20% for men and 25% for women in the US
and 58% for men and 77% for women in urban Western Samoa. However, despite the
patchiness of the available information, prevalences were framed in both reports as
uniformly rising. In TR894, current prevalences and secular trends for each region of
the world were discussed, despite the fact that the US and countries such as England,
Finland and the Netherlands were unusual in having national survey data
demonstrating increasing prevalences; for most African, Asian and Middle Eastern
countries such data was very fragmentary or simply did not exist. These gaps
demonstrate that the framing of obesity as a global public health problem seems to
have largely preceded the data.

7.4 Particular populations as exemplars

As the discussion of the prevalence data given above suggests, some populations were
seen as particularly prone to both obesity and overweight and the physiological
consequences of excess weight. These populations functioned as exemplars within
obesity science, warnings of future possibilities for societies currently with lower
rates, and ways of understanding the causes of excess body weight and the chronic
diseases associated with it. The Pima Indians (see section 5.3) and the Samoans were
regularly referred to in such discussions. For example, in TR854 analyses of
populations among whom overweight was now very common - the Pima Indians in
the US, Fijians, Maltese, Melanesians, Nauruans and Samoans - suggested that ‘the
condition [of obesity] was uncommon before the adoption of sedentary lifestyles and
high-fat diets’ (WHO, 1995: 317). Because they referred to particular populations,
such discussions could readily shift from the role of economic development to genetic
explanations. In this way, descriptions of the effects of increasing affluence on
average body weight were combined with explanations focusing on the role of genetic
causes in differential rates of obesity and overweight:

239
 The expression of overweight requires a certain level of food availability
above which the relative contributions of genetics and environment probably
vary within and across populations. High-fat diets combined with low levels of
physical activity play an important role in the increase of overweight that
accompanies the transition from poverty to affluence…In an affluent
population of individuals with similar socio-economic values and resources,
genetic factors will become relatively more important in determining which
individuals will become obese (WHO, 1995: 317).

This allows the authors to mention the ‘“thrifty genotype” hypothesis’ - that
‘populations exposed to inadequate or fluctuating food supplies are genetically
selected for a high level of efficiency in caloric utilization or fat storage…[which]
may lead to an increase in the prevalence of overweight and non-insulin-dependent
diabetes mellitus’ (ibid.) - as one example of such genetic factors (see section 9.5).

In the 1990s, the debate about the use of BMI to measure body fat transformed into a
narrower argument about the extent to which different populations are characterised
by differing body proportions. The BMI classification, and its cut-off points, was used
in both the later WHO technical reports, demonstrating how it had become, in some
senses, a globally standard definition. However, the fact that there were still ongoing
discussions about the appropriateness of this definition for different population groups
suggests that this process of standardising the definition was partial and incomplete.

Body mass index appears to be a good indicator of the deposition of excess

energy as fat in adult white men and women living in Europe and North
America. It is probably less appropriate in other population who differ in body
build and body proportions’(WHO, 1995: 327).

BMI had become generally accepted as an appropriate measure of body fat for
Caucasian populations, but discussions were beginning about how to devise
appropriate ethnically specific definitions and classifications for other populations.
Despite the centrality of BMI to the anthropometry of overweight, this debate was
mentioned only in this quote and a brief discussion of the differential risks of cardio-
vascular disease and diabetes amongst different overweight populations (WHO, 1995:
328). It was, however, another way in which obesity and overweight were not yet
fully global phenomena (see section 7.3).

240
This acceptance of a BMI-based definition meant that the authors of TR894 could cite
the earlier report, give the same cut-off points and, therefore, spend less time arguing
for a particular definition and measure of obesity.100 They also acknowledged some of
the criticisms of BMI, and stressed its use as a population rather than individual
measure to forestall others.

BMI can be considered to provide the most useful, albeit crude, population-
level measure of obesity. The robust nature of the measurements and
widespread inclusion of weights and heights in clinical and population health
surveys mean that a more selective measure of adiposity, such as skinfold
thickness measurements, provides additional rather than primary information.
BMI can be used to estimate the prevalence of obesity within a population and
the risks associated with it, but does not, however, account for the wide
variation in the nature of obesity between different individuals and
populations. (WHO, 2000: 9).

Waist circumference and WHR (waist to hip ratio), were discussed but there were also
problems with developing standard cut-off points due to significant differences
between populations:

For instance, abdominal fatness has been shown to be less strongly associated
with risk factors for CVD and non-insulin dependent diabetes (NIDDM) in
black women than in white women…Also people of South Asian
(Bangladeshi, Indian and Pakistani) descent living in urban societies have a
higher prevalence of many of the complications of obesity than other ethnic
groups…These complications are associated with abdominal fat distributions
that is markedly higher for a given level of BMI than in Europeans. Finally,
although women have almost the same absolute risk of coronary heart disease
(CHD) as men at the same WHR…they show increases in relative risk of
CHD at lower waist circumferences than men. Thus, there is a need to develop
sex-specific waist circumference cut-off points appropriate for different
populations (WHO, 2000: 10).

Because of such arguments, these authors stressed that the BMI cut-off points given
were specific to Caucasian populations (WHO, 2000: 11). Despite ongoing
sociological debates as to its validity (Epstein, 2007, Fausto-Sterling, 2004), race and
ethnicity continued to operate as variables in this area of biomedical research.
Amongst the ‘non-genetic’ factors in the development of overweight and obesity were
sex - women have, on average, a higher percentage of body fat - and ethnicity:
100
The fact that this report used the terminology overweight and obesity, rather than just overweight
like TR854, may reflect an increasing consensus around the use of BMI or it may be due to the
differing composition of the expert committees that produced the two reports.

241
 Ethnic groups in many industrialised countries appear to be especially
susceptible to the development of obesity and its complications. Evidence
suggests that this may be due to a genetic predisposition to obesity that only
becomes more apparent when such groups are exposed to a more affluent
lifestyle (WHO, 2000: 138).

Such groups included the Pima Indians of Arizona, Australian aboriginals, indigenous
Hawaiians and South Asians overseas. Obesity and its complications were seen as the
result of genetic susceptibilities combined with the transition from a ‘traditional’ to a
‘more affluent and sedentary lifestyle and its accompanying diet’(WHO, 2000: 139).
The example of African Americans in the US ‘where the highest rates of obesity are
found in the poorest communities’ (ibid.) was also given to argue that other
environmental factors may be important.

In these populations, fat-rich, energy-dense diets are likely to be the cheapest,

and reduced levels of activity stem from unemployment. Other factors
associated with poverty may also be involved (ibid.).

Only the African American population in the US was described as poor, and its high
rates of obesity related to this poverty. It would have been equally valid to describe
many of the communities listed above – especially the Pima Indians in Arizona,
Australian aboriginals and indigenous Hawaiians – as poor. This leads to the question
of why the prevalence of obesity in these communities was not explicitly linked to
high levels of poverty in the same way (see section 9.5).

7.5 Relative risk and the expanded list of health consequences

As I described in chapter 3, the individual paradigm contained a list of health

consequences that was relatively short, usually consisting of heart disease, diabetes,
mechanical problems (joint pain and breathing difficulties) and psychological
problems (see section 3.4). In the 1990s this list became extended and standardised
using information from Framingham (see section 2.5), and the American Cancer
Society studies (see section 3.4). This new data not only increased the number of
conditions that were seen to be associated with obesity, but was also used in the
development of a wider understanding of the health consequences of overweight and

242
obesity. In contrast to earlier, largely clinical, knowledge, it consisted of statistics of
relative risk, derived from large population studies, which were combined with
existing knowledge to produce a new understanding of wider metabolic effects of
obesity than had been previously recognised, and the greater damage to health,
particularly over the long term.

When this new understanding was first developing, authors typically explained these
relationships at greater length and outlined models accounting for these effects. Thus,
Peter Kopelman discussed the hypertensive consequences of obesity using
Framingham data to illustrate the relationship between the two conditions: ‘greater
degrees of obesity were associated with progressively higher levels of blood pressure
and hypertensive cardiovascular disease’ (Kopelman, 1984: 623). The causes of
hypertension in these individuals were not known but

The alterations in endocrine function, which accompany weight gain, may

contribute to an increase in blood pressure and there appears to be a
relationship between plasma insulin and catecholamine101 concentrations, fat
cell size and the development of hypertension (Kopelman, 1984: 630).

Wide ranging endocrinological changes were now seen to result from obesity which
increased rates of hypertension, and, so, stroke and heart disease. Other authors of the
time listed the conditions with which increased body weight was associated - coronary
heart disease, hypertension, diabetes, gall bladder disease, arthritis, cancer, disorders
of the respiratory and menstrual systems and some psychological conditions – using
the 1983 RCP report (Hautvast and Deurenberg, 1987: 67). However, obesity was
also seen to be associated with high blood cholesterol, reduced glucose tolerance and
high blood pressure, but it was not clear whether these were caused by obesity or its
cause (Hautvast and Deurenberg, 1987: 67-8), and demonstrating this uncertainty,
these conditions were described as confounding factors (see section 6.5). Writing in
the same period, Garrow provided a detailed discussion of the direction of causality in
these associations:

This survey [the Seven Countries Study] showed that overweight men were
more likely to have heart attacks, but that when age, smoking, blood pressure

101
This is another name for a group of hormones such as adrenaline, noradrenaline and dopamine that
are released by the adrenal glands in response to stress.

243
 and serum cholesterol were taken into account, relative weight did not
significantly predict coronary heart disease among men aged 40-59 years. This
finding may be interpreted to mean that weight is irrelevant to coronary heart
disease, or alternatively that it increases the risk of death from coronary heart
disease by contributing to hypertension and hypercholesterolaemia (Garrow,
1988: 11).

Evidence from feeding studies done on prisoners (see section 6.4) was inconclusive
but data from Framingham demonstrated that obesity, even if not associated with high
blood pressure and high cholesterol, was not ‘benign’; in fact, it was a ‘significant and
independent predictor of disease, especially among women’ since it was the third
most reliable indicator of heightened risk of cardiovascular disease for women, after
age and blood pressure (Garrow, 1988: 13). However, these relationships were not
linear: the highest values of blood cholesterol and coronary atheroschlerosis were not
found amongst the heaviest patients (ibid.). Other conditions associated with obesity
and overweight were diabetes, gallbladder disease, impaired reproductive function
(including menstrual irregularity, ovarian failure and polycystic ovary syndrome),
cancers of the rectum, colon and prostate in men and the gallbladder, endometrium,
cervix, ovary and breast in women102, mechanical problems such as osteoarthritis in
joints, and social and psychological disadvantages (see below) (Garrow, 1988: 13 -
15).

Garrow’s list of the conditions associated with excess body weight was widely
reproduced in the next ten years. In 1994, West gives the same list of diseases as
Garrow103 which are associated with obesity (West, 1994: 13 - 14) and the 1995
Health of the Nation report discusses the topic briefly using the 1983 RCP report,
Garrow’s 1988 textbook and an American report on diet and health. The authors
argued that

There is a large body of epidemiological evidence that links obesity to

increased mortality and increased risk of chronic diseases such as
cardiovascular disease, certain cancers, diabetes, gallbladder disease,
hypertension and a range of bone joint and skin disorders…Risk of disease
increases with BMI throughout the range of BMI, but is particularly marked at
high BMI (DoH, 1995: 3).

102
The first ACS study is cited for these conditions.
103
Again, the authors cite the first ACS study.

244
As there is a significant overlap in the evidence used by two of these publications, it is
arguable how large this body of epidemiological actually was, but these authors no
longer had to make a case for this link, they could merely state it.

Similarly, in the SIGN report, the health complications of weight gain were listed as
diabetes; raised blood pressure; stroke; hyperlipidaemia (and low levels of HDL
cholesterol); coronary heart disease; gallstones; cancers (post-menopausal breast,
endometrial, ovarian, gallbladder and colon cancers); breathlessness, respiratory
disease and sleep apnoea; menstrual abnormalities, pregnancy complications
including perinatal mortality and gestational diabetes; musculo-skeletal disorders and
osteoarthritis (back joint and foot conditions); stress incontinence; psychological
problems (social isolation, low self-esteem, binge eating and night eating), and
disability (SIGN, 1996: 10-11). The list is long and daunting, but derived largely from
Garrow’s 1988 textbook, and included ‘reduced employment prospects’ and ‘early
retirement’ which are important, but not health, consequences. Such blurring of the
medical and the social gives the impression of constructing an impressive list partly
for rhetorical purposes. But, for the first time, these authors discussed the hazards of
intentional weight loss - gallbladder disease, the risk of which increases with
increasing weight loss, and decreased bone density (SIGN, 1996: 14). Apart from
Garrow’s acknowledgement that weight loss does not reduce the incidence of
gallstones (Garrow, 1988: 19), this is the first reference, within mainstream obesity
discourse, I have found to the negative health consequences of losing weight.

In the BNF report, the complications of obesity were given as non-insulin-dependent

diabetes mellitus (NIDDM), coronary heart disease, cancer, osteoarthritis, gallstones,
sleep apnoea, reproductive disorders, complications of pregnancy, psychological
disorders and social penalties (BNF, 1999: 7-12). Evidence for the link between
diabetes and obesity came from studies of the metabolic consequences of overfeeding
both humans and rhesus monkeys. In the human studies, the volunteers increased their
body weight by roughly 20% in 6 months:

There was a significant increase in fasting insulin, glucose and triaglyceride,

cholesterol and amino acids, along with a decrease in oral and intravenous
glucose tolerance. These are all changes of the type that characterise the non-
insulin-dependent diabetic. The subjects did not become frankly diabetic but

245
 their BMI at maximum weight was only about 28. The overfeeding was then
stopped, they lost weight and the insulin sensitivity reverted to normal (BNF,
1999: 7).

In rhesus monkeys the development of obesity and insulin resistance was followed by
NIDDM. There was also epidemiological evidence for the link between obesity and
diabetes:

In epidemiological studies, the prevalence of diabetes increases with

increasing severity of obesity, with increasing duration of obesity and with
increasing age. Analysis of the National Health and Nutrition Examination
Study (NHANES) data shows that, for each kilogram increase in weight of the
population, the risk of diabetes increases by 4.5% (BNF, 1999: 7-8).

However, the association between obesity and diabetes ‘underestimates the

contribution of obesity to the incidence of diabetes’ (BNF, 1999: 8) because
individuals often lose significant amounts of weight when they develop overt
diabetes.

The evidence for the link between coronary heart disease – ‘the main cause of excess
mortality among obese people’ - and obesity was epidemiological. However, in this
case, the report’s authors had to construct an argument incorporating the findings of
the Seven Countries study, which had shown that obesity was not an independent risk
factor for CHD when age, cigarette smoking, blood pressure and serum cholesterol
were already accounted for (in a calculation to predict heart attacks in men aged 40-
59). To do this, they argued that

This paradoxical solution arises because obesity itself is strongly related to

hypertension and stroke, particularly in young people…Obese people also
have an adverse blood lipid profile, and therefore, in Key’s multiple regression
equation, part of the effect of obesity has already been ‘explained’ by the
hypertension and cholesterol (BNF, 1999: 9).

The data for increased cancer risk, once again came from the first ACS study. It
showed that the mortality rate for cancer among men who were 40% overweight was
1.33 and the mortality rate for women at the same level of overweight was 1.55
(ibid.). The difference in increased risk between men and women was not discussed,
but the increased risks included post-menopausal breast cancer in women as well as

246
cancers of the endometrium, uterus, cervix, ovary and gall-bladder in women and
those of the colon, rectum and prostate in men (ibid.). Overweight and obesity were
also seen to have negative effects on the female reproductive system: ‘Significant
associations are seen in reproductive endocrinology between excess body fat and
ovulatory dysfunction, hyperandrogenism and hormone sensitive carcinomas’ (BNF,
1999: 10). Excess body weight was also seen to increase the risk of complications in
pregnancy: ‘even moderate degrees of obesity are associated with an increased risk of
hypertension, toxaemia, gestational diabetes, urinary tract infections and fetal
macrosomia’ moreover, ‘perinatal death is three times more common in obese than in
thin women’ (ibid.). The effect of obesity and overweight on the male reproductive
system was also discussed, and the authors stated that decreased levels of testosterone
were found in massively obese men (ibid.). These arguments also illustrate an
increased understanding of the endocrinological effects of overweight and obesity.

In the two WHO technical reports (TR854 and 894), evidence for the negative health
consequences of obesity and overweight was discussed in a new way that focused
more on the increased relative risk of associated conditions rather than on establishing
plausible biological mechanisms. In 1995, the authors of TR854 discussed the health
consequences of overweight under the headings of the major conditions with which it
was understood to be associated – heart disease, stroke, hypertension, NIDDM, gall
bladder disease, osteoarthritis and cancer (WHO, 1995: 323 - 7)104 - by summarising
the available evidence. For diabetes this evidence was particularly strong:

During a 8-year follow-up of 113 861 women in the USA, aged 30 – 55 years,
the risk of developing NIDDM increased with increasing BMI…Compared
with women with a BMI below 22, risk was increased 20-fold for women with
a BMI between 29 and 31, and more than 60-fold for those with a BMI above
35. Within the total cohort, 90% of diagnoses of NIDDM were attributable to
a BMI greater than 22 (WHO, 1995: 324).

In the case of gallbladder disease, the relative risks quoted also showed substantial
increases: data from the American Nurses Health Study demonstrated that women
with a BMI of above 32 had a six fold increase in their risk, compared with women

104
They list other associated disorders such as varicose veins, infertility, psychosocial problems, hiatus
hernia, sleep apnoea and ‘important social and economic disadvantage’ (WHO, 1995: 326-7) without
any further discussion.

247
with a BMI of less than 20 (WHO, 1995: 325).105 However for other conditions the
evidence seems more equivocal. In the case of hypertension research linked weight
loss to decreased systolic and diastolic blood pressure, but body fat distribution was
also seen to be related to hypertension, independently of BMI (WHO, 1995: 324),
which undermines the earlier association. For coronary heart disease the evidence also
seems mixed. The most direct statement of the association between heart disease and
overweight was that:

The relationship between BMI and CHD has usually been found to be linear,
but the level of risk is modified by ethnicity, age, sex, and smoking habits.
Elsewhere, it has been calculated that about 40% of the incidence of CHD was
attributable to a BMI of above 21 and was therefore potentially preventable
(WHO, 1995: 323).

Framingham data showed that a 10% reduction in bodyweight corresponded to a 20%

reduction in the risk of developing heart disease (Ashley and Kannel, 1974), but
studies following intentional weight loss had not been able to demonstrate this
reduction in risk (WHO, 1995: 323), which is unexpected given the strength of the
relationship between overweight and coronary heart disease that was being asserted
here.

For specific cancers the results described were also equivocal as a review of studies of
the relation between weight and six types of cancer showed only one definite increase
in risk for endometrial cancer, a probable increase in risk for post-menopausal breast
cancer, and that the ‘relationships between overweight and cancer of the colon,
rectum, ovaries, and prostate are uncertain; reported associations are inconsistent
between and within sexes and across populations’ (WHO, 1995: 326).

Some of this evidence even seems distinctly unhelpful to the arguments the authors
were trying to make: ‘Despite the clear relationship between overweight and
hypertension… it has been concluded that overweight is not one of the major risk
factors for stroke… although three prospective studies…have shown that abdominal
fatness may be associated with increased risks for stroke independently of BMI’

105
But in order to assess the importance of such increases, it would be necessary to know the sizes of
the absolute risks involved.

248
(WHO, 1995: 323). The overall finding was that abdominal fatness was an
independent risk factor, which undermines a wider case about the negative health
consequence of overweight and obesity per se. This description of equivocal or non-
supporting evidence was partly a demonstration of transparency and impartiality: the
report’s authors were publicly weighing all the available evidence, not just the results
that supported their case.

Five years later, in TR894, the list of negative health consequences, derived from
Framingham and the ACS surveys, remained the same. However, the evidence for
these increased risks was discussed in more detail as the authors devoted an entire
chapter of the report to this topic. They reproduced the now standard list of the major
health consequences – diabetes, heart disease, hypertension, gallbladder disease,
psychosocial problems and certain cancers (WHO, 2000: 39) - but for the first time,
they divided them into different categories based on increase in relative risk - see
table 7.5 (below).

!"#$%& ' ( /( !" $%& ' ( !$,- !" $%& ' (

) ! $* $'+ , - % . ) ! $* $'+ 0# . ) ! $* $'+ #0.

! " #
%
$
& ' "
( " %

) %

Table 7.5 (WHO, 2000: 43)

Many of the new conditions now included in the broader understanding of the
negative health consequences, such as cancers, impaired fertility and anaesthetic
complications, showed only slight increases in relative risk. This list was reproduced
in subsequent policy documents (Health Committee, 2004, Branca et al., 2007b), but
this classification was removed, and these conditions were listed as if they were all
equally likely. The authors themselves acknowledged that some of these risks were
more serious than other, arguing that the ‘more life-threatening chronic health

249
problems’ could be classified into four different types – cardiovascular conditions,
those associated with insulin resistance, certain cancers and gallbladder disease
(WHO, 2000: 43), but this qualification was often also omitted when the information
was reproduced in later documents.

For cardiovascular disease and hypertension, the authors argued that

Obesity predisposes an individual to a number of cardiovascular risk factors

including hypertension, raised cholesterol and impaired glucose tolerance.
However, longer-term prospective data now suggest that obesity is also
important as an independent risk factor for CHD-related morbidity and
mortality…The Framingham Heart Study ranked body weight as the third
most important predictor of CHD among males, after age and
dyslipidaemia…Similarly, in women, a large scale prospective study in the
USA106 found a positive correlation between BMI and the risk of developing
CHD (WHO, 2000: 47).

Data from Framingham and the Nurses Health Study was still crucial for obesity
researchers constructing arguments about the negative health consequences of
overweight and obesity. Without large scale prospective data demonstrating these
associations, they could not have made these arguments convincingly.

For hypertension, the authors used data from NHANES II to argue that the prevalence
of hypertension in overweight adults was nearly three times that for the non-
overweight, and moreover that

The risk in those aged 20-44 years is 5.6 times greater than for those aged 45-
74...which in turn is twice as high as that for non-overweight adults. The risk
of developing hypertension increases with the duration of obesity, especially
in women, and weight reduction leads to a fall in blood pressure (ibid.).

The reasons for the association between increased weight and hypertension was ‘not
clear’ but the suggestion was that it might be related to higher levels of insulin in the
blood increasing the amount of sodium retained by the kidneys (WHO, 2000: 48), an
example of increasing understanding of the metabolic effects of obesity discussed
above.

106
This refers to the Nurses Health Study.

250
The specific cancers discussed in TR894 were the same as those discussed in TR854
and the evidence from the American Cancer Study was still cited. Increased risks of
endometrial, ovarian, cervical, prostate and post-menopausal breast cancer among the
overweight and obese were thought to be ‘a direct consequence of hormonal changes’
resulting in part from ‘excess’ abdominal fat (WHO, 2000: 48). However it was
difficult to separate out the effects of body size, fat distribution and weight in
assessing relative risks for these cancers:

In addition to overall obesity, intra-abdominal fat distribution and adult weight

gain have been independently associated with an increased risk of breast
cancer. For example, it has been reported that an increase in intra-abdominal
fat accumulation increases the risk of postmenopausal breast cancer,
independently of relative weight and particularly when there is a family
history of the disease. Furthermore, weight gain during adulthood has
consistently been associated with increased risk of breast cancer, even in
cohort studies that showed no association between baseline relative weight and
subsequent risk of breast cancer (WHO, 2000: 48 - 9).

Data from two prospective studies was used to make the striking claims that ‘about
64% of male and 74% of female cases of NIDDM could theoretically have been
prevented if no one had a BMI over 25’ (WHO, 2000: 50). However, the detail of the
evidence provided seems more equivocal, as intra-abdominal fat accumulation ‘in
some studies, has been an even stronger predictor of NIDDM than overall fatness’ and
the prevalence of NIDDM was two to four times higher in the least physically active
individuals, irrespective of body mass (ibid.). Both these pieces of evidence call into
question the precision of the authors’ earlier claims.

Obesity and overweight was now seen as having important effects on a wide range of
hormones because fat had become understood as a metabolically active substance:

Recent research has shown the adipocytes (fat cells) are more than just fat
depots. They also function as endocrine cells, producing many locally and
distantly acting hormones and as target cells for a great many hormones.
Altered hormonal patterns have been observed in obese patients, especially
those with intra-abdominal fat accumulation (WHO, 2000: 51).

The endocrine disturbances associated with intra-abdominal fat accumulation were

held to lead to conditions such as insulin resistance and polycystic ovary syndrome,

251
but the main metabolic consequence of overweight and obesity was seen to be
dyslipidaemia, where blood levels of triglycerides and LDL cholesterol are raised and
those of HDL cholesterol are lowered: ‘This metabolic profile is most often seen in
obese patients with a high accumulation of intra-abdominal fat and has consistently
been related to an increased risk of CHD’ (WHO, 2000: 53). These two types of
disturbances - the metabolic and the endocrinological - were often found in
combination with other heart disease risk factors, such as high blood pressure and
abdominal obesity. In this period, the new diagnosis of ‘metabolic syndrome’ was
beginning to be applied in such situations.107 There was no authoritative international
definition of this condition yet but epidemiological evidence confirmed that it ‘occurs
commonly in a wide variety of ethnic groups including Caucasoids, Afro-Americans,
Mexican Americans, Asian Indians and Chinese, Australian Aborigines, Polynesians
and Micronesians’ (WHO, 2000: 54). This was another new way of framing high rates
of chronic disease in particular population groups (see section 7.4).

A final set of health problems associated with excess body weight was labelled
‘psychological’ or ‘social’. For the authors of the BNF report, four different studies
were available, some of which showed no significant relationship between depression,
anxiety and obesity. However, the severely obese participants in the Swedish Obese
Subjects (SOS) study

showed very poor ratings for mental well-being and anxiety and depression
than the reference population. The scores on psychometric scales were as bad
as, or worse than, those of patients with chronic pain, generalised malignant
melanoma or tetraplegia after neck injury (BNF, 1999: 11).

This was a significant level of psychological distress, demonstrating that severe

obesity was more hazardous to mental health than other degrees of overweight, but
this finding was merely cited and not discussed further. In the case of ‘social
penalties’, the authors stated

There is compelling evidence that our society discriminates against fat people.
This is particularly damaging to the psychological well-being of obese
children, who are believed by their peers at school to be lazy, dirty, stupid,
ugly, liars and cheats…Social discrimination continues into adult life. Sonne-

107
This was also known in this period as syndrome X or the insulin resistance syndrome.

252
 Holm and Sorensen showed that for a given parental social class, intelligence
and education, obese people achieved less than non-obese. Overweight in
adolescence is associated with less social success in later life (BNF, 1999: 12).

In total five different studies were cited to provide evidence for this point, but no
further argument was made. This sort of evidence would have allowed discussion of
the high levels of psychological distress found in the SOS study, by linking it to the
discrimination and prejudice experienced by severely obese individuals. But this was
not done, presumably as the authors felt it was out with their area of expertise (see
chapter 9).

The authors of TR894 also considered psychological problems, including body shape
dissatisfaction and eating disorders, as well as those resulting from dealing with
prejudice and discrimination. Such psychological problems were categorised
differently from the other health consequences:

It is important to note that the mechanisms leading to impaired psychological

health are different from those underlying physical illness. The psychosocial
problems associated with obesity are not the inevitable consequences of
obesity but rather of the culture-bound values by which people view body fat
and “unhealthy” and “ugly” (WHO, 2000: 56).

Whilst explicitly arguing against the ‘negative stereotypes and attitudes of health
professionals’ (ibid.), these authors did not discuss such issues further. Researchers
critical of conventional understandings of obesity contend that biomedical research
and practice is a key site of discrimination against the overweight and obese which
legitimates such attitudes and practices (Bacon, 2008, Campos, 2004, Gaesser, 2002).
Official acknowledgement of the stigmatisation of these individuals required careful
expression in order to forestall such analyses or accusations of further stigmatisation.

7.6 Dissenting accounts of the health effects of overweight and obesity

Within these reports there was no disagreement about the negative health effects of
obesity and overweight – it can be described as a generally accepted consensus. The
findings of Framingham and the American Cancer Study were accepted and

253
incorporated into an explanation of the effects of overweight and obesity that
understood it as having profoundly negative effects on the individual metabolism.

Again, Ernberger and Haskew were the most significant internal dissenting voices that
I have been able to locate. Not only did they dispute the mainstream interpretation of
the evidence about the negative consequences of overweight and obesity, but, more
radically, they were prepared to argue that there are health benefits to both overweight
and obesity:

The beneficial effects of moderate adiposity on longevity and the relatively

modest hazard associated with extreme adiposity present a paradox when
considered alongside the associations between obesity and a variety of
diseases. This has led other researchers to speculate that there may be health
benefits to adiposity [here they give three references] (Ernsberger and
Haskew, 1987: 13).

They argued that greater body weight was associated with decreased mortality rates
for cancer, hypertension and cardiovascular disease (see section 4.7), but also that it
was associated with a lower incidence of respiratory disease, infectious diseases
(notably tuberculosis), bone disease, anaemia and type 1 diabetes, as well as a more
favourable prognosis for NIDDM, hyperlipidaemia, hypertension and rheumatoid
arthritis (Ernsberger and Haskew, 1987: 18). They cited results from specific studies
as evidence for each of these conditions, but I will quote only their argument for
cardiovascular disease, since this is such a classic and canonical example.

Fat people have been shown repeatedly to exhibit elevations in literally every
coronary risk factor known: total cholesterol, ratio of high-density to low-
density cholesterol, uric acid, triglycerides, glucose and insulin…[however]
adiposity is not related to the diseases that these risk factors are purported to
predict. The incidence and extent of atheroschlerosis in coronary and other
vascular beds is entirely unrelated to body fatness as demonstrated in autopsy
studies (Ernsberger and Haskew, 1987: 26).

Furthermore, they also suggested that the positive effects of obesity may be
underestimated in certain conditions because studies do not classify individuals
according to their fat distribution (see section 4.4). According to their argument,
‘nearly all “obesity–related” diseases are associated with accumulation of fat in the
upper body’ and so ‘individuals with predominantly lower body fat may benefit from

254
the protective effects of adiposity but escape the onus of “obesity-related” disease’
(Ernsberger and Haskew, 1987: 24). However, this not to say that adiposity was
beneficial to health:

Even though fatness appears to be protective in some disorders, the overall

impact of significant adiposity (BMI of approximately 30 or greater) is
undeniably negative. However, the severity of the associated risk is probably
less than generally assumed…Furthermore, it is no longer appropriate to
consider obesity as a disease if it has benefits as well as hazards (ibid.).

In this framing, the negative health effects associated with overweight and obesity
were caused by a cyclical pattern of weight loss and regain. Evidence showed that
obese animals did not suffer from hypertension or other cardiovascular ‘abnormality’
which led to the question

Why then is obesity associated with high blood pressure and heart disease in
humans, alone among the animal species? The answer may be that humans are
the only animals that diet to lose weight. When dogs, swine, rats or mice are
repeatedly deprived until they lose 20 percent or more of their body weight,
then are allowed to regain the weight, they develop high blood pressure,
undergo damage to their blood vessels and develop heart disease similar to
that seen in fat humans… However, animal studies fail to duplicate the major
hazards of human obesity, cardiovascular disease, unless the animals are made
to lose and gain weight repeatedly (Ernsberger and Haskew, 1987: 38).

Additional human and animal evidence was cited to support this argument, including
studies of the effects of the Siege of Leningrad, and of 200 fat men on hospital-run
total fasts (Ernsberger and Haskew, 1987: 39). Ernsberger and Haskew were knitting
together laboratory feeding studies on rats with clinical data and large scale study
data, in the same way as other authors, to provide an alternative narrative of the
effects of starvation on human and animal physiology.

However, another dissenting voice was Ancel Keys, who set up the Seven Countries
study (see sections 2.5 and 4.3). He argued that insurance company statistics and data
produced by the first ACS cancer prevention study were seriously flawed:

This is another example of the dangerous idea that serious defects in the
quality of data can be disregarded if the numbers of persons is very
large…The Cancer Society questionnaire, distributed by local non-

255
 professional volunteers, asked the age, sex, height, weight and smoking habits.
For six years Cancer Society volunteers tried to find out who had died and
what was reported as the cause of death. No measurements or medical
examinations were involved. This approach leaves some uncertainties: Are the
answers accurate? Are the subjects representative of the population? Are their
results reliable? (Keys, 1980a: 18).

Because of these flaws, he argued that such data did support the conventional analysis
that overweight increases the risk of heart disease or early death. In fact he used ‘more
sophisticated’ analyses of data from the Seven Countries Study to argue that ‘it is bad
to overweight, worse to be underweight’ and ‘the best prospect of avoiding early
death is to be somewhat above the average in relative weight’ (Keys, 1980a: 20). The
professional interests of a public health scientist like Keys, who pioneered the use of
statistics in health research, could easily lead to attempts to devalue the data collection
methods used by the insurance industry and voluntary bodies such as the ACS.
However, it was an important critique of a body of data fundamental to the
development of obesity science, and the article was reprinted in the British Medical
Journal in 1986. But such criticism, from a major figure in the field of chronic disease
epidemiology, appeared to have little effect on contemporary discussions.

7.7 Conclusion

The understanding of obesity and overweight, and chronic disease more broadly, in
WHO reports changed significantly between the 1970s and the 1990s. The concept of
diseases of affluence was largely abandoned, and it was argued that increasing rates of
these conditions were the result of processes of industrialisation and modernisation
(see section 5.5 and chapter 8). This breaking of the links between chronic disease and
affluence meant that even relatively poor populations could be seen as at risk, and was
a precursor to ideas of the ‘double burden of disease’ developed by researchers
associated with Harvard School of Public Health and the WHO (see chapter 8). It was
also consonant with the new framing of increasing rates of obesity and overweight as
a global phenomenon developed in WHO publications from the mid 1990s.

The negative health effects of overweight and obesity were one of the most important
elements of the arguments of obesity science. In this period, descriptions of the
consequences shifted from earlier discussions focusing on heart disease and

256
mechanical side effects, to a much longer list of metabolic and endocrinological
conditions that drew on more detailed understandings of the physiological processes
associated with excess weight gain. This increased list was crucially dependent on
data from large scale epidemiological studies such as Framingham. Such
understandings were being linked to genetic explanations in an attempt to understand
the very high rates of excess body weight, and chronic diseases, occurring in
populations such as the Pima Indians of Arizona or the Samoans.

257
CHAPTER 8: THE WHO REPORT OF 2000 AND THE ENVIRONMENTAL
PARADIGM OF OBESITY

8.1 Introduction

This chapter continues my analysis of the two WHO technical reports, TR854 and
TR894, although it focuses mostly on the latter. It builds on the previous chapter’s
description of the way increasing rates of obesity were understood as a global
phenomenon to describe how the move into the WHO led to an expanded
understanding of its causes. Mobilising the concerns of international development
central to the work of UN organisations, these gave increasing weight to the effects of
industrialisation and modernisation on average body weights. This chapter also
develops chapter five’s discussion of the limitations of treatment by describing the
growing importance of preventive measures in these reports. I describe the increasing
attention given to the costs of obesity and overweight to healthcare systems and some
of the new data from health economics and related research that was available to the
authors of TR894. I give some details of one of these programmes – the Global
Burden of Disease project – in order to explain the importance of its work to changing
understandings of chronic disease, and ideas of obesity as a global public health
problem. Finally, I give a brief summary of the new understandings of obesity and
overweight – the environmental paradigm – that I have described in this and the
previous chapter.

8.2 A renewed focus on the effects of modernisation and

industrialisation

By the late 1990s obesity science researchers routinely referred to broader, more
social causes of excess body weight (see section 5.4). These were seen to derive from
the significant social changes that had occurred first in developed and then in
developing countries since the Second World War. For example, the authors of
TR854 contrasted affluent societies, where overweight was associated with low
socioeconomic status and education level, to poorer countries where the converse was
true. They argued that there had been a shift in European patterns of male overweight
in the last forty years (see section 6.3). Studies carried out from the 1950s to the

258
1970s showed a positive relationship between obesity and socioeconomic status,
whereas those carried out in the 1980s and 1990s show an inverse association: for
women the majority of studies have shown an inverse relation throughout this period
(WHO, 1995: 318). In less affluent countries:

Overweight may be seen as a visible indicator of wealth and status in countries

where food is scarce. Brazil is an example of a country in which there is a
clear positive association between socioeconomic status…and BMI…It has
been suggested that this association is mediated by the fat content of the diet…

And furthermore,

In some traditional societies, there are pressures on women both to gain

weight and remain overweight during reproductive life. An example of this is
the custom of ‘fattening huts’ for elite pubescent girls in certain communities
in West Africa. Such practices reflect cultural perceptions and values related
to overweight (WHO, 1995: 319).

This brief discussion of the socio-cultural determinants of overweight consisted of

one page of a thirty page chapter. The behavioural determinants were discussed with
similar brevity in this report. They included smoking which was initially associated
with lower BMI but since the 1980s had become linked to other ‘unfavourable’ habits
such as high levels of consumption of alcohol and saturated fat and therefore higher
BMIs. There was seen to be a ‘clustering’ of health damaging habits amongst
particular groups in the population:

in populations in which there is a growing health awareness and an increasing

proportion of people who stop smoking, the remaining smokers are those
whose lifestyles carry significant health risks (WHO, 1995: 320).

Despite a significant body of high profile research in this field (e.g. Townsend and
Davidson, 1982), no mention was made of in this report of the link between such
health behaviours and socioeconomic status.

Five years later in TR894, the causes of obesity were considered first in terms of diet
and physical activity: ‘high-fat energy-dense diets and sedentary lifestyles are the two
factors most strongly associated with increased prevalence of obesity worldwide’
(WHO, 2000: 108). Evidence from laboratory studies and clinical studies was cited as

259
demonstrating that increased amounts of fat in the diet and increased energy intake
were both strongly associated with excess body weight. Because dietary fat has a
higher energy density than other macronutrients, individuals eating a high fat diet
often experienced ‘passive overconsumption’ (see section 5.3):

The body does compensate for the overconsumption of energy from high-fat
foods to some extent, but the fat-induced appetite control signals are thought
to be too weak, or too delayed, to prevent the rapid intake of the energy from a
fatty meal (ibid.).

Eating fatty food was seen as overwhelming the immediate regulatory mechanisms,
and control of intake depended on longer term mechanisms that responded much
better to underfeeding than overfeeding. The body’s capacity for fat storage was
thought to be virtually unlimited, and so when fat balance was disrupted by changes
in energy balance, it was re-established by a change in body fat mass (WHO, 2000:
109). Palatability promoted overeating and fat, especially combined with sugar, was
particularly palatable: ‘Fat appears to be the key macronutrient that undermines the
body’s weight regulatory systems since it is very poorly regulated’ (WHO, 2000:
110).

In the case of physical activity, cross-sectional data had shown that increasing BMI
was associated with decreasing levels of physical activity, but, according to the
authors, it still had not been possible to establish causality. However, this caveat was
quickly ignored to argue that other research demonstrated the role of ‘decreased
physical activity and/or increased sedentary behaviour’ in both weight gain and the
development of obesity (WHO, 2000: 112). For the first time, the authors of TR894
provided a detailed analysis of the contribution of physical activity to total energy
expenditure, and outlined three components of physical activity – occupational work,
household and other chores, leisure-time physical activity (exercise and sport).
Physical inactivity or sedentary behaviour was also defined much more carefully as
including participation in physically passive behaviours such as watching TV,
working on a computer, driving, talking on the phone and eating (WHO, 2000: 113 ).

Another reflection of increased research in this area was the new concept of physical
activity level (PAL), a numerical measure to ‘express daily energy expenditure as a

260
multiple of BMR’ (WHO, 2000: 114). Individuals could now be classified into
different categories based on their average levels of physical activity – a sedentary
lifestyle had a PAL value of 1.4, ‘limited activity’ 1.55 – 1.6 and ‘physically active’
greater than 1.75 (WHO, 2000: 114). However, this narrow numerical range hid a
significant difference in activity levels since to move from the limited activity to the
active category required a daily extra hour of moderate activity108 or half an hour of
vigorous activity (ibid.). To reduce the possibility of becoming overweight,
individuals should be advised to sustain a PAL of 1.75 or above throughout their adult
life (WHO, 2000: 118). However, the authors acknowledged that in industrialised
societies, PALS were ‘drifting downwards’ and that:

People who make extensive and increasing use of motorized transport,

automated work and sedentary pursuits, may find it difficult to attain PAL
levels at or above 1.75 simply by increasing activity during “leisure
time”…Increasing a PAL of 1.58 to one of 1.76 requires approximately 1 hour
and 40 minutes of extra walking (at 4km/h) per day…As these activity
requirements are additional to a 24-minute period of active leisure…already
required for a PAL of 1.58, it follows that urban sedentary populations are
likely to attain a PAL of 1.75 or more only if supported by vigorous national
policies that encourage physical activity (ibid.).

Arguments like this led to a shift from advocating increased leisure time ‘exercise’ to
promoting ‘physical activity’ as an integral part of everyday life (e.g. Foresight,
2007). Dietary composition and physical activity levels were also seen to interact as
regular physical activity led to an increased capacity to burn fat rather than
carbohydrate:

moderate physical exertion allows free-living volunteers to consume ab

libitum a 40% fat diet without storing excess fat, whereas the same
individuals, when sedentary are in positive fat and energy balance and thus
have a greater risk of becoming overweight and obese with time. If, however,
they are offered a 20% fat diet, they remain in balance even when sedentary
(WHO, 2000: 116).

This interaction meant that individuals with high levels of physical activity can
maintain their weight on a high fat diet, whereas those who were sedentary needed to
prevent weight gain by minimising their fat intake. The authors assumed that people

108
Moderate activity is defined as brisk walking, cycling (12km/h) or gardening and vigorous activity
as running, football rugby or hockey (WHO, 2000: 114).

261
in developed countries were mostly sedentary and so needed to maintain a fat intake
of 30% or less (ibid.). Little data was available for developing countries, and the
comparisons that had been made showed few differences from more affluent
countries. Nevertheless, the authors were prepared to argue that:

it is reasonable to conclude that people in less developed countries who spend

a considerable portion of their time in finding food for their next meal and on
personal chores are expending more energy in work and physical activity for a
given body size than those in more developed countries (WHO, 2000: 115).

Common-sense assumptions that increasing average body weights were a direct result
of industrialisation and economic development seem to have overwhelmed the limited
data that did exist.

Another body of theoretical material analyses the genetics of obesity and overweight:

Epidemiological, genetic and molecular studies of populations all over the

world suggest that some people are more susceptible than others to becoming
overweight and obese, and that such susceptible individuals exist in countries
differing widely in lifestyle and environmental conditions (WHO, 2000: 133).

In the case of genetic susceptibility, it did not seem to be that single or multiple gene
effects cause overweight and obesity, rather that ‘the genes involved in weight gain
increase the risk or susceptibility of an individual to the development of obesity when
exposed to an adverse environment’ (WHO, 2000: 134). Obesity was therefore
understood as multi-factorial and polygenic i.e. the result of the interaction of many
genes and their environment (see section 5.5).

Estimates of the heritability of obesity vary widely: the heritability of BMI was
thought ‘likely to be in the range of 25-40%’, the heritability of fat distribution to be
50%, and the heritability in the amount of abdominal fat to be 50 to 60% (WHO,
2000: 134 -5). It is not clear what such a range of numerical values adds to the
understanding of obesity, especially as the authors argued that there was a lack of data
concerning the level of risk of developing obesity for relatives of the overweight and
obesity, despite the knowledge that obesity tended to run in families (ibid.). Various
mechanisms were suggested for the operation of this genetic susceptibility including

262
low resting metabolic rate (studies of the Pima Indians have shown RMR clusters
within families, and that those with lower RMR have greater risks of weight gain);
low fat free mass (which was a risk factor for weight gain since it depressed the level
of RMR); and poor appetite control which the authors link to research into the role of
leptin as a satiety factor (ibid.). Other factors that were seen to promote weight gain
included smoking cessation, excess alcohol intake, drug treatment, genetic and
endocrinological disorders, major reductions in activity levels (due to chronic illness
or disabling injuries) and ‘changes in social and environmental circumstances’ that
include marriage, the birth of a child or a new job (WHO, 2000: 139 - 42). This last
heading was a catch-all category which was too broad to have any significant
explanatory power, but resulted from the exclusion of sociological arguments from
such accounts (see chapter 9).

As policy writing on the public health problem of obesity developed in the late 1990s,
(see above) understandings of the causes of obesity became more detailed and
significantly wider in scope. The authors of TR894 discussed these models under the
heading of ‘environmental and societal influences’ that focused on the wider
environmental causes, especially those deriving from processes of modernisation.
When compared to the headings in TR854’s label of ‘socio-cultural’ and
‘behavioural’ determinants of overweight, this is evidence of a shift to a less
individualistic and more environmental understanding of the causes of increasing
average body weights.

Environmental and societal influences were explicitly linked to changes in social

structures resulting from modernisation and industrialisation. Early in a discussion of
the regulation of body weight, the authors argued that:

In traditional societies, where people tend to be more physically active, and

provided food supplies are not limited, few adults are either underweight or
overweight despite the interaction of seasonal cycles of work, individual
susceptibilities to obesity for physiological or genetic reasons, and the wide
range of varying physical demands within a society (WHO, 2000: 105).

This distinction between ‘modern’ and ‘traditional’ societies played an important

explanatory role throughout this report. The section on environmental influences

263
contains nine separate references to traditional cultures, diets, lifestyles, foods and
values as well as extended discussions of the effects of societal change (WHO, 2000:
119-33).

The authors also framed obesity as a major social problem, by placing it alongside
other long-standing policy concerns such as unemployment, poor housing and
‘broken’ families:

The trend towards industrialization and an economy based on trade within a

global market in most of the developing countries has brought about a number
of improvements in the standard of living and in the services available to the
population. However, it has also had various negative consequences; these
have led, directly and indirectly, to deleterious nutritional and physical activity
patterns that contribute to the development of obesity. Changing societal
structures resulting form this economic transition have given rise to new
problems associated with unemployment, overcrowding and family and
community breakdown (WHO, 2000: 118-9).

Industrial food production was seen, by these authors, to have led to improved food
availability but ‘it has not necessarily solved the problem of undernutrition in many
poorer countries, nor has it improved the nutritional quality of the diets of the
affluent’, while motorised transport and mechanisation have lessened the physical
burden for many individuals but ‘leisure time dominated by television viewing and
other physically inactive pastimes has increased’ (WHO, 2000: 120). Other changes
discussed included modernisation, the transition to market economies, increasing
urbanisation, changes in the role of women, changes in labour markets and the
globalisation of world markets (WHO, 2000: 120-4). The transition to market
economies had led to increased food imports and increased consumption of processed
foods; urbanisation to reduced levels of physical activity and increased consumption
of protein and fat; rising numbers of women in paid employment to the demand for
labour saving devices and convenience foods; changes in labour markets to an
increasing proportion of populations working in sedentary occupations; and
globalisation to increasing competition between multinational companies to sell
processed foods. The underlying assumption was that these changes had resulted in
negative health consequences, but no evidence was produced to substantiate this
argument. Of these factors, modernisation was seen as most important, and given
most explanatory power:

264
 Most adults who still have a ‘traditional’ lifestyle appear to gain little or no
weight with age. Anthropometric studies have reported an absence of obesity
in the few remaining hunter-gatherer populations of the world, since energy
expenditure is generally high and food supplies are scarce in certain periods of
the year…For the majority of the world’s population, however, the process of
‘modernization’ has had a profound effect on the environment and on
lifestyles over the last 50-60 years (WHO, 2000: 120).

Obesity was labelled as ‘the first of the so-called “diseases of civilisation” to emerge’
(ibid.). More broadly, the negative health effects of modernisation were labelled
‘New World syndrome’:

Obesity can be seen as the first wave of a defined cluster of NCDs [non-
communicable diseases] now observed in both developed and developing
countries. This has been called the ‘New World syndrome’ and is already
creating an enormous socioeconomic and public health burden in poorer
countries. High rates of obesity, NIDDM, hypertension, dyslipidaemia and
CVD, coupled with cigarette smoking and alcohol abuse, are closely
associated with the modernization/acculturation process and increasing
affluence. The New World syndrome is responsible for disproportionately
high levels of morbidity and mortality in the newly industrialised countries,
including Eastern Europe, as well as among the ethnic minorities and the
disadvantaged in developed countries. …Thus, while obesity is viewed by
health professionals from a medical perspective, it also needs to be recognized
as a symptom of a much larger global social problem (WHO, 2000: 122).

New World Syndrome was related to ideas of the double burden of disease (see
section 8.6) and appears to have been developed in the mid 1990s. However, the term
was not widely adopted and does not seem to be currently used.109 The shift in dietary
composition and levels of physical activity was also labelled ‘the nutrition transition’
(Drewnoski and Popkin, 1997). This characterised such changes as inevitable, a stage
through which all societies must pass. Unlike the New World Syndrome this concept
is still being used and developed (Popkin and Doak, 1998, Popkin, 2004, Popkin and
Mendez, 2007).

Modern societies seem to be converging on a diet high in saturated fats, sugar

and refined foods and low in fiber – often termed the “Western diet” – and on
lifestyles characterized by lower levels of physical activity. These
developments…are reflected in nutritional outcomes such as changes in

109
A search on Medline found three articles containing the term, all dating from the mid-1990s.

265
 average stature, body composition, and morbidity patterns (Popkin and
Mendez, 2007: 68).

As in the previous WHO report, the effects of modernisation were seen to lead to a
changing relationship between rates of obesity and socioeconomic status. More
information was now available to demonstrate how high socioeconomic status was
negatively correlated with obesity in developed countries, and ‘as the less developed
countries attain higher levels of affluence, the positive relationship between
socioeconomic status and obesity is slowly replaced by the negative correlation seen
in developed countries’ (WHO, 2000: 124). In poor countries thinness was still
associated with lack of sufficient food and the performance of physically demanding
manual work and so ‘thin adults are considered poor, and overweight and obesity are
a sign of affluence’ (WHO, 2000: 125). By contrast in richer countries there was an
inverse relationship between obesity and socioeconomic status, meaning that it was
associated with poverty rather than affluence. Because few people were short of food
and the amount of manual labour performed had decreased significantly, members of
lower socioeconomic groups were not, on average, more physically active or short of
food than more affluent individuals. But there was evidence that poorer families were
less physically active than affluent ones and the poorest households ate cheaper and
more energy-dense processed foods rather than more expensive and nutritious food
such as fruit, vegetables and whole grains (ibid.). The inverse link between education
and average body weight found in industrialised countries (WHO, 2000: 126) was
seen to reinforce this relationship. Social research was beginning to be incorporated,
in order to provide explanations for class differences in rates of obesity and
overweight.

The last major cause of obesity and overweight discussed in this report was ‘cultural
influences’. The authors admitted that these were not well defined or understood, but
then went on to argue that

Cultural factors are among the strongest determinants of food choice. They
include peer group pressures, social conventions, religious practices, the status
assigned to different foods, the influence of other members of the household
and individual lifestyles…Human beings value food for much more than its
nutrient content, and it is used to express relationships between people as well

266
 as in celebrating religious festivities, weddings and other important social
occasions (WHO, 2000: 127).

The vagueness and generality of this discussion demonstrates the limited

understanding of ‘the social’ mobilised in such writing. In particular, culture was
often understood as a barrier to modernisation, so its influence led to a rejection of
recreational physical activity and a valuing of large body size as a sign of prosperity
and health.110 The discussion concluded with a brief outline of changing attitudes to
body shape and weight in industrialised countries and the link between the increasing
denigration of fatness and rising incidence of eating disorders and the use of
‘unhealthy weight control practices’ such as diuretics and self-induced vomiting
(WHO, 2000: 128). These problems were forecast to increase in developing countries
as they adopted the ‘values and ideals’ of industrialised countries.

8.3 Continuing problems with treatment

There were no new treatments for overweight and obesity discussed in TR894: weight
loss diets, physical activity, behavioural therapy, pharmacological treatment and
surgery remained the available options, as they had been in previous reports (see
section 5. 5). However, the ineffectiveness of dieting was becoming more widely
accepted, leading to an increased emphasis on obesity and overweight as life-long
chronic conditions that can be managed, but rarely cured.

The authors of TR894 covered the topic from a clinical perspective to aid the
development of appropriate treatment strategies, making a distinction between the
different techniques used for different aspects of ‘weight management’: weight
maintenance and the prevention of weigh gain require ‘healthier eating’ and a ‘more
active lifestyle’, whereas weight loss requires a temporary negative energy balance
through decreased consumption and increased energy expenditure (WHO, 2000: 206).
A BMI-based classification of individual patients to decide on appropriate
management strategies was also given. A partial outline of this scheme is given in
table 8.1.

110
The latter includes another mention of African ‘fattening huts’ for pre-pubescent girls (WHO, 2000:
127) which is consistent with an old fashioned and colonial understanding of culture as a barrier to
modernisation.

267
BMI Overall Management strategies
health risk
18.5 – 24.9 average Healthy diet and advice on preventing weight gain
25 – 29.9 increased Weight maintenance, healthy diet, exercise
30 – 34.9 moderate Goal of 5-10% weight loss
35 – 39.9 severe Use full therapy (diet, exercise, behaviour therapy) to
achieve > 10% weight loss
BMI 40 very severe Refer to specialist for separate management and
consideration of surgery if treatment fails. Aim for 20-
30% weight reduction

Table 8.1 (adapted from WHO, 2000: 207)

A successful weight management programme should include components of personal

support, dietary assessment and advice, modification of physical activity patterns and
behavioural advice (WHO, 2000: 208). Regular monitoring was described as crucial,
as part of a long-term responsibility for healthcare practitioners: ‘it should not cease
when patients have reached agreed goals but should form part of continuing care’
(ibid.). Obesity was seen here as a chronic condition requiring long-term, maybe even
lifelong, medical supervision.

However, once again there was an argument against returning to an ‘ideal’ body
weight:

[This approach] has for too long been considered by the medical profession to
be both a possible and a mandatory target for obese people. This
misconception has been transmitted to the public, and has been reinforced by
the promotion by the mass media of slenderness as the ideal body image. As a
result, there is now considerable pressure on the overweight individual to
return to his/her ideal, often at the lower end of the normal (18.5-25) BMI
range (WHO, 2000: 201).

This was the most explicit critique of past medical practice in this report. This
changed approach derived from the fact that weight gain was now seen as a health
risk independent of BMI, important health benefits could follow from weight loss of 5
or 10 kg (see section 5.5 ), repeated failure to sustain weight loss might result in
further weight gain and limiting weight gain was important to long term health

268
(WHO, 2000: 201 - 2). The dangers of repeated dieting and the health benefits of
weight stability appear to have become more widely recognised.

As in earlier reports, the authors acknowledged that calorie-restricted diets were not a
very effective treatment for obesity:

Dietary restriction represents the most conventional “treatment” for

overweight and obesity. It usually induces weight loss in the short term, but its
poor long-term effectiveness, especially when used in isolation is widely
recognised…Diets based on healthy eating principles, including the
individualized modest energy-deficit and the ad libitum low-fat diet appear to
have better long-term outcomes (WHO, 2000: 211).

Acceptance of the low success rate of short-term dietary restriction led to the idea of
longer-term modification of eating habits as a necessary treatment for overweight and
obesity. The two approaches mentioned in the above quote – the individualised
modest energy-deficit diet and the ad libitum low fat diet – were still staples of
contemporary weight loss advice. Low fat diets were initially used for their effects on
cardiovascular risk factors, but it was found that patients managed to lose weight on
them: one study found a reduction of 10% in dietary fat produced an average 5 kilo
weight loss in obese patients (WHO, 2000: 212). Other possible dietary treatments
were severe/moderate energy-deficit diets of 1000 to 1200 kcal/day which can
produce significant short-term weight loss, but could lead to insufficient levels of
certain nutrients and weight regain in the longer term, and very low calorie diets
(VLCDs) which could cause rapid weight loss but ‘do not seem particularly
conducive to long-term weight maintenance’ and so should only be used for rapid
short-term weight loss in obese patients for reasons such as preparation for surgery
(WHO, 2000: 213).

The combination of diet and exercise was seen as more effective than either method
alone for promoting fat loss, and the necessary approach was very similar to that
recommend in the 1996 SIGN report (see section 5.6).

Physical activity strategies should aim at encouraging higher levels of low-

intensity activity and reducing the amount of leisure time spent in sedentary
pursuits. The main aim is to convert inactive children and adults to a pattern of
“active” living (WHO, 2000: 214).

269
‘Appropriate’ levels of physical activity could be achieved by increasing activities
such as walking or cycling by up to three hours daily, or by undertaking 45 to 60
minute sessions of fitness training three times a week (WHO, 2000: 215). Because of
the perceived benefits of ‘modest’ exercise, and its acceptability to patients, walking
was recommended as a way of getting overweight and obese patients to increase their
physical activity levels (ibid.). Physical activity was also seen as important in the
prevention of age-related weight gain (WHO, 2000: 320).

As in all the recent reports, behaviour modification was seen as an essential element
of obesity treatment (WHO, 2000: 215). The core features of such a programme were
self-monitoring, stimulus control, an emphasis on improved nutrition (‘rigid dieting is
discouraged in favour of balanced and flexible food choices’), cognitive restructuring,
improved interpersonal skills and relapse prevention (WHO, 2000: 216). However,
the initial enthusiasm about behaviour modification evident in the early reports (see
section 5.5) has now faded:

Behavioural treatment has been more intensively researched, and its effects
more thoroughly documented, than any other obesity intervention. It is
effective in changing behaviour in the short term and consistently produces
significant weight loss in patients with mild to moderate obesity. In the long
term, however, results are not encouraging, virtually all adult patients
returning to their pre-treatment baseline within 5 years (ibid.).

This failure was seen as requiring further research into ways of making such methods
more effective, and as an argument for the treatment of obesity as a lifelong
endeavour. Just as the patient category was widened to include all those ‘at risk’ of
obesity or the ‘pre-obese’ (see section 4.4) so the treatment became not a matter of a
few months, or years, but of a lifetime.

Drug treatment was described as ‘controversial, largely because of failure to

understand how it should be used’ (WHO, 2000: 217). However,

Due to the paucity of data, no particular strategy or drug can yet be

recommended for routine use. However, the availability of new evidence of
long-term efficacy and safety of several drugs currently awaiting approval is
likely to change the situation (ibid.).

270
In spite of the regularity with which anti-obesity drugs have been taken off the
market due to unacceptable side effects (Ernsberger and Haskew, 1987), such
optimism about exciting new drug treatments has been regularly re-stated. Drugs
should be used in the treatment of patients with a BMI of 30 in conjunction with
diet, exercise and behavioural regimes, as part of long-term treatment programme,
under medical supervision and continued only if the patient has managed a weight
loss of at least 10% (WHO, 2000: 217-8).111 These principles repeated a well
established orthodoxy that had been present throughout obesity science (see section
5.5). The four available weight management drugs were ephedrine and caffeine,
tetrahydrolipstatin, phentermine and sibutramine (WHO, 2000: 219). The controversy
over the safety of fenfluramine and dexfenfluramine was mentioned (WHO, 2000:
218), but phentermine (see section 5.5) was still listed as an appropriate drug to
prescribe. Despite the side effects listed for each drug,112 the issue of long-term usage
of such substances was addressed only in a relatively bland paragraph stating that

While the clinical tolerance of most drugs appears to be acceptable, their long-
term use raises some safety concerns…As with drugs prescribed for long term
treatment in other chronic diseases (e.g. hypertension, NIDDM), the risk
associated with long-term drug use for weight management must be weighed
against the potential benefits for each individual (WHO, 2000: 221).

This argument relies on giving obesity the status of a chronic disease like high blood
pressure or diabetes. If it is not framed in this way then the hazards of drug treatment
might be seen as unacceptable by both doctors and patients. If critics such as Keys
and Ernsberger and Haskew were correct and only significant obesity was dangerous
to health, then the side effects and risks of long term drug treatment for excess weight
could not have been equated with relatively successful drug treatments for chronic
conditions such as hypertension or diabetes.

The final type of treatment discussed was gastric surgery:

111
The authors admitted that this last criterion has been described as ‘unrealistic in most cases’ but
described it as a current UK guideline.
112
These include abdominal pain, flatulence, insomnia, irritability, agitation, tension, anxiety, nausea,
constipation, dizziness and increased blood pressure (WHO 2000: 219-20).

271
 Surgery is now considered to be the most effective way of reducing weight,
and maintaining weight loss, in severely obese (BMI>35) and very severely
obese (BMI>40) subjects. On the basis of cost/kg of weight lost, surgical
treatment has been estimated, after 4 years, to be less expensive than any other
treatment (ibid.).

Given the low success rates for other treatments (see above), it is not surprising that
surgery was seen as the most effective treatment for these patients. Whether its
relatively low cost outweighs the significant risks to patients was an issue that was not
directly addressed here. Vertical-banded gastroplasty and Roux-en Y gastric bypass,
were considered ‘effective and safe’ after studies that followed patients for more than
15 years (WHO, 2000: 221-2). Such treatment was suitable for patients with a BMI
greater than 40,113 who should only be considered for such treatment when other
treatment have not worked, and should be ‘well informed and motivated with
acceptable operative risks’ (ibid.). According to the authors, the Swedish Obese
Subjects study showed that weight loss of 30 to 40 kg usually occurs in the two years
after surgery, and

surgical treatment produced remission of NIDDM in 68% of obese patients

and of hypertension in 43%. For those who did not have risk factors at
baseline, a 30-kg weight loss was associated with a 14-fold risk reduction for
NIDDM, and 3-4-fold risk reductions with respect to the development of
hypertension and cardiovascular risk factors (WHO, 2000: 222).

This list of improvements is impressive, but many of them describe reductions in risk
factors rather than diseases themselves. However, the risks of such surgery were also
significant and included micronutrient deficiencies, neuropathy, post-operative
complications, ‘dumping syndrome’ and post-operative depression (WHO, 2000:
223). The authors argued that most of these complications could be treated with
behavioural therapy, and, without citing any supporting evidence, that ‘Operative
mortality in experienced centre is a fraction of the mortality observed in unoperated
patients’ (ibid.). Developments in surgical techniques, combined with the perceived
failure of other treatment methods, was leading to a situation where surgery was seen
as the most effective treatment option for obesity, if not (yet) overweight.

8.4 The growing importance of prevention in the 1980s and 1990s

113
Or ones with a BMI of 35 plus ‘high-risk, life-threatening comorbid conditions’ (WHO, 2000: 222).

272
The ineffectiveness of available treatments combined with projection of rapidly rising
rates of obesity and overweight, led to a stress on prevention. As excess body weight
was seen as difficult to treat once developed and affecting increasing numbers of the
population, so prevention became a logical focus for public health policy. This topic
was not written about regularly in early obesity science, but by the 1980s some
authors were discussing it. Philip James stressed the need for prevention, as well as
treatment:

A primary care physician in Britain can expect at least a third of his adult
patients to be overweight and this constitutes an extraordinary workload if he
decides to undertake treatment individually. Therefore the problem is often
ignored. Yet recognizing that an individual has Grade I obesity could trigger a
whole strategy of simple management which would prevent a worsening of the
condition, highlight problems of lifestyle which might need to be corrected
and alert the physician to the need for preventive action with other members of
the family (James, 1984: 636).

A 1987 British textbook (Bender and Brookes, 1987) contained an entire section on
prevention – one of the earliest textbooks to do so – which included chapters on
obesity as a population issue by epidemiologist Geoffrey Rose (see section 7.2), the
role of exercise, Norwegian nutrition policy and the North Karelia project. Rose
argued that the problem of obesity was an ‘adverse shift of the whole risk factor
distribution’ and therefore potential solutions must ‘correct the underlying causes of
this shift’ which were low average energy expenditure, changes in the availability and
nature of food and social attitudes’ (Rose, 1987: 181). His recommendations were that
overweight should be prevented from occurring in childhood and early adulthood, that
individuals should maintain a ‘healthy’ weight until retirement by increasing their
energy expenditure, and that weight loss should be encouraged only in those with
other risk factors such as hypertension, a family history of heart disease or diabetes or
the severely overweight (those in the top 20% of the weight distribution) (Rose, 1987:
186).

The chapter on Norwegian nutrition policy summarises the country’s setting of

dietary goals to reduce the average amount of fat in the population’s diet and increase

273
the amount of starchy foods and fibre, using various types of subsidies. The author
argued that

The consumption of cereals, vegetables, fruit, fish and low-fat milk has
increased. Furthermore, there has been a reduction in the intake of margarine
and whole milk. However, there has been no change in consumption of
potatoes, meat and sugar…Concomitant with the changes in the national diet
there has been a reduction in deaths from coronary heart diseases and related
disorders…It is tempting to believe that the Norwegian Nutrition and Food
Policy may be one of the reasons for this decline in heart diseases (Norum,
1987: 209).

This sort of government policy was possible only in a country like Norway where the
government sets the prices for a significant number of food stuffs and there was some
consensus about its level of involvement in food production, manufacture and
retailing.

The final chapter discussed the North Karelia Project as a practical example of a
community-based health education programme (see section 2.5).114 Using Rose’s
ideas the authors argued that:

The central aim is to influence the risk factor levels of the community through
comprehensive intervention, using the available community structures and
naturally occurring interactions. This approach will be more effective in
reducing the community disease rates than would a restricted, even intense
intervention among high-risk people alone, because most cases of disease
come not from the relatively few people with high risk factor levels but from
the large majority of the population having moderately elevated levels, but
usually several risk factors. Risk-related behaviour is deeply embedded in
ways of life and in the social and even physical environments. Thus, to
influence permanently the health-related lifestyles of large numbers of people
must involve a general process of social change (Puska et al., 1987: 211).

This is the logic for population wide prevention of overweight and obesity, rather than
prevention at a purely individual level. It also acknowledged, if only in passing, the
ways in which habits of eating and physical activity were socially embedded.

114
Despite the fact that it was set up as a preventive programme aimed at cardiovascular disease in
general, rather than as an obesity prevention programme, North Karelia is commonly cited as the only
successful example of an obesity prevention programme.

274
The 1995 Health of the Nation report contains the next extended discussion of
prevention: as part of a ‘population approach’ to obesity, the majority of its
recommendations were preventive measures. These recommendations targeted key
organisations. For example, the NHS should improve nutrition training for health
professionals and develop schemes to prescribe exercise; local authorities should
increase provision of safe cycle routes and footpaths and make facilities for physical
activities more accessible to those on low incomes; schools should develop
programmes for safe walking and cycling to school and devote more teaching time to
PE; and the food industry should develop educational resources to explain the idea of
energy balance (DoH, 1995: 21). Whilst these were worthwhile activities, all were
framed as voluntary and none addressed socio-economic structural issues around diet
and physical activity.

The authors of the SIGN report argued that prevention was necessary to decrease the
increasing burden on ill-health that rising rates of obesity will generate:

In most countries the rate of obesity is doubling every 5-10 years. This means
that in a Scottish practice of 10,000 perhaps 80 extra patients will become
obese every year unless steps are taken to prevent this (SIGN, 1996: 27).

A preventive approach was recommended despite the fact that there was no evidence
of what measures should be used to prevent obesity and no formal trials of obesity
prevention had yet been carried out (ibid.). The author was limited to discussing
healthy eating advice, and the desirability of providing weight management advice to
those in situations, such as enforced immobility, steroid therapy and smoking
cessation, that led to weight gain.

The authors of the BNF report also began their discussion of the topic by stating that,
despite the simplicity and obviousness of the solutions to obesity, there was very little
evidence of the effectiveness of health education behaviours in promoting such
behaviours (BNF, 1999: 139). The traditional health education approach, used in the
1992 UK Health of the Nation White Paper, was not seen to be working:

275
 This met with a resounding lack of success. The public received and
understood the message, but the prevalence of obesity continued to increase
(BNF, 1999: 141).

The other example of a population wide prevention programme discussed was the
Minnesota Heart Health Programme, which, as its title suggests, was designed to
reduce cardiovascular risk rather than obesity per se. This programme also showed no
effect on the average BMI of participants (ibid.). Alternatives to these approaches
mentioned in the BNF report were individual incentives such as financial sanctions
for overweight and obesity or an ‘ecological’ approach which seeks to alter the wider
environment ‘to promote lower energy intake and greater physical activity’ (BNF,
1999: 144). Referring to the SIGN guidelines, the authors argued that healthcare
professionals had begun to address the issue of prevention, but that,

At present, opinion formers in the media see obesity as mainly a social and
cosmetic problem, not a serious public health hazard…Progress on prevention
will not be made until obese people are seen to have a physical disability, for
which they need help from both health-carers and from society in general
(ibid.).

The understanding of obesity as a physical disability does not seem to have become
widely accepted, perhaps due to the persistence of moralistic ideas that it results from
excessive consumption due to poor self-control.

8.5 Prevention in WHO Technical Report 894

The increasing stress on prevention in obesity science, which began in the 1990s,
developed out of projects such as North Karelia, but also fed back into WHO reports.
By the publication of TR894, obesity and overweight was being understood as a
global problem (see section 7.3) concerning the health of populations. This, combined
with acknowledgement of the limitations of treatments, meant that prevention
techniques became seen as important strategies, and were discussed in more detail.
The problems with current individual approaches were that

The benefit of nutritional knowledge per se appears to be limited. Surveys

indicate that, although some people know what constitutes a healthy diet, they
prefer in practice to consume a relatively unhealthy one…Obesity rates

276
 continue to climb, despite the increased frequency of dieting among obese
people, suggesting that knowledge and frequent attempts to slim are
insufficient for successful weight control. However, without these widespread
attempts to control body weight, the prevalence of obesity in industrialized
countries might be much higher (WHO, 2000: 126).

In a skilful rhetorical turn, the problem of widespread lack of success in dieting

became fresh evidence for the importance of the issue as a public health problem.
This led to arguments for preventative approaches that were framed in the context of
increasing obesity rates and newly identified environmental causes (see section 8.2).
Prevention became seen as an integral element of health professionals’ management
of overweight and obese patients:

Until recently, obesity prevention and obesity management were perceived as

two distinct processes, the former being aimed at preventing weight gain and
the latter concerned with weight loss. Management was seen as the role of the
clinician, whereas prevention was considered to be the domain of health
promotion or the public health departments. However, it is now realized that
obesity management covers a whole range of long-term strategies ranging
from prevention, through weight maintenance and the management of obesity
comorbidities, to weight loss…The individual strategies are interdependent, so
that truly effective obesity management must address all of them in a
coordinated manner (WHO, 2000: 156).

This co-ordinated approach could be seen as a group of clinical medical researchers

trying to extend the use of their concepts and methods into areas of public health
medicine. In the area of diet, prevention could be seen as the concern of several other
professional groups including those working in health promotion, primary care and
public health. Such fragmentation allowed new groups, like the obesity science
coalition, to also claim ownership of the issue. At the same time the aims of obesity
prevention were expanding beyond preventing normal-weight individuals becoming
obese, to preventing the development of overweight in these individuals, preventing
the overweight from becoming obese, and preventing weight regain in those who had
lost weight (WHO, 2000: 158 -9). Such interventions were also seen to operate at
several different levels - universal or public health prevention directed at the whole
community, selective prevention directed at high-risk individuals and groups, and
targeted prevention directed at those with existing weight problems (WHO, 2000:
160). This multi-level perspective increased the scope of such measures and,
therefore, their encroachment into the public health domain.

277
The authors admitted that there had been little research on the effectiveness of such
strategies and increasing prevalences ‘cast doubt on whether it is even possible to
prevent excessive gains in body weight in the long term’ (WHO, 2000: 157). This
lack of research and doubt about effectiveness led to detailed discussion of a ‘public
health approach’ to overweight and obesity:

A public health approach to obesity concentrates on the weight status of the

populations as a whole, in contrast to interventions that deal exclusively with
factors influencing the body fatness of individuals. In many developed and
developing countries, underprivileged minority groups have to bear a
disproportionately heavy burden of higher than average levels of obesity.
Thus, in efforts to remove inequalities of health status as one of the main aims
of public health, it is necessary to consider the causes that make particular
groups more vulnerable to weight gain (WHO, 2000: 174).

This statement contains one of the few mentions of health inequalities in this report
where public health was seen to be able to ‘remove’ health inequalities. This
optimism seems profoundly at odds with the earlier description of the environmental
causes of overweight and obesity.

One of the arguments for such a population-based approach comes from the
relationship found between the average BMI of a population and increases in the
prevalence of obesity in that population

[Geoffrey] Rose115 found that a 4.66% increase in the prevalence of obesity

for every single increase in the population’s average BMI above 23, resulting
in a strong correlation between the average adult BMI of a population and the
proportion of adults with obesity…In the United Kingdom between 1980 and
1993, the mean BMI increased from 24.3 to 25.9 for men and from 23.9 to
25.7 for women. Over the same period, the rates of overweight increased by
one-third, whereas those of obesity doubled. This implies that further increases
in mean BMI are likely to result in even more dramatic rises in the rate of
obesity (WHO, 2000: 178).

This argument led to the concept of optimum population BMI: to minimise the
numbers of both underweight and obese adults an average BMI of 23 was thought to
be optimum, whereas to minimise only the numbers of overweight an average BMI of

115
Rose is a British epidemiologist (see section 7.2).

278
21 was thought to be optimum (WHO, 2000: 178-9). Giving such figures seems futile
in the face of earlier descriptions of relentless increases in average bodyweights
caused by widespread environmental change.

However, the social differentiated nature of obesity rates provided indirect evidence
‘that there are environmental conditions as well as genetic factors that can protect
populations’ (WHO, 2000: 157-8). This was a strangely circular argument that used
the evidence for the public health problem of obesity to also argue for potential
preventive solutions. Data from American and Finnish116 populations was given as
evidence that individuals of higher social class showed smaller weight gains than
those of lower social classes and that these lower weight gains might be levelling off;
a parallel was drawn between obesity and other epidemics of chronic disease, such as
heart disease, which were ‘abating’ in countries where preventive measures had been
implemented (WHO, 2000: 158). However, far from showing that such measures
were effective, such figures might just have illustrated an exacerbation of existing
health inequalities, as affluent groups were more able to follow preventive health
advice and their mortality and morbidity rates were declining accordingly. Obesity
and overweight were in the process of becoming even more strongly associated with
poverty and disadvantage.

The kinds of public health interventions envisaged by the WHO report were either
those that aimed to educate and improve the skills of individuals, or those that aimed
to modify the environmental causes of obesity. Interventions of the first kind had not
generally been very successful:

Communities are generally well aware of the problems associated with

obesity, and many individuals are actively attempting to control their weight.
Participation rates in these kinds of programmes are usually high, and many
succeed in reducing their weight in the short term. Nevertheless, there is
generally little impact on the overall average BMI of the community and a
negligible effect on obesity prevalence, so that preventive strategies are
obviously of great importance (WHO, 2000: 180).

The second kind of intervention was, therefore, seen as potentially more ‘effective’,
but, ‘there have not been any well evaluated and properly organized public health

116
The Finnish figures are derived from the North Karelia study.

279
programmes aimed at the population-level management or prevention of obesity’
(WHO, 2000: 183). Singapore was the only country that had attempted to tackle rising
obesity rates by means of a national healthy lifestyle programme targeting particular
population groups.

The Trim and Fit programme was launched in 1992 and is aimed at all
children in Singapore. It combines progressive nutrition changes in school
catering, and nutrition education together with regular physical activity in
schools…Recent results indicate that the number of children successfully
completing the fitness tests annually is increasing, and that obesity rates fell
from 14.3% in 1992 to 10.9% in 1995 for primary students, from 14.1% to
10.9% for secondary students, and from 10.8% to 6.1% for junior college
students…However, it should be noted that this decline in obesity rates may
have been somewhat exaggerated because of the new weight-for-height norms
introduced by the Ministry of Health in 1993 (WHO, 2000: 184).

Singapore is a small city state with an authoritarian government that was willing to
use legislation to alter the personal behaviour of its citizens. Presumably, this was an
important factor in the programme’s success and makes it unlikely that the results
could be replicated in other countries. This was the only national intervention that the
authors could cite and the rest of the discussion used examples of heart disease
prevention programmes117 which included reduction of BMI as one of their outcomes.
These all showed reductions in CHD risk factors in treatment groups, but only one
showed any effect on obesity rates. In fact, in one

BMI showed a strong secular increase despite such innovative weight-control

programmes as adult education classes, a workplace weight-control
programme, weight loss by correspondence course and a weight-gain-
prevention programme (WHO, 2000: 185).

The North Karelia Project was repeatedly cited as a successful anti-obesity

programme, yet ‘Despite remarkable reductions in CHD risk factors, which were still
declining in 1992…the average BMI and the level of obesity remained similar
throughout the project’ (WHO, 2000: 185 - 6).

117
These included the Stanford Three Communities Project, the Stanford Five Cities Study, the
Minnesota Heart Health Programme and the North Karelia Project.

280
Various reasons were suggested for this lack of success including a focus on heart
disease risk rather than obesity, the failure to reach a sufficiently large proportion of
the population and trying to target too many behaviour changes at once. But the one
reason that showed the most congruence with the causes outlined in other parts of the
documents was that:

Powerful societal and environmental obesity-promoting factors have

developed rapidly in many societies in the last few decades, and the
intervention programmes have not been strong enough or sufficiently well
coordinated to overcome them (WHO, 2000: 186).

Such arguments make it difficult to see how obesity prevention programmes based
solely on changing individual behaviour could ever be successful. Conceding this
point, the authors argued instead for programmes that ‘reduce the exposure of the
population to obesity-promoting agents by concentrating on environmental factors
such as transportation, urban design, advertising and food pricing that promote the
availability of high-fat, energy-dense diets and physical inactivity’ (WHO, 2000:
190).

Population level interventions were seen as a ‘shared responsibility’ that should

‘involve the active participation of governments, the food industry/trade, the media
and consumers’ (WHO, 2000: 167). However, the effects of government activities
were described as contradictory: in principle, governments’ regulatory activities
should contribute to the health of the population. However,

Modernisation and the competing demands of economic development and

health have sometimes created a situation where actions by governments have
contributed to a decrease in physical activity and an increase in the intake of
energy-dense food, contrary to their own health guidelines (WHO, 2000: 128).

Recognition of this contradiction did not to lead to any discussion of the problems
inherent in calling for government action to provide public health solutions (see table
8.2 below). The effect of the food industry could also, at best, be labelled
contradictory, but this issue was also not addressed. Instead the authors described how
technological advances increased food availability, but in carefully neutral language
then argued that:

281
 Advances in food technology have also contributed to the consumption of
diets increasingly dependent on processed foods…food characteristics are
often manipulated to such an extent that it is difficult for individuals to
associate visual, textural or taste cues with the energy content of meals. This is
especially important given the increasing trend towards prepackaged foods
and the concomitant decline in the use of natural and basic foods in food
preparation in the home. Consumers are losing control over the preparation of
the foods that they eat, and food composition is increasingly being placed in
the hands of manufacturers (WHO, 2000: 131).

Very little evidence was produced for this claim – just one reference to a report on
trends in consumer food choice – and its appeal to nostalgia was also part of the link
made, again without any evidence, between rising obesity rates, decreasing
consumption of home prepared foods and increasing consumption of take-away foods
(WHO, 2000: 132). A stress on the role of consumer choice also blunted the impact of
a discussion of the role of marketing and advertising in the selling of highly processed
foods (WHO, 2000: 133).

Such considerations led to the list of possible strategies reproduced in Table 8.2. A
substantial proportion of the entries relate to consumer education, presumably because
it was one the most politically palatable options for an organisation like the WHO to
recommend in a consensus report.

1/ & $/% 2 34! /1 4/''$5! ' ,$ '

* $ ,
+ % "
$ % " " %
- .$. "
# %% % % $
.$. % / %
%% $ % % " $ $
0 $
% $ $ $ %
' $ %%
$ ' $ $
1 % % $ % %
$
2% " " $ "
# 2 " % 2 $ $% %

( # / % $ $ $ $

1 % %%
1 $ % " %

) $ $ % % $
.$. " $

282
# # % $ " $ %% %
% $ %
% $ 2
' $
* $ .$. $
1 %
$ " % $ $
1 % $
2
1 % %
, %
) % 1 $ % % 3% 4$ $ % $ $

Table 8.2 (WHO, 2000: 193 - 4)

The discussion ended with another call for structural change addressed to developing
and newly industrialised countries that, again, calls for action going well beyond the
list given above.

As in developed countries, obesity in the developing and newly industrialized

countries will not be prevented simply by telling individuals and communities
to change their diet and exercise behaviours. What is needed is a radical
improvement in the social, cultural and economic environment through the
combined efforts of government, the food industry, the media, communities
and individuals (WHO, 2000: 193 - 4).

This strikingly depoliticised view of potential solutions recurs throughout much

subsequent policy writing in this area (e.g. Foresight, 2007). It contrasted with the
much more mundane and pragmatic nature of the measures advocated, such as those
outlined in table 8.2. These two approaches co-existed in such documents without
their contradictions causing problems because of their different functions (see chapter
9).

8.6 Estimating the costs of obesity and the growth of health economics

The argument that obesity and overweight gives rise to substantial costs to health care
systems had a dual function within obesity science writing. Firstly, it established
obesity and overweight as a serious health problem, legitimating the expertise of
researchers and clinicians who work in this area. Secondly, it formed an important

283
part of the economic argument for both prevention and treatment. Figures estimating
the large costs of obesity and overweight have been a notable feature of the obesity
epidemic discourse of the last ten years (WHO, 2000, Health Committee, 2004,
Foresight, 2007). Previously, they had been included only in the reports published by
the Organisation for Health Economics, presumably due to its specialist interest in the
topic (OHE, 1969, West, 1994).

However, marked differences in the ways the two OHE reports described these costs
demonstrate that there had been significant developments in methods of analysing and
forecasting the economic impact of chronic diseases. This growth occurred as part of
the growth of the academic discipline of health economics. A study of the field,
carried out in the late 1980s, argued that ‘there was no community of health
economists in Britain before 1970; even though there were undoubtedly several
academic economists at that time who were professionally interested in health’
(Ashmore et al., 1989: 5). The authors argued that an examination of academic
discourse, textbooks, courses and research centres and professional organisations and
conferences showed that by the period of their study ‘an organised community has
come into existence’ (ibid.). The existence of such a community is further
demonstrated by the increasing availability of economic analyses of healthcare
expenditure and their growing sophistication.

In contrast to detailed modern accounts of the enormous expenses incurred by treating

obesity, the author of the 1969 OHE report used a very simple breakdown and
described the costs of obesity to the NHS as relatively insignificant (see section
3.4.6). Twenty five years later, when the second OHE report was published, much
more detailed figures were available. The direct costs to the NHS were estimated to
be £29.3 million (West, 1994: 38). This figure was calculated from estimates of the
costs of GP practice consultations (which make up nearly half of this figure), in- and
out-patient hospital treatments, prescriptions for appetite suppressants and treatment
by dieticians (West, 1994: 38-9). However, West went on to argue that:

The additional costs to the National Health Service for treating conditions
triggered by obesity will, however, be much larger and would include a
proportion of the cost involved in post-operative care of surgical patients,

284
 treating heart disease, stroke, diabetes, some cancers, osteoarthritis and
hypertension among others (West, 1994: 39).

The negative health consequences of overweight and obesity (see section 7. 5) meant
that the costs of overweight and obesity were much higher than these direct costs.
Calculations of cost for various conditions associated with overweight and obesity–
heart attack, stroke, NIDDM, arthritis and hypertension – and percentages due to
obesity were used to estimate the ‘obesity cost’ which ranged from £100 million for
diabetes to £7.75 million for heart attacks (West, 1994: 41). Approximately 10% of
the overall cost arising from these conditions was due to overweight and obesity, but
this varied from 5% for stroke to 80% for diabetes (ibid.). In total the annual costs of
obesity to the NHS in the late 1980s were estimated to be £165 million (West, 1994:
40-1). These were not the only costs that could be attributed to obesity and
overweight – British spending on meal replacements and diet magazines, only a
couple of the available products, amounted to £80 million and £5 million (West,
1994: 41).

West used as a model for his calculations an article published by Graham Colditz in
1992 (Colditz, 1992). In this article, Colditz estimated the total costs of obesity to the
US economy in 1986 to be between $39 and $56 billion dollars. He linked obesity to
an increased risk of diabetes, hypertension, CVD, gallbladder disease and certain
cancers to calculate these figures, and included both the direct costs to the healthcare
system and the indirect costs to the American economy such as loss of output and
productivity due to morbidity and mortality resulting from overweight and obesity.
This initial calculation produced the figure of $39 million, but West argued that this
‘may be an underestimate since several cancers have been omitted as have
musculoskeletal disorders such as osteoarthritis’ and if these were included ‘the total
costs of obesity amount to 7.8 per cent (or $56.3 billion) of the total costs of illness in
the US’ (West, 1994: 39-40). These percentages were derived using disease-specific
estimates of American healthcare costs and then assigning a percentage of causality to
overweight and obesity on the basis of epidemiological studies such as the Nurse
Health Study. The prevalence of diabetes is closely associated with increasing
bodyweight and so 57% of the healthcare costs due to diabetes were attributed to
obesity, whereas the figure for gall balder disease was 30%, for hypertension 77%, for

285
CVD 19%, for breast cancer and colon cancer combined 2.5% (Colditz, 1992: 504S -
506S). The ‘geological’ layering of assumptions involved in the production of such
estimates makes them vulnerable to gaps and inaccuracies in the original data and
means that they contained larger elements of uncertainty than were usually
acknowledge when they were quoted in different contexts (MacKenzie, 1990).

By the publication of TR 894, there was a well developed analysis of the costs of
obesity that was given an entire chapter of this report. This illustrates the continuing
growth of the relatively new discipline of health economics and the development of
new techniques for analysing the costs of healthcare, particular those associated with
specific conditions. Using a rhetorical technique that was very characteristic of these
reports, the authors begin this chapter by stating that

To date, there have been only a few attempts to quantify the economic burden
of obesity-related morbidity and mortality. This is in marked contrast to
smoking and alcohol consumption, where a large number of international
studies have been undertaken to determine the magnitude of the economic
burden that they impose on the community (WHO, 2000: 78).

This statement contains the familiar device of proclaiming the paucity of evidence
whilst using it to make firm arguments. It also continues the comparison between
overweight and obesity and alcohol consumption and smoking that I have discussed
earlier (see sections 5.5 and 6.5). In the context of economic costs, this comparison
was particularly important because it framed excess body weight as a problem of self-
discipline and therefore an illegitimate reason for increased healthcare expenditure
(Petersen and Lupton, 1996).

Total costs attributed to obesity in developed countries began to increase sharply: in

TR894, the authors estimated them at between 2% to 7% of total healthcare costs
(WHO, 2000: 84). These percentages translated into estimated direct costs of 464
million dollars (Australia 1989/90), 1 000 million gilders (Netherlands 1981-9) 12
000 million francs (France 1992) and 45 8000 million dollars (USA 1994) (WHO,
2000: 83). The American costs derived from the 1992 Colditz study (see above) and a
revised estimate published in 1994 that increased the number of obesity-associated
diseases included in the analysis. Despite this expansion the authors still argued that

286
this estimate ‘should still be considered conservative’ (WHO, 2000: 86). The first
Colditz study was replicated to produce the estimates for the Australian healthcare
system, whereas the Finnish, French and Dutch studies used different methods.

The scope and methodology of the various studies vary considerably in terms
of the diseases costed, the definition of obesity, the cost categories used and
the epidemiological assumptions as to the relationship between obesity and
disease (WHO, 2000: 83 - 84).

Reliable comparisons between different countries were difficult to make, but such
caveats, along with considerations of the uncertainties inherent in producing such
estimates, were included as these figures became reproduced in different contexts.
Like lists of health consequences (see section 7.5) such figures were stripped of their
uncertainties and caveats as they moved from one context to another by being
repeatedly cited in research articles and policy documents. This parallels processes
undergone by other forms of scientific knowledge which are described in the image of
the certainty trough (MacKenzie, 1990), where those who are nearest to and furthest
from the knowledge production processes are least certain about its accuracy, whereas
those in the middle, who are disseminating the information, are most certain of its
accuracy.

The growth in health economics is shown by the number of new techniques that were
now used to analyse the costs of obesity. The authors of TR894 focused most of their
discussion on the cost of illness study but also referred to the disability-adjusted year
(DALY) in their attempt to quantify these costs (see section 8.6). However, the
DALY had not at that point been used to estimate the costs of obesity118 – and, again,
the authors used the parallel with smoking to argue for further research to be done in
this area:

Obesity and overweight, in the same way as tobacco use, contribute to several
NCDs. Thus, the total DALY loss attributable to obesity and overweight
would represent the attributable fraction of the total loss of DALYs due to
NCDs associated with excess body weight. A number of estimates of the
attributable fraction associated with tobacco use have been made, thus
facilitating national and regional comparisons. Efforts should therefore be

118
The Global Burden of Disease analysis (see section87.6) considered malnutrition, hypertension and
physical inactivity as causes of disease but not obesity or overweight (Murray and Lopez, 1996: 28).

287
 made to generate similar estimates of the attributable fraction associated with
obesity and overweight (WHO, 2000: 83).

The main technique that was discussed in TR894 is the ‘cost-of-illness’ studies. Such
studies calculate the direct cost to the community (healthcare costs of obesity and
related diseases), intangible costs (cost to the individual in terms of health and quality
of life) and the indirect costs (welfare and economic benefits lost by other members of
society) using measures such as PAF (population attributable fraction) or DALY
(WHO, 2000: 78 - 9). The PAF was defined as ‘the proportion of total events (e.g.
death or morbidity) in a population that could be prevented if a particular risk factor
e.g. obesity could be eliminated’ or more simply the ‘preventable proportion’ (WHO,
2000: 81-2). It was calculated using the prevalence of a risk factor such as obesity in a
population and the relative risk of the incidence of a particular associated disease in
an obese person, compared to that of a non-obese person, and its accuracy was highly
dependent on estimates of both the prevalence of the risk factor and the relative risk
of the disease in particular populations. This means that it is vulnerable to the same
criticisms as assertions of the health consequences effects of overweight and obesity
(see section 7.5). In yet another plea for further research, the authors also argued for a
systematic review or meta-analysis ‘to provide a clearer understanding of the
relationships found in such studies between excess weight and the diseases’ (WHO,
2000: 82).

Cost of illness techniques were seen to have two major limitations. The first was that
‘intangible costs and many of the direct costs of disease management and prevention,
especially those incurred outside the formal health care system, tend to be ignored’
(WHO, 2000: 78). Although studies of the broader economic costs of overweight and
obesity were cited (WHO, 2000: 86- 87), there were only four of them,119 mostly from
Scandinavian countries, and it was not clear whether their results could be
generalised. The second was that such studies can be ‘misused’ since diseases with
high costs should not necessarily be given high priority, because of ineffective
existing treatments or insufficient impact on overall mortality and morbidity: dental
disease incurs high costs but low morbidity and mortality compared to youth suicide
which incurs low health care cost but a high burden in terms of mortality and

119
This included the Swedish Obese Subjects study.

288
reduction in quality of life (WHO, 2000: 81). Such judgements about the burden of
disease and quality of life could not be reduced to the financial costs and benefits of
particular treatments.

The report’s authors admitted that there had not been any comparable studies on the
economic impact of obesity and overweight in developing countries, so they cited
studies by the WHO and the World Bank analysing the ‘increasing burden associated
with the rapidly emerging adult NCDs.’ (WHO, 2000: 88)

In developing countries, about 50% of deaths in 1990 were caused by NCDs,

but by 2020 that proportion is expected to rise to almost 77%...In contrast in
developed countries 87% of deaths in 1990 were from NCDs, and the
proportion is expected to rise only slightly – to 90% - by 2020 (ibid.).

These figures come from the Global Burden of Disease study and the World Health
Report of 1997. But the fact that these were the only figures the authors could quote
demonstrates another facet of the omission of developing countries (and, therefore, a
large percentage of the world’s population) from this discourse.

The authors of TR894 also gave preliminary analyses of the costs and benefits of
different obesity treatments. Very little research had been done on this topic apart
from the Swedish Obese Subjects study which compared a group of obese individuals
who had undergone weight loss surgery with a control group who did not. The results
showed that the prevalence of diabetes decreased by 68% in the group that had
surgery, and there was a four fold reduction in the risk of developing hypertension
and raised blood cholesterol (WHO, 2000: 89). Applying the reduction of diabetes
prevalence ‘would decrease the total cost of obesity in [France] by approximately 3%,
while in the USA costs could be reduced by almost 20%’ (WHO, 2000: 90). These
figures were highly provisional as they consisted of aggregates of estimates and
assumptions piled on top of each other. More prosaically, they did not include the
costs of surgical treatment, or subsequent medical care, which given the risks of such
surgery might have been more substantial that the authors allowed.

The economic evaluation of obesity prevention and treatment was also an under-
studied area and the authors could only summarise the results of two studies of

289
interventions for the prevention and treatment of diabetes in Australia and of the cost-
effectiveness of different weight loss treatments in the Boston area (WHO, 2000: 91 -
95), but neither of them provided useful information for assessing the financial
benefits of obesity prevention programmes. Again, there had been no studies done in
developing countries:

No analyses have been made of the economic costs of obesity treatment in

developing countries. However, other analyses of the costs of health
interventions show that prevention is more cost-effective than treatment once a
disease is diagnosed (WHO, 2000: 95).

The cost of the clinical services required to treat chronic diseases was seen to be
entirely beyond the resources of low income developing countries, and a major cost
for the healthcare systems of middle income countries, therefore ‘it would appear to
be more cost-effective for money spent on obesity and other NCDs to be used for
prevention rather than for expensive treatments during the advanced stage of disease.’
(WHO, 2000: 96). A final point echoes earlier writing on the ‘epidemic’ of coronary
heart disease (see section 2.2).

In developing countries where NCD epidemics are emerging or accelerating, a

large proportion of NCD deaths occur in the productive middle years of life, at
ages much younger than those seen in developed countries. The health burdens
attributable to excess weight gain in societies in transition are likely to be huge
because of the absolute numbers at risk, the large reduction in life expectancy
and the fact that that the problem affects, in particular, individuals with a key
role in promoting economic development (WHO, 2000: 97).

Despite a lack of evidence this rhetoric argument still managed to construct excess
body weight as a serious economic problem for developing countries.

8.7 The Global Burden of Disease Project

A contemporary research programme whose results were incorporated into obesity

science by the authors of WHO reports was the Global Burden of Disease (GBD)
project, begun in 1990 by the Harvard School of Public Health, WHO and the World
Bank (Murray and Lopez 1996). It would have been politically expedient to use GBD
results in a WHO technical report, since they came out of the organisation’s

290
collaboration with two very prestigious organisations. Two important concepts
deriving from GBD research were used by the authors of TR894. The first of these
was the ‘double burden of disease’ that was forecast to occur in developing countries.
This came from figures showing rates of non-communicable disease increasing in
poorer countries:

The next two decades will see dramatic changes in the health needs of the
world’s populations. In the developing regions where four-fifths of the
planet’s people live, noncommunicable diseases such as depression and heart
disease are fast replacing the traditional enemies such as infectious diseases
and malnutrition as the leading causes of disability and death. By the year
2020, noncommunicable diseases are expected to account for seven out of
every ten deaths in the developing regions, compared with less than half today
(Murray and Lopez 1996: 1)

According to these authors, increasing rates of NCDs now co-existed with continuing
high rates of infectious disease to create levels of disease and mortality beyond the
resources of poor countries’ health care systems (see section 8.6). This idea of the
double burden of disease was important to the authors of TR894 because it could be
used in accounts of obesity as a global public health problem, not just a condition of
rich industrialised countries. BMI distribution in adult populations was thought to
vary according to the country’s stage of development, and rising incomes led to a
population shift where overweight replaced thinness. In the early stages of
development, there would be an increase in the numbers of affluent people with high
BMIs, whereas poor people would mostly still be undernourished: ‘in countries in the
early stages of [this] transition, overweight can co-exist with underweight, so that the
burden of disease may be doubled’ (WHO, 2000: 31). It was another important means
of breaking the association between non-communicable diseases and affluence, and
was also used in discussions of the causes of obesity (see section 8.3).

The second of these concepts was the DALY (disability-adjusted life years) measure.
The DALY was devised as a measure to estimate the health burden resulting from a
particular condition, by calculating the years lost due to increased mortality and the
decrease in quality of life due to serious illness. Intended to be an ‘internationally
standardized’ version of the QUALY, it was developed as part of the GBD project in
order to ‘capture the impact of both premature death and disability in a single

291
measure’ (Murray and Lopez 1996: 6). Such a measure was useful for the authors of
TR894, as it could be used to estimate more precisely the costs of the conditions
associated with obesity that were not fully captured in mortality figures. As chronic
diseases develop over many years and often have disabling symptoms, including the
effects of morbidity as well as mortality allowed for a more complete estimate of
costs to healthcare systems and society as a whole.

8.8 The environmental paradigm of obesity in TR894

The shifts in understandings of obesity and overweight that I have outlined in the
previous four chapters are best considered as combining to form a new paradigm
(Kuhn, 1970). However, this new paradigm did not replace older understandings of
obesity, instead the approaches co-existed. The authors of TR894 still located their
understanding of the causes of overweight and obesity within the framework of
energy balance:

In simple terms, obesity is a consequence of an energy imbalance – energy

intake exceeds energy expenditure over a considerable period. Many complex
and diverse factors can give rise to a positive energy balance, but it is the
interaction between a number of these factors, rather than the influence of any
single factor that is thought to be responsible. In contrast to the widely held
perception among the public and parts of the scientific and medical
community, it is clear that obesity is not just a result of overindulgence in
highly palatable foods, or of a lack of physical activity (WHO, 2000: 101).

Multiple and incompletely understood physiological processes were seen to control

the regulation of body weight, but these processes could be ‘overwhelmed’ by a
positive energy balance arising from a combination of increased energy intake and
decreased energy expenditure. In particular, the body was seen to have stronger
defences against under-nutrition and weight loss than against over-nutrition and
weight gain, and a small but consistently positive chronic energy balance was again
seen as capable of producing large increases in body weight (WHO, 2000: 105). This
was all rooted in the earlier understandings I have labelled the individual paradigm of
obesity.

292
Between 1970 and 1995, understandings of obesity expanded and shifted focus, and
by the publication of TR894 in 2000, this new approach could be considered a
separate paradigm, that I have labelled the environmental paradigm. Although these
two approaches co-existed, partly because they operate on different levels, I believe
they are best described as different paradigms because the new approach used new
sources of data, developed new classification schemes, provided new causal
explanations and suggested new solutions. The new sources of data were results from
epidemiological studies such as Framingham and the American Cancer Study that
were incorporated into obesity science to produce a list of negative health
consequences of overweight. New techniques in health economics had provided an
increasing amount of statistical data on the costs of chronic diseases, and so the costs
to healthcare systems had become an important element of the argument that obesity
was a major public health problem. Bray and Garrow developed the new classification
system of BMI cut-off points that were now used as the standard definition of obesity
and overweight, despite ongoing debate about their relevance to different populations.
New causal explanations focused on the environmental causes of obesity and
overweight, stressing the increased availability of energy dense foods and decreased
energy demands of daily living resulting from processes of modernisation and
industrialisation. Individual treatment options remained the same, but the long-term
failure of these measures, especially weight loss diets, was increasingly
acknowledged. Combined with rising prevalences, this led to increasing formulation
of preventive measures that were quite conservative despite the rhetoric of radical
structural change.

In this way, researchers framing public health policy recommendations for the British
government and the WHO transformed excess body weight from an individual
condition caused by over-eating and treated by weight-loss diets into a global public
health problem caused by changes in the food supply, patterns of work and physical
environments, that required significant structural changes in order to reduce the
potentially huge costs to healthcare systems.

8.9 Conclusion

293
By the beginning of the twenty first century, the causes of obesity and overweight
were being framed in obesity science as environmental ones, resulting from processes
of industrialisation and modernisation. The availability of energy-dense processed
food, mechanisation in the workplace, sedentary leisure activities and motorised
transport were the factors usually highlighted. Unlike the authors of TR854, the
authors of TR894 did not directly refer to the ‘thrifty gene’ hypothesis. However, the
importance they give to the distinction between modern and traditional lifestyles,
combined with their focus on particular populations (see section 7.4) is consonant
with such accounts (see chapter 9).

The authors of TR854 and TR894 also did not use the term ‘obesogenic environment’
(Swinburn et al., 1999) like the authors of the contemporary BNF report (see section
5.4), but such a concept is consonant with their analysis, and would subsequently be
widely used in obesity science (WHO Regional Office for Europe, 2006, Grant et al.,
2007, Foresight, 2007) and popular writing on the topic (Tsichlia and Johnstone,
2010). The more radical preventive measures discussed by the authors of TR894 –
regulating advertising to children, nutritional standards for institutions, improved food
labelling, encouraging physical activities and providing facilities for pedestrians and
cyclists - were subsequently also widely discussed in the popular press. In the UK,
they have also become part of a mainstream discussion of policy responses to the
public health problem of obesity and overweight (see for example Foresight, 2007).

The increasing amount of data on the direct and indirect costs of conditions such as
heart disease and diabetes (and to a more limited extent obesity) illustrates the growth
in healthcare bureaucracies and the increasing range of statistical information that was
collected and calculated by governments and biomedical researchers. Other authors
have given historical accounts of these developments in the 18th and 19th century
(Desrosieres, 1998, Hacking, 2006a) but the continuing expansion of the discipline of
health economics demonstrates the further development of such techniques in the
second half of the twentieth century.

294
CHAPTER 9: INDIVIDUAL AND SOCIO-POLITICAL EXPLANATIONS OF
EXCESS BODY WEIGHT

9.1 Introduction

In the 1990s, the obesity science coalition expanded its sphere of influence markedly.
Obesity and overweight became (temporarily) incorporated into central UK
government health target setting in the early 1990s (DoH, 1992: 49). Concern about
obesity as a public health problem had moved up the governmental hierarchy to be
considered by the Department of Health as a whole (DoH, 1995) (see section 3.5),
rather than by one of its expert committees, as in 1977. The coalition’s activities then
extended further onto the global stage as members contributed to two WHO technical
reports in the late 1990s. The first, TR854 on anthropometry, included a chapter on
overweight (WHO, 1995), but the second, TR894, was entirely on obesity, labelling it
as a global epidemic (WHO, 2000). This was produced under the chairmanship of
Philip James. As I have argued previously, TR894 was widely referred to in both the
popular and the medical press, and an important resource for legitimating ideas of
obesity as a global public health problem (Fletcher, 2007). The authority of the WHO
made this report an easily cited piece of evidence for writers in the popular press. But
this phenomenon was also found in the scientific and medical press. Referring to the
existence of an ‘obesity epidemic’ and citing TR894 became an easy way of
demonstrating the importance and relevance of many different kinds of research. It
became a bandwagon in the mainstream sense, as well as in the narrower STS sense
(Fujimura, 1988).

Despite its high profile, this report remains something of an oddity. Since its
foundation the WHO had prioritised treating epidemic disease and building health
capacity in poor countries (see section 9.4). Initially, this led to a focus on malaria,
tuberculosis, and venereal diseases and on promoting prevention through work on
maternal and child health, environmental sanitation, and nutrition (Staples, 2006:
138). In the 1990s, the organisation’s declared priorities were fighting malaria,
HIV/AIDS and tuberculosis and the regulation of tobacco marketing (Lee, 2009: 84-5,
94). The WHO had begun work on the global control of cardiovascular diseases in the
1970s (Litsios, 2008: 221) and since the 1960s, occasional technical reports had been

295
produced on coronary heart disease, and then also on chronic diseases (WHO, 1982,
WHO, 1986, WHO, 1990). The WHO’s European regional office had worked in the
area of cardiovascular disease since the 1960s and developed a long-term programme
of research into CHD and its risk factors120 (Kaprio, 1991: 35-6). But these precursor
activities were irregular and limited to one of the WHO’s six regions: they do not
fully explain why TR894 was produced at this time. From the mid 1990s the WHO
had also been heavily criticised within the medical community as over-bureaucratic
and inefficient, and its Director-General at that time, Hiroshi Nakajima, was
unpopular and widely seen as ineffective (Anon, 1997, Godlee, 1994, Godlee, 1997).
Moreover, as some rich donor countries had withheld or frozen their budget
contributions, the organisation had been attempting to economise and improve its
efficiency (MacGregor, 1994, MacGregor, 1996, MacGregor, 1999). A report into the
global epidemic of obesity seems to be oddly disconnected from the rest of the
WHO’s work at the time and its declared priorities.

One of the factors that might have made the production of this report possible was its
relevance to debates about future increases in health costs that would result from
rising rates of chronic disease (see section 8.5). Questions of cost form a significant
portion of TR894 (see section 8.4), and they are prioritised by appearing in one of the
earlier chapters. Rising concern about rates of chronic disease combined with
increasing forecasting of future demands on resources by healthcare systems
increased debates about the rationing of healthcare from the 1970s onwards. Such
debates had long been central to the discourse of preventive health (Lupton, 1995,
Petersen and Lupton, 1996). In the UK in the 1980s and early 1990s, these took place
in the context of widespread reductions in government health spending due to a
combination of neo-liberal policies and economic recession (Berridge, 1999a: 97,
Baggott, 2000:50-51). However, the British researchers who were most active in the
public health coalition around obesity were mostly consultants in large teaching
hospitals and researchers in universities or government funded research institutions.
They were, thus, insulated from the worst of these funding reductions (Berridge,
1999a, Webster, 2002), and had the resources to develop the institutional networks
required for such a coalition to function successfully.

120
The North Karelia Project (see section 2.5) was part of this research programme.

296
9.2 Public health expertise and boundary-work in obesity science

As I have described above, the literature on obesity, and especially the WHO
literature on the global epidemic, articulates shifts towards an environmental
paradigm from the 1980s. The causes of obesity and overweight were increasingly
framed as the undesirable effects of contemporary patterns of eating, working and
travelling (see section 8.2). These effects were first thought to be occurring in
industrialised countries, but later in poorer countries as well (see section 8.6). There
were two aspects of these changes – changes in dietary composition and changes in
physical activity levels. Rising average incomes had led to increasing consumption of
protein, fat and sugar (see section 8.2). As part of a shift away from under-nutrition
these changes do not seem inherently negative to health, but they were thought to be
accompanied by a shift away from diets based on high levels of unrefined
carbohydrates, and towards more energy dense diets containing much higher levels of
processed food (see sections 5.3, 5.4 and 8.2). Declining levels of physical activity
resulted from economic development: increasing mechanisation, industrialisation and
urbanisation meant that employment became less physically demanding, sedentary
leisure activities grew in popularity, and individuals relied more on motorised
transport (see sections 5.4 and 8.2). Arguments about changing patterns of food
consumption were based on FAO national food balance sheets – statistics of aggregate
production and consumption (Drewnoski and Popkin, 1997). However claims about
changing levels of physical activity levels were less well documented, and it has been
argued that the existing evidence does not clearly demonstrate such a decrease (Gard
and Wright, 2005: 119-24).

Obesity science researchers labelled visible signs of contemporary urban life - fast
food, cars and television and computer games - as important causes of ill-health, often
in the absence of concrete evidence. This was a new version of the ‘diseases of
civilisation’ argument (Rosenberg, 1998) which appears to have expanded, filling the
gaps in existing data (see section 5.4 and 8.2). Moreover, in TR894, there is a sense in
which of ‘traditional lifestyles’, and even poverty, were being framed as healthier than
affluence. By necessitating physically demanding work, a less processed diet and
transport by foot or bicycle, a lack of resources meant that individuals avoided the
chronic diseases associated with ‘affluent’ lifestyles (WHO, 2000: 120, 123-4, 126).

297
This is a crude over-simplification of these authors’ arguments, but, by not explicitly
acknowledging the health benefits of ‘affluent’ lifestyles, such accounts represented
the health costs and benefits of changes deriving from industrialisation and
modernisation in a very partial fashion.

These changes in the patterns of daily life were pervasive and highly interconnected.
To reverse the effects in line with the analyses presented in these reports would have
required changes on an enormous scale - for example, the restructuring of agriculture
and the food processing and manufacturing industries. Moreover, the commercial
interests involved - such as the motor and oil industries - form a significant portion of
the global economy. These factors made it unlikely that such a programme would be
carried out by contemporary governments, even those of rich industrialised countries.
In particular, economic growth is often seen as a greater political priority than
population health, even in affluent countries (Baggott, 2000: 56). Such (unexpressed)
considerations led to an important gap in the rhetoric of obesity science which was
characterised by an expansive description of the possible socio-economic causes of
excess body weight, in contrast to a relatively conservative range of potential
solutions.

Only in the area of prevention were some of the more wide ranging causes addressed
and then their political implications were blunted by being largely framed in the
health education language of individual choice. Although increased regulation of the
food industry or provision of financial incentives for weight loss were sometimes
suggested, most proposals involve advising, encouraging or educating consumers,
medical professionals and the general public (see section 8.3). This gap is the product
of a crucial form of boundary-work (see section 1.6.2) undertaken by members of the
obesity science coalition who understood that it was beyond their authority to call
directly for such wide ranging social change. Individual behaviour, especially in the
area of diet and physical exercise, has long been seen as seen as coming under the
authority of doctors (Shapin, 2003). A focus on solutions based on individual
education and voluntary codes of practice was, therefore, a form of political self-
limitation to ensure that these researchers would not damage their credibility by

298
overstepping the boundaries of their acknowledged expertise.121 The professional
interests of epidemiologists and others working in this area may have been to improve
population health by expanding their sphere of influence, but even in the contest of
modern preventive healthcare, there were political limits involved in the framing of
policy.

9.3 The effects of global health inequalities: poverty as the missing

factor

There are problems within the causal explanations of obesity and overweight provided
by the environmental paradigm, and approaches from health inequalities research may
be able to provide a fuller explanation of the ill-health of the exemplary populations
of obesity science. The discourse of modernisation (see section 8.2) and its negative
effects on the health of populations has lacunae, and there appears to be an unhelpful
idealisation of the traditional taking place in this writing. Some critics of the dominant
explanation of obesity have argued that poverty and social disadvantage offer an
alternative and perhaps more satisfactory account of the obesity suffered by
exemplary cases of obese populations including Pima Indians and Australian
aborigines. The authors I discuss below do not extend this explanation to obesity in
rich countries, but other researchers have as part of their wider critique of obesity
science (Gard and Wright, 2005: 124-5).

Genetic differences were often discussed as possible explanation of the high rates of
diabetes and obesity in such populations (see section 7.4). By describing the high
rates of overweight and obesity in the Pima Indians or Australian aborigines as due to
a genetic mismatch between their hunter-gatherer past and their modern present, such
theories wipe out a history of displacement and poverty that have contributed equally
to their community’s poor health: ‘diabetes has been figured as the process aboriginal
people have paid for civilization rather than the penalty exacted by colonization’ (Fee,
2006: 2994). In a discussion of high rates of chronic disease amongst Australian
aboriginals, McDermott argued that

121
Such boundary-work also partially explains the neglect of ‘social’ explanations in obesity science
(see section 6.6).

299
 If the…“epidemiological” narrative of diabetes in Aborigines is expanded to
include social history, early life experiences of malnutrition, poverty and
illness followed by later experiences of welfarism, poverty, physical inactivity
and obesity, explanations other than genetic ones might emerge (McDermott,
1998: 1193).

Modern genetic explanations are a narrowing of James McNeel’s original concept of

the ‘thrifty genotype’ which saw the ill-health resulting from a western lifestyle as a
universal phenomenon (McDermott, 1998, Fee, 2006). This narrowing has led to a
focus on research into the genetic determinants of chronic disease and a neglect of
research into understanding the social determinants of ill-health (McDermott, 1998:
1194). By incorporating the work of Barker (2007) into the effects of maternal
deprivation on the adult health of children, McDermott showed that an account of the
health of these groups could be constructed that accorded with contemporary
epidemiological consensus and acknowledged their history of colonisation,
displacement and poverty (ibid.). At the risk of over-simplifying complex histories,
many of the groups listed above, and routinely cited in the global health literature as
examples of under-developed populations who nonetheless suffer very high levels of
obesity, can be described as having experienced colonisation and, therefore, their
collective histories will show some of the same features as those of Australian
aborigines.

Instead of naturalising health inequalities through genetic explanations, other accounts

describe the social inequalities that produce them. Such inequalities are a feature of
many populations with high rates of obesity and overweight. Exemplary populations
in obesity science included Pacific Islanders, the Pima Indians, Black women in the
US and urban South Africa, and Australian aboriginals (see section 7.4). One feature
that many of these exemplary populations shared was relative poverty. They shared
experiences of displacement, high levels of economic dependency and financial
insecurity. For example, the Pima have been displaced from their land and live on
reservations in South Arizona, unable to farm the desert land they irrigated: decades
of medical research have been conducted on them that appear to have brought very
little benefit to their community (Gaesser, 2002: 172). Samoans and other Pacific
Islanders are geographically and economically marginal: their economies are often
dependent on a public sector funded by development aid and much of their food is

300
imported (McMurray and Smith, 2001, discussing the case of the Marshall Islands).
For Samoans and Marshallese, this has resulted in diets based on highly processed
nutritionally poor foods items such as tea, white rice, instant noodles and sugared
drinks, and low grade fatty meats such as turkey tails and lamb flaps (Cheung, 2010:
2207). Black women are one of the poorest socioeconomic groups in America, and
this is reflected in their health status.

The links between poverty and poor diet have been part of British health inequalities
research since the Black Report (Townsend and Davidson, 1982, see also section
1.2.1). In the 1990s this topic began to be written about more frequently (Leather,
1996, see also Dowler et al., 2007). Nutritionists contributed to this discussion,
including Philip James who co-authored an article for the BMJ on the topic, with Suzi
Leather from the Food Poverty Network. They argued that

The diet of the lower socioeconomic groups provides cheap energy from foods
such as meat products, full cream milk, fats, sugars, preserves, potatoes, and
cereals but has little intake of vegetables, fruit, and wholewheat bread. This
type of diet is lower in essential nutrients such as calcium, iron, magnesium,
folate, and vitamin C than that of the higher socioeconomic groups (James et
al., 1997: 1545).

Such a diet of high fat and high sugar processed foods was precisely the kind of diet
that was supposed to lead to increasing rates of obesity and overweight (see section
8.2). Given that these arguments were readily available to obesity science researchers,
and even produced by one of the coalition’s major participants, it is necessary to
explain why they were not drawn upon to account for the causes of excess
bodyweight. One important reason is that they were politically contentious and
another that they did not fit with the ideas of individual responsibility embedded in
chronic disease epidemiology (see section 2.5)

In an alternative framing, these differences in average body weights amongst

populations were the result of global health inequalities. The highest rates were found
in groups that were living in relative poverty in rich industrialised countries or in
poorer regions that were dependent on rich countries. Members of such groups ate
diets high in processed food because they were cheaper, readily available or high
status because of being ‘western’. They may have led more sedentary lifestyles due to

301
mechanisation and motorised transport, but individuals often worked long hours in
bad conditions and did not have the time, space or cultural permission for leisure
activities. These were still relatively poor populations who did not have the resources
or the political clout to accrue the health benefits that other, more affluent groups or
countries could derive from economic development.

9.4 The social and the biomedical: competing models of health and
disease within the WHO

The contested boundary between the ‘medical’ and the ‘political’ that demanded
political self-limitation on behalf on obesity science authors has been a pervasive
feature of modern public health policymaking. In the context of the WHO it is
illustrated by a tension between biomedical and social models of health that has
existed from the founding of the organisation. Biomedical models explain the causes
of ill-health in more individualised terms and seek technical solutions (Blaxter, 2010:
12-16). In contrast social models focus on the link between ill-health and social
structures, and seek socio-political change to improve health (Blaxter, 2010: 16-19).
These different models contain a different understanding of the boundary between the
‘social’ and the ‘medical’ and therefore differing assessments of the legitimate
authority of the medical profession.

Like other UN bodies, including the World Bank and the FAO, the WHO was
founded on the basis of a post-war optimism about possibilities of improving the lives
of individuals and populations (Staples, 2006). The work of these organisations
derived from a founding narrative of progress and economic development, a reaction
to the devastation of the Second World War. In the case of the WHO, this optimism is
illustrated by its famous definition of health as ‘a state of complete, physical, mental
and social well-being and not merely the absence of disease and infirmity’ (Siddiqi,
1995: 226). Such an expansive definition derived from the ideals of a social model of
health. The public health experts responsible for setting up the WHO saw it as a
universal membership organisation that would act collaboratively to address a wide
range of health needs. This was a vision of social medicine guided by ideas of
humanitarianism and social equity (Lee, 2009: 16-7). However, Lee argues that, from

302
its founding, there was another narrower and more biomedical conception of the
organisation’s role (Lee, 2009: 46-7).

This competing vision, framed by the political context of the Cold War and
functionalist political theories, was a more circumscribed understanding of its role in
promoting international health co-operation. It involved the WHO focusing on the
prevention of disease, especially epidemic infectious diseases, and the provision of
technical advice to poorer countries. Two factors helped to promote this approach.
The first of these was recent developments in medical knowledge and practice that
seemed to hold the promise of significant improvements in health through more
effective treatments, and, sometimes, the complete elimination of diseases (Lee, 2009:
46, Siddiqi, 1995: 195). The second was the collective professional identity of WHO
representatives and staff, as members of a well-connected international medical
community:

When these internationally minded doctors came together within the World
Health Organization, they combined the scientifically based authority of their
profession with a commitment to apolitical internationalism in order to garner
a global authority that facilitated the often unquestioning acceptance of their
recommendations and ensured that countries clamoured for the organization’s
advice and assistance (Staples, 2006: 137).

Walt also describes how this deployment of professional expertise meant that in the
1950s and 1960s ‘WHO was stable and pragmatic, largely disease oriented and
dominated by medical professionals’ (Walt, 1994: 137).

Despite the wider ideals embedded in its definition of health, the biomedical model
was dominant in the early history of the WHO (see section 1.2.1). It enabled the
organisation to operate across national boundaries and in varying political contexts: a
narrower technical approach helped to avoid accusations of compromising the
authority of national governments. An early focus on technical activities, such as the
revision of international sanitary regulations, allowed the organisation to maintain a
low profile in its early years, but ‘support for tackling the social factors influencing
health and disease was never far from the surface’ (Lee, 2009: 18). However, for the
next 20 years, the most important and prestigious of the WHO’s activities was the
Malaria Eradication Programme (Siddiqi, 1995). This was the epitome of a ‘vertical’

303
programme that used the newly developed technology of DDT insecticidal sprays and
a globally standard approach to eliminate one infectious disease endemic in poor
countries.

By the 1980s the Malaria Eradication Programme was considered to have failed (Lee,
2009: 49) and criticisms of the social power of the medical profession and the
biomedical model of disease had become more mainstream (Berridge, 1999a: 56-7).
Such a climate enabled the WHO to broadened its activities and focus on the social as
well as the individual determinants of health (Lee, 2009: 71-2). It switched from a
vertical model of tackling individual diseases to a ‘horizontal’ model of primary
healthcare that aimed to meet the healthcare needs of poor populations (Siddiqi, 1995:
196). Major initiatives of this period included Health For All, which came directly out
of the primary healthcare movement, and attempts to strengthen global regulation of
the marketing of baby milk, pharmaceuticals and tobacco (Lee, 2009: 71-98). This
change resulted from a shift towards the priorities of developing countries (Siddiqi,
1995: 209), and the increasing influence of campaigning NGOs and other civil society
organisations (Walt, 1994: 140, Lee, 2009: 87-95). It was so comprehensive that,
writing in 1993, Siddiqi could argue that primary health care had become the
‘dominant philosophy’ within the WHO (Siddiqi, 1995: 196). Tackling the structural
causes of ill-health involved political controversy as the WHO attempted to curb the
activities, and profits, of the tobacco, pharmaceutical and food industries.

Taking on such powerful commercial interests led to criticism from the US and a
long-term budgetary freeze by major donor countries, which has been described as
part of a ‘sustained effort to keep the WHO focused on biomedicine’ (Lee, 2009: 98).
This was combined with sustained accusations of politicisation (Siddiqi, 1995) and
the relative unpopularity of later Director-Generals, compared to the charismatic
Brock Chisholm and Halfdan Mahler (Lee, 2009: 15-16, 74-5). By the 1990s, the
biomedical model was once again dominant, partly due to renewed focus on
individual diseases such as TB and HIV/AIDS by rich donor countries (Lee, 2009:
84). These countries were increasingly funding disease-specific special projects,
which were taking up an expanding proportion of the organisation’s budget, giving
them greater influence over the WHO’s priority areas (Lee, 2009: 39, 85). In this way,

304
the rich countries managed to re-orient the organisation’s agenda, and re-instate a
narrower model of its activities (Lee, 2009: 98).

The WHO founding model of health, as including social well-being, was closely
related to social medicine which was influential in Britain in the 1940s and 1950s
(Porter, 1997a). Porter describes how social medicine in Britain in the 1940s aimed to
become ‘a medicine of society for society’, and that ‘its mission was to facilitate
progressive human social and biological evolution’ (Porter, 1997a: 99). Social
medicine lost much of its influence in the UK after the 1960s (see section 1.2.1), and
research from the social sciences began to be applied to medical topics as part of the
developing sub-discipline of medical sociology, rather than in the main body of
medicine (Porter, 1997b). The social model seems to have remained part of the
mainstream within the WHO for longer (see above), and this may be because much of
the organisation’s work is in the area of public health. Because it studies population
health, public health contains more of the ‘social’ than other medical specialities, both
in terms of its understanding of the role of structural factors in health, and in its
history of political engagement. However, the boundary between the ‘biomedical’ and
the ‘social’ shifted regularly in post-war chronic disease epidemiology, as new forms
of behaviour became to be seen as the concern of public health (see section 3.2 for
one example). This continuing tension between wider socially engaged models of
health and narrower more technical and biomedical ones was demonstrated in the
1990s by the gap between expansive understandings of the causes of excess body
weight and much narrower recommendations for treatment and prevention.

9.5 Making an American and European health problem global?

Within this changing political orientation, concern with chronic diseases was inserted
into the WHO agenda in the 1980s by developed countries, and projected onto
developing countries as these diseases cease to be described as ‘diseases of affluence’.
By the technical report of 2000, obesity had become one of the ‘chronic diseases’ that
was projected onto poor countries (in the absence of much evidence) at that time.

The WHO is a global health body which has historically carried out much of its work
in poor, less developed countries. Because of lack of ‘backstage’ information (see

305
section 1.8.4), my arguments about WHO internal processes contain an element of
speculation. However, in this context, a report on a global health problem presumably
would have carried more weight inside the organisation than one that focused on a
condition found mostly in rich countries and regions, especially if the condition was
normally understood to be the result of over-consumption and individual lifestyle.
Such considerations lead to the framing of obesity and overweight as a worldwide
problem: the ‘everyone, everywhere’ approach (Gard and Wright, 2005: 17). This
framing also fitted well with the development of new categories of ‘at risk’
individuals and populations that takes place within risk factor approaches to chronic
disease (Petersen and Lupton, 1996: 18-22). However, to construct an argument for
obesity and overweight as a global health problem required significant rhetorical
work. Some of this had already been accomplished in other areas. The labelling of
chronic diseases such as heart disease, stroke, diabetes and cancer shifted between the
1960s and the 1980s. In the 1960s and 1970s Morris described them as ‘modern
epidemics’ (see section 2.3). In the 1980s they became a focus of attention as
‘diseases of affluence’ (Hardy and Tansey, 2006: 427-8). By the 1990s, WHO reports
contained arguments for the name to change from diseases of affluence to ‘chronic
diseases’122 (WHO, 1990: 10). This shift made it possible to argue that poor
populations and countries could experience high rates of such conditions. Such
framings were most fully articulated in ideas of the ‘double burden of disease’
developed by researchers from the Global Burden of Disease project (see section 8.6).
As I outlined above, this work was heavily cited within TR894, as it provided an
authoritative source that argued chronic disease was a global health problem.

According to the prevalence data cited by the authors of TR894, increasing rates of
excess body weight were found in European and North and South America
populations, and some other smaller groups such as Pacific Islanders, middle class
Indians and urban Black South African women (see section 7.3). However, it did not
seem to be an issue for the majority of the population of Africa and Asia, who account
for approximately two thirds of the world’s population. This may have been partly
due to a lack of information: as the authors pointed out several times, good quality

122
In this report such diseases were described as being caused by an ‘affluent diet’ (WHO, 1990: 10-
14), so the idea of affluence was still being used, but gradually this usage fades as the components of
such a diet become cheaper and its highly processed or industrial nature became seen as more
important.

306
data was available only for Europe and North America (WHO, 2000: 16-30). The
description of increasing rates of excess body weight as an urgent public health
problem followed closely by calls for more research has been a consistent feature of
obesity science. The certainty that obesity was an important public health problem
had been juxtaposed with recurrent descriptions of lack of data since the 1960s.123

Modelling and forecasting were important in the arguments used by the authors of
TR894 to frame obesity and overweight as a global health problem. In this report such
techniques were used in two important ways. Firstly, in the absence of representative
data, populations were modelled on the basis of what data did exist. In some cases this
included generalising from obviously unrepresentative populations: Thai government
officials represented their entire country; data from only six countries (including sub-
populations such as urban black South Africans and Mauritians) represented the entire
population of continental Africa (see section 7.3). These were obvious attempts to
compensate for missing data. A more subtle form of modelling occurred when this
data was presented as if increases in the past represent the only possible future pattern
– it was taken for granted by these authors that average body weights not only have
been rising, but would keep on rising, in all regions and populations. The possibility
that increases in average body weight might begin to level off was not even
considered.124 Combined with estimates of the costs of obesity and overweight
produced by health economics (see section 8.5), this extrapolation produced a vision
of inexorably increasing average body weights and spiralling health care costs that
was at odds with other contemporary health data for industrialised countries, showing
rising life expectancies and decreasing mortality rates (Wilkinson, 1996: 29-30).

It can be argued that the data cited in TR 894 did not demonstrate that obesity and
overweight was a global problem. Data did not exist for a large proportion of the
world’s population, and the limited data that did exist for much of African, China and
other parts of South East Asia showed very low rates of excess body weight (see

123
The acceptability of such framings may be linked to traditions within public health of raising
concerns in situations of incomplete information and, therefore, high uncertainty.
124
There is some evidence that average body weights among Canadian and American children are
beginning to plateau (Gard, 2008).

307
section 7.3). Instead the data included in these reports demonstrated that obesity was a
problem for European and American populations, and, especially for particular sub-
populations. The sub-populations that were worst affected were those in industrialised
economies who were marginal for reasons such as poverty, ethnicity or geography.
Pacific islanders, Pima Indians, some aboriginal Australians, Black populations in the
US and south-east Asian populations in the UK showed very high rates of excess
body weight, often accompanied by high rates of chronic diseases such as diabetes
(see section 9.5). Outside of Europe, North America and these specific populations,
average body weights either remained low or they were increasing, but by relatively
small amounts given the restrictiveness of the BMI-based definitions.

9.6 Conclusion

Obesity science’s heavy reliance on ideas of modernisation in its explanations posed

problems for how to explain obesity in under-developed populations, such as
Australian aborigines, in a way that agreed with how it was explained in developed
countries. Attributing it to poverty, both in developed countries and globally, would
be politically sensitive, particularly in view of the corporate pressure on WHO to
draw back from a political understanding of public health problems. In this respect, it
was fortuitous that obesity had been placed on the WHO agenda not by public health
experts but by biomedical researchers and clinical practitioners, who were disinclined
to propose ‘social’ causes. Instead, they were able to adapt a model already
developing in their own work. The dominant model in developed countries was still
the individual model, though in view of the failure of medical interventions obesity
science authors had started to acknowledge some ‘environmental’ causes. This
account was expanded to include a wider social or environmental account of the
causes of obesity globally, which was broadly acceptable within the WHO’s remit,
but which did not make reference to poverty.

Framing obesity and overweight as a global epidemic was a useful rhetorical tool for
members of the obesity science coalition as it validated their work. The attribution of
that epidemic to modernisation, and to the kinds of behaviour associated with
modernity, was useful to governments and institutions like the WHO, as it had the
effect of de-politicising the health consequences of poverty. It acted as a pressure

308
valve, allowing discussions that had developed from earlier concerns about high rates
of heart disease to develop without disturbing the relationship between expert
scientists and their government, or the WHO and its member governments. The use of
modernisation as an explanation allowed for discussion of current population health
issues and future forecasted problems to continue without directly referring to
contentious political issues such as health inequalities, poverty or the rights of
marginalised social group and ex-colonies. Although structural issues were raised in
discourses about the causes and prevention of excess body weight in this period, they
had little influence on the recommendations of such reports. Even when calls for
reform were made, for example, in agricultural subsidies or the food processing
industry, were made they were framed within a discourse of individual responsibility
so that their political impact was largely blunted.

An alternative reading of material contained in TR894 would still see excess body
weight as resulting from the processes of industrialisation and modernisation, but it
would acknowledge that these effects vary between populations and groups within
populations. It would acknowledge that economic development has brought health
benefits, reflected in increased life expectancy, but that these benefits have been
unequally distributed, with affluent populations and groups within populations
continuing to benefit disproportionately from such development. Crucially, this more
nuanced reading would accept that although body weight may be partially under
individual control, it is also the result of structural factors such as changes in food
production and retailing, in work patterns and in urban transport. Such changes have
been encouraged by governments of many political persuasions since they are
produced by, and, in turn, produce economic growth. In addition, a productionist
model of agriculture (Lang and Heasman, 2004) using mechanisation and petroleum
derived fertilizers and pesticides to increase productivity had been combined with
agricultural subsidies in order to keep food prices low in Europe and North America.
This subsidised and industrialised production of certain crops, combined with
innovations in food manufacturing, has led to diets with high levels of fat sugar and
salt. Decreasing average levels of physical activity can be linked to mechanisation and
motorisation in workplaces and public transport. These are also processes that have
been subsidised in many different ways by governments, from grants to specific
industries to road building.

309
The environmental paradigm of obesity partially addresses such socio-political factors
and potentially offers an enormous number of potential solutions such as reforming
the entire food system, from agriculture through to processing and retailing, or
redesigning cities and towns to encourage (or compel) individuals to walk and cycle
rather than drive or take public transport. However, such solutions are definitely
‘social’ rather than ‘medical’, and political self-limitation meant that these authors
reverted back to ideas of individual responsibility when proposing solutions to the
problem of excess body weight.

310
CHAPTER 10 CONCLUSION

10.1 Introduction

In the preceding chapters I have described how medical concern about obesity
developed out of the post-war growth and re-orientation of British and American
public health towards chronic rather than infectious disease. I showed how research
into the causes of heart disease led to the development of the large scale prospective
epidemiological study and the risk factor approach to chronic diseases. I outlined the
development of a public health coalition around obesity and of a policy-orientated
body of expert knowledge on the causes and treatment of excess body weight that I
have labelled obesity science. I summarised the initial model of obesity - the
individual paradigm - that was developed by British and American researchers in the
1970s and early 1980s. Through an examination of the development of the body mass
index and of the role of the energy balance model in obesity science, I showed how
their model of obesity continued to develop in the 1980s and 1990s.

Next I shifted focus to describe the activities of an overlapping and international

network of WHO expert scientist in the area of chronic disease and diet. I showed
how members of the British obesity coalition became active in this arena, and how,
due to the incorporation of new large scale epidemiological data and the move to a
global forum, the model of obesity developed in this period. Lastly I analysed a WHO
technical report on the topic of obesity that shows the development of an entirely new
model of obesity – the environmental paradigm. I argue that there are problems with
this model because it is not clear that obesity is a global health problem, and because
processes of political self-limitation meant that the causes of increasing rate of obesity
were attributed to the effects of industrialisation and modernity, rather than more
politically contentious explanations such as poverty.

10.2 Policy activity on obesity 2000 to the present

The obesity band wagon not only continued rolling after 2000, but also gained greater
momentum. The global epidemic became a high profile issue that was widely written
about in the popular and scientific press. This in turn appeared to generate more

311
policy activity as governments and international organisations needed to be seen to be
addressing such a pressing problem. In Britain, the issue was part of high level
political discussion. The National Audit Office produced a report on the topic in 2001,
and in 2002 organised the Joining Forces to Tackle Obesity conference. The House of
Commons Health Committee held an inquiry into obesity in 2003/4 and produced a
substantial report with 69 recommendations that, despite its ‘optimistic’ framing,
stressed the role of social norms in the development of obesity and seemed sceptical
about current health promotion-based interventions (Health Committee, 2004).
Obesity was once again included in the government health targets. ‘Reducing obesity
and improving diet and nutrition’ was one of the ‘overarching priorities’ of the
Choosing Health initiative (DoH, 2004: 4), as was increasing exercise. The DoH also
promised ‘a new cross-government campaign to raise awareness of the health risks of
obesity and the steps people can take through diet and exercise to prevent obesity’
(DoH, 2004: 5) as well as a ‘definitive’ NICE guideline on the prevention,
identification, management and treatment of obesity (DoH, 2004: 15). Obesity was
also discussed in the 2007 King’s Fund report into NHS funding (Wanless et al.,
2007) and used as a case study in the Nuffield Council on Bioethics report on ethical
issues in public health (Nuffield Council, 2007).This policy attention culminated in
the launching of the government Office for Science’s Foresight project Tackling
Obesities: Future Choices. This report argues that ‘By 2050 Britain could be a mostly
obese society’ (Foresight, 2007: 6) and further develops the environmental paradigm
to provide an account of the key ‘drivers’ of obesity and an ‘obesity system map’ of
impressive complexity, illustrating the multiple and interacting causes that have led to
the development of ‘passive obesity’.

In the international arena, a joint FAO/WHO technical report on diet, nutrition and the
prevention of chronic disease was published in 2003 (WHO, 2003). Prominent
members of the international obesity coalition sat on the committee, including Philip
James125 and Boyd Swinburn,126 while temporary advisors included Anna Ferro-
Luzzi and Peka Puska. This report continued to stress the increasing burden of chronic
disease affecting developing countries, resulting from industrialisation, modernisation
and globalisation. Obesity was also one of the conditions it considered in detail along

125
He was listed as chairman of the International Obesity Taskforce (IOTF).
126
Swinburn and a colleague coined the term ‘obesogenic environment’ (see section 6.4).

312
with diabetes, cancer and cardiovascular disease. In 2004, the WHO also launched a
global strategy on diet, physical activity and health at the World Health Assembly
(WHO, 2004). The role of unhealthy diets as one of two major causes of increasing
rates of non-communicable disease has been translated into a policy concern with
obesity in the reports produced by this programme. It remains active: a global
ministerial conference on healthy lifestyles and non-communicable disease control
will take place in Moscow in April 2011. There were major initiatives taking place at
the European level as well. The European Union launched its own Platform on Diet,
Physical Activity and Health in 2005 in co-operation with the WHO and the IOTF
(EC, 2005), and a year later, in Istanbul the WHO Europe office held the European
Ministerial Conference on Counteracting Obesity to address ‘the growing challenge
posed by the epidemic of obesity to health, economies and development’ (WHO,
2006, see also Branca et al., 2007a). The policy momentum has died down since this
period of frantic activity, but the issue still resurfaces regularly in the popular press,
when relevant research results or public health data are released.

10.3 The creation of policy knowledge

In this section, I want to examine what my case study can say about the creation of
public health policy knowledge. In particular I want to address three interrelated and
overlapping questions. Firstly what does the development of obesity science tell us
about the relationship between policy and epidemiology? Secondly what does it say
about the role of public health coalitions? Thirdly, what does it add to sociological
accounts of the new public health?

The relationship between epidemiology and policy is complex partly due to the
unusual character of epidemiological science. It is a large and relatively
heterogeneous body of knowledge – as well as biomedicine it incorporates elements
of non-human biology, geography, health economics, psychology and sociology.
Understanding patterns of health in human populations also involves analysing
complex and ill-defined systems.127 These factors may explain why, since the

127
I think there are several parallels between epidemiology and environmental science in this respect.
This may be one reason why, in the last few years, comparisons have been made in the press and policy
literature between the obesity epidemic and climate change (Foresight, 2007).

313
beginning of the twentieth century, there have been recurring internal doubts about
epidemiology’s status as a science (Amsterdamska, 2005). Since the nineteenth
century, British epidemiology has involved both policy-oriented and policy-engaged
research, to use contemporary terminology. But individuals acting as expert scientists,
on behalf of governments or of international organisations such as the WHO, have to
be careful not to exceed their professional authority when making policy
recommendations. Epidemiology contains such a broad range of accounts of the
causes of health problems that multiple forms of professional expertise can
legitimately be claimed in areas such as the relationship between nutrition and chronic
disease. In addition, such accounts could readily be mobilised to support proposals for
solutions to health problems, ranging from providing behavioural advice to
individuals at the micro-level to recommendations for wholesale social change at the
macro-level.

I have outlined above the effects of a political self-limitation whereby the

recommendations of expert scientists in the area of diet and health are largely
restricted to initiatives aimed at improving individual’s food choice or level of
physical activity (see section 9.2). I have also described how the causes of excess
body weight in particular populations are seen to derive from the effects of modernity,
rather than inequality or poverty (see section 9.3). In these kinds of ways, expert
scientists in this area limit what they say according to what is politically acceptable or
expedient. This self-limiting has important implications not just for overt policy
recommendations, but also for the content of obesity science more generally –
including what kinds of approaches, findings and arguments are incorporated into that
literature and presented as authoritative. Insofar as epidemiology is a policy science,
then, it is powerfully informed by the expectations and expediencies of policy, as
much as by the original scientific research on which it draws.

My research demonstrates that public health coalitions are a normal and productive
element of biomedical research. Lobbying for funding for research or treatment
programmes is a normal professional activity, and so is attempting to influence the
formation of public health policy and raise public awareness of particular health
conditions. Coalitions form as a normal aspect of the academic research process when
individuals and groups perceive their work as having a common concern with

314
important health issues. As well as a means of sharing information, these coalitions
also develop out of the sense of responsibility that often propels individuals to work in
public health and related research areas. The work of such coalitions is motivated by a
desire to improve society, and this accounts for some of the sense of campaigning that
their activities generate.

However, in the UK, the formation of such policy collations was influenced by the
NHS re-organisation in 1974 that severed the links between public health and local
authorities (see section 1.2.1). Throughout the period of my research, the resulting
institutional weakness of public health meant that clinical medicine researchers and
practitioners could claim expertise in areas that previously would have been the
domain of public health professionals. British health policy has typically been
developed by a relatively closed community of invited experts and, given the relative
prestige of clinical medicine, this facilitated such a takeover. Public health coalitions
appear to have developed around many of the major risk factors for heart disease,
including smoking, hypertension and physical activity. The coalition that developed
around excess body weight has been very successful but I think in one sense it can be
described as opportunistic: from the late 1960s, nutrition researchers, whose research
into malnutrition had suddenly become much less fashionable were looking for
interesting new research opportunities and formed alliances with research working on
heart disease and diabetes so that they could participate in the new area of chronic
disease epidemiology. British public health science went through a significant
reorientation when, following the undermining of the old public health institutions,
clinicians effectively got together to colonise the public health policy space. And this
had important implications for the way that public health policy science conceived of
health problems, including obesity and chronic disease.

Finally, my analysis of changes in the epidemiological and public health

understanding of obesity provides an empirically rich account of the emergence of
one expression of ‘new public health’. This supplements the work of authors such as
Petersen and Lupton who analyse and criticise the growth of individualism in
contemporary public health and its focus on the individual management of lifestyle to
prevent chronic diseases. Whilst I agree with many of their criticisms of
contemporary public health – particularly the stress on individual rather than

315
structural causes of ill-health, and the responsibility that is put on individuals to
maintain their own health – the lack of empirical detail in their account explaining
how the new public health came to replace older forms of public health is
problematic. The historical and professional contexts of the development of this
knowledge is omitted which gives their account a slightly detached and faintly
conspiratorial character. This is compounded by the fact that individual and group
motivation to adopt these new ideas and practices remains unaddressed throughout.
My research, however, provides a historical and institutional context for such
changes. It shows how such changes are the result of the shifting disciplinary
constitution of public health following the dismantling of environmentally-oriented
institutions of public health in the UK, the colonisation of public health by clinicians
and biomedical research scientists, and the exportation of that model into a WHO that
was under pressure from wealthy donor countries to identify public health problems
and solutions relevant to those countries.

10.4 Boundary-work and policy science in public health nutrition

Thomas Gieryn’s (1983, 1995, 1999) concept of boundary work argues that
researchers rhetorically define and re-define the boundaries of science in order to
present their knowledge claims as credible, reliable and trustworthy (see section 1.6).
These boundaries are context-specific, he argues, and thus highly contingent and
mutable, but they are constructed in order to maintain the epistemic authority of both
the individual making the claim, and their discipline, by demonstrating the knowledge
as well-founded and relevant to the specific context. In the preceding account of the
development of obesity science, I have outlined the ways in which researchers define
and redefine boundaries between public health nutrition and other kinds of knowledge
in order to ensure that the claims that they make could be seen as legitimate and
authoritative. In chapter 6, I described how a focus on individual behaviour led to the
exclusion of social and structural factors as explanations of eating behaviour. This
means that the explanations of obesity science drew on a relatively narrow
understanding of health (similar to the biomedical model discussed in section 9.4),
where the boundaries between the social and the biomedical were considered to be
quite firm. This narrower model was a central feature of the risk factor analysis that

316
came out of the Framingham study (see section 2.5) and, therefore, was a readily
available and prestigious approach for these researchers to use.

In chapter 8, I outlined how the failure of earlier models to fully explain increasing
rates of excess bodyweight, led to the incorporation of wider environmental causes
into a new model of obesity and overweight. Factors such as the widespread
availability of highly processed foods or decreasing requirements for energy
expenditure came to be seen as key causes of increasing average bodyweights. This
model involved a much wider understanding of the causes of health and disease more
akin to the social model of health (see section 9.4), and would seem to imply that
wide-ranging, large-scale changes and economically damaging changes would be
required to halt or reverse this trend. However, as I argued in section 9.2, the authors
of these reports did not call for such changes. Instead they recommended more
politically palatable changes relying on individual choice or voluntary codes of
practice, such as improved food labelling, increased regulation of advertising for
products targeted at children, nutritional standards and guidelines for institutional
catering, increased education in nutrition and in food preparation, and promotion of
exercise and physical activity.128 A narrower focus on individual causes allowed
obesity scientists to frame recommendations that largely relied on accepted health
promotion techniques in order to persuade individuals to change their behaviour, or
opportunistically incorporated facilities developed for other purposes such as road
safety or physical education in schools. This gap between an expansive set of causes
and a narrow and conservative set of solutions was a key example of the boundary-
work undertaken by obesity science researchers. It meant that their recommendations
largely focused on the area of personal behaviour, traditionally an area in which
medical expertise was considered authoritative, rather than exceeding the accepted
boundaries of their authority by advocating large-scale social and economic change.

Having illustrated that boundary-work has played a key role in the development of
obesity science, in this concluding discussion I will reflect upon the significance of
this phenomenon in greater depth. In doing so, I will also describe how this thesis has

128
More intrusive measures such as the creation of urban pedestrian walkways and cycle paths or
subsidies for fruit and vegetable growers were also listed but these were a minority of the measures and
were still relatively conservative compared to the potentially wide-ranging interventions suggested by
their analysis of the causes of excess body weight.(WHO, 2000: 269).

317
developed Gieryn’s concept of boundary-work and why this is a necessary and
important move in the analysis of obesity science.

The material that I outlined in section 1.3 demonstrates that such processes have been
present throughout the history of nutrition science. Since the late nineteenth, when
they contributed to debates about the link between poverty and poor nutrition or
minimum wage levels, nutrition researchers have had to work in an area that is
characterised by shifting boundaries between the political and the technical. This has
meant that their recommendations have often been controversial, and that individual
expert scientists have had to engage in ongoing sophisticated boundary-work to
maintain the distinction between nutritional and political judgements, and maintain
their credibility by demonstrating their scientific credentials or lack of commercial
funding. For example, the controversy about the protein levels necessary for health
accompanied the beginnings of nutrition science in the late nineteenth century and
recurred until the 1950s. It was linked to arguments about minimum wage levels,
since animal protein was an expensive and desirable component of poor peoples’ diet,
but also ideas about appropriate eating. In this period, several generations of nutrition
researchers tried, and often failed, to persuade the public to accept novel pulse-based
proteins as inexpensive substitutes for meat. This socially conservative form of
activism saw the role of nutrition as advising individuals on ‘intelligent’ food choices
rather than linking improved nutrition to increased wage levels: nutrition science
could be used to advise individuals on how to choose correctly in order to obtain the
maximum nutrients for the minimum cost, but, according to these accounts, it could
not be used to argue that wage levels were insufficient to obtain adequate levels of
nutrients. The latter was deemed to be a matter of political economy – to be decided
by employer and legislators - and so outwith the legitimate authority of nutrition
researchers.

The diet-heart disease hypothesis of Ancel Keys and others (see section 2.5) led to a
protracted controversy around the health effects of consuming foods such as butter,
eggs, milk and meat. This controversy (mostly) abated despite the absence of
definitive scientific proof, and in spite of the lobbying of influential food producing
groups, such as the dairy and egg industries (Garrety, 1997). Other researchers in the
field as well as representatives of particular food producer groups persistently

318
questioned the scientific evidence linking high rates of cholesterol consumption to
increased risk of heart disease and labelled low-fat diets as ‘food faddism’, that is
non-scientific and based on irrational fears (Garrety, 1997: 751). As Garrety points
out, because of the wide dissemination and practical relevance of nutrition science
‘the boundary between “food fads” and orthodox nutrition is particularly difficult to
maintain. Considerable energy has been – and is – devoted to keeping fad diets under
control’ (Garrety, 1997: 750-1). Because of their reliance on weight loss diets as a
treatment, obesity science researchers when working in an official capacity also have
to be careful to maintain the boundary between orthodox nutrition and fad diets.129
They have needed to give details of the biochemical research findings underlying their
dietary recommendations and outline the clinical results demonstrating their
effectiveness in order to differentiate them from wide a range of alternative
recommendations. These can be based on particular kinds of food (raw foods, whole
foods or commercial weight loss products), recommend abstention from specific
foods (vegetarian, vegan or fruitarian) or advocate a particular balance of foodstuff
(high-protein or macrobiotic). As many of these unorthodox diets also use nutritional
science to claim legitimacy, researchers have often had to spend time rebutting the
claims made on behalf of these ‘fad’ diets as well making very detailed evidence for
the legitimacy of their own recommendations.

Thus obesity science researchers produced their policy advice in the knowledge of
previous controversies about expert dietary advice and, in particular, of the role
played by commercially powerful interests within the food and agriculture sectors in
inciting and sustaining such disputes. Such commercial interests often promoted a
narrower model of health (in this case body size or physical fitness) that focused on
the role of individual responsibility rather than wider factors such as food availability
or advertising. Individual researchers needed to be able to successfully engage in
various kinds of boundary work in order to maintain their authority and avoid their
advice being discredited by being described unscientific. Examples of such boundary-
work included drawing on particular kinds of research results – often biochemical and
physiological - in order to substantiate their arguments; emphasising their credentials
129
However several of my participants did manage to translate their nutritional expertise into best-
selling diet books, including Alan Howard and Ian MacLean-Baird who developed the Cambridge Diet
(now the Cambridge Weight Plan).

319
as members of state-funded research centres and as governments scientists
recommending relatively uncontroversial dietary treatments and policy initiatives (see
above); and largely distancing themselves from campaigning groups or commercial
interests that might be seen as ‘contaminating’ their findings.

By making use of Gieryn’s concept of boundary-work in my research in this fashion, I

have extended his argument into a new empirical area. When describing boundary-
work Gieryn (1983, 1999) used examples from the history of the physical and
biological sciences, whereas my empirical material concerns the scientific knowledge
created in the making of public health policy.

In conclusion, I want to return to Sheila Jasanoff’s ideas about the distinctive nature
of policy science. As I outlined above (see section 1.6), Jasanoff (1990: 227)
describes policy science as a hybrid entity composed of many elements assembled in
response to the demands of answering key policy questions, but also as an important
influence on subsequent policy. I argue that one of the important characteristics of
policy science is that it necessarily requires more boundary-work than other forms of
science. There are two main reasons for this. Firstly, policy science does not appear to
be like other kinds of science or like the accepted model of scientific knowledge
creation: it does not take place in a clinic or a laboratory, is based on the results of
several different kinds of research and makes broad claims about the nature of
contemporary society. This means that extra work has to be done to make policy
science appear more scientific. Such activities may include stressing the expert
credentials of participants, reference to previous reports in the same area, an
invocation of moral languages or reference to rules of thumb such as, in
environmental contexts, the precautionary principle or guidelines for sustainability.

Secondly policy science often relates to contested scientific/technical arenas which

are also inhabited by other powerful interest groups who contest the scientific
legitimacy of specific claims and the individuals experts making them: examples of
this in public health nutrition case include food producers contesting specific nutrition
claims and producing counter-evidence about nutritional properties of particular
foods, such as meat or diary products, or of particular patterns of consumption, such
as low fat or whole-food diets. Nutrition researchers have often had the problem of

320
producing authoritative advice whilst negotiating the interests of powerful food
producers. However, this need for increased levels of boundary-work – including
providing evidence of elite scientific status, conventional political activity and
acceptable interactions with commercial interests - is not unique to public health
nutrition, as in many areas of environmental policy science claims made by
researchers are countered using similar methods.

I have argued that the nature of policy science means that it will require more
boundary-work (than other kinds of science) and some authors (Weinberg, 1972,
Jasanoff, 1990, Wynne, 1996) state that such boundary-work will often be based on
extra-scientific moral arguments, which can, in principle always be challenged by
alternative interpretations, which in turn leads to further boundary-work. Examples of
such extra-scientific moral arguments, or meta-scientific judgements, include the
precautionary principle in environmental policy and ideas about the un-healthiness of
certain body sizes in obesity science. These kinds of arguments are a routine feature
of science, but are used mostly frequently in contentious situations when evidence is
limited as they function to ‘fill in the gaps’ of particular arguments. But because
disputes involving such judgements cannot be settled by recourse to more scientific
knowledge – they cannot be settled at the level at which they are being conducted -
they are prone to recur regularly, as examples outlined in section 1.3 from the history
of nutrition clearly demonstrate.

I have demonstrated how boundary-work was pervasive in the attempt to provide

governments with credible public health recommendations in the area of diet and
nutrition and argued that this is because it is inevitably part of the creation of policy
science. However, my case study demonstrates that the increased levels of boundary-
work performed in this area of policy science have led to problems in the ways in
which research scientists participated in the making of recent public health policy on
diet and nutrition. Pressure to maintain expert authority in this highly contestable area
of science has led to attempts to erect firm boundaries around what counts as expert
knowledge in a way that has deleted important social factors. Following Wynne
(1993), I argue that this situation could be mitigated if obesity scientists were given
the political space in which to reflect on this problem i.e. the rhetorical strategies of
boundary-work that they have to use in order to maintain their authority as experts.

321
Such reflection might allow them to realise and acknowledge the ways in which their
own institutional processes lead to the deletion of important kinds of explanations and
sources of evidence, which in turn might make it possible to redefine the boundaries
of what counts as expert or authoritative public health policy knowledge and extend
the parameters of discussion to consider structural factors more fully. This might
involve incorporating existing national and international research into health
inequalities into their explanations of increasing average bodyweight. It could also
involve contacting patients and other advocacy groups in order to gain alternative
perspectives into the causes of excess body weight and the effect of current
treatments. Finally, it might involve examining the link between overweight and ill-
health and analysing how robust these links, and at what point the increased risks of
conditions such as diabetes and hypertension are more significant than the risks of
current treatments (see section 7.6)..

These pressures have been exacerbated by the tradition of basing participation in

policymaking on professional achievement and history of public service that has led
to a socially unrepresentative group of elite scientists developing British public health
policy in this area. Due to the post-war weakness of British public health (Lewis,
1986, Berridge, 1999a, Berridge, 2007a) these elite scientists have also largely been
inclined towards a narrow model of health that focused on individual behaviour as a
cause of excess bodyweight (see sections 6.6 and 9.2). Moreover, the UK approach of
private deliberation leading to the publication of consensus reports (see section 1.3) –
in contrast to the more open American system (Jasanoff, 2005: 262) - has also
accentuated the lack of wider public participation in the development of this body of
knowledge.

This pattern of private deliberation amongst elite researchers leading to the production
of consensus reports appears also to have been a feature of the production of WHO
technical reports. Despite the availability within the organisation of alternative, more
social, models of health (see section 9.4), this mode of knowledge production,
combined with the active participation of British and American nutrition researchers
and external political pressures on the WHO, led to partial accounts of the causes of
excess body weight that omitted important potential causes such as poverty and social
inequality. As in the examples I discussed above, in this case, the legitimate use of

322
scientific evidence from nutrition research results (and other biomedical research
results) is restricted to explanations of increasing average body weights largely
framed terms of individual behaviour. Explanations framed in terms of wider social
and economic change risk being described as ‘political’ and therefore not
scientifically legitimate. The authority of such partial explanations has been
reinforced by their congruence with culturally pervasive and highly moralised
‘common sense’ accounts of modernity, health and bodyweight (Gard and Wright,
2005). These accounts provide explanations of excess body weight that place undue
emphasis on factors such as irrational food choice or inappropriate leisure activities,
as opposed to more structural factors such as restricted food choice, due to low
income, or changes in work practices resulting from widespread mechanisation.

In the case of medical knowledge, the social processes of creating new policy science
have also been different from those in other areas of science, partly due to the
historical durability of medical authority in areas of health and individual conduct. As
accredited and government appointed medical experts, research scientists who are
members of consultations and expert committees have multiple and overlapping forms
of expertise upon which they can draw - including those of a medical practitioner, a
research scientist, a government scientists and as a member of the ‘great and the
good’ (Jasanoff, 1990: 229, Jasanoff, 2005: 268). This can be described as an ‘excess’
of authority upon which to draw, and by virtue of their position they are given the
authority to make very wide ranging judgements about the proper conduct of
individuals and the organisation of society that go well beyond their specific research
expertise (Jasanoff, 1990: 229, 237, see also Wynne, 1996). This is not an abuse of
their role, since the form of embodied expertise practiced in the British political
system is based on the exercise of such judgement (see section 1.6.2). Recruitment of
the ‘great and the good’ on the basis of professional and personal standing legitimates
the fact that these individuals and bodies are often asked to make political judgements
that go beyond the available science.130

130
In my account I have emphasised the difference between policy science and other kinds of science
(see section 15). However, policy science has remained an analysts’ rather than an actors’ category. It
is therefore an empirical question requiring further research to establish how expert scientists
understand their participation in such committees, other work involved in the development of public
health policy and the knowledge created during such processes. In particular it would be useful to
establish whether there is an actors’ category that maps onto the analyst’s one of ‘policy science’?

323
As I have explained earlier (see section (see sections 6.4 and 9.5), there are also
important gaps in contemporary biomedical knowledge about the causes and long-
term health effects of excess bodyweight in both individuals and populations. In this
way, obesity science fits Funkowitz and Ravetz’s (1993) definition of post-normal
science. However, their call for an ‘extended peer community’ to review such science
seems to be warranted not by the uncertainties of these bodies of knowledge, but by
the social justice issues deriving from the limited democratic participation in the
making of public health policy in the British political system. Precisely because many
of these judgements rely on extra-scientific and often moral arguments, they should
be assessed by a more socially inclusive body of participants to avoid the
stigmatisation of particular individuals and social groups.

In these respects the construction of current ideas about the obesity epidemic can
itself be seen as the outcome of the political process – driven by the work of well-
intentioned research scientists and medical practitioners concerned to improve the
health of the population, but unwittingly and implicitly shaped by political forces that
have led to the writing-out of overtly politically sensitive solutions from public health
policy.

10.5 Conclusion

Dominant accounts of present-day public health are framed in a broad historical

narrative that emphasises the importance of a number of ‘transitions’, all of which
serve to underline the idea that we live in a time of both peculiarly good health and of
specific forms of illness that can be attributed to ‘modernity’, itself understood rather
vaguely as a condition of economic and material plenty that is quite distinct from the
constitution of earlier ‘traditional’ societies. Obesity science has added to this
narrative the further concept of a ‘nutritional transition’, which brings both an
increase in the amount of food available, but also a shift from healthy traditional diets
and patterns of activity to more harmful lifestyles and forms of consumption.

324
The aim of my thesis has been to show how the story of this transition was
constructed through the work of a specific coalition of scientists, working within a
particular social and economic context of policy science. And in so doing, I have
sought to open up the possibility that other, more political accounts of obesity and
chronic disease might be seen as equally or more valid and appropriate to address the
health problems that confront the modern world.

325
APPENDIX 1

Principal primary sources

DATE TITLE AUTHOR

1969 Obesity: Medical and Scientific Aspects: Edited by I.M. Baird and A
Proceeding of the First Symposiums of Howard
the Obesity Association
1969 Obesity and Disease Office of Health Economics
1969 Obesity and its Management Denis Craddock
1977 Research on Obesity Department of Health and
Social Security/Medical
Research Council
1978 Obesity and its Management (3rd edition) Denis Craddock
1978 Energy Balance and Man John Garrow
1982 The Prevention of Coronary Heart WHO (technical report 678)
Disease
1984 Obesity (Clinics in Endocrinology and W P T James
Metabolism volume 13 no 3)
1985 Body Weight Control Edited by AE Bender and LJ
Brookes
1986 Community prevention and control of WHO (technical report 732)
cardiovascular diseases
1988 Obesity and Related Disease John S Garrow
1988 Healthy Nutrition: Preventing nutrition- WPT James with A Ferro-Luzzi,
related diseases in Europe B Isaksson and WB Szostak
(WHO Regional Publications,
European Series no 24)
1990 Diet, nutrition and the prevention of WHO (technical report 797)
chronic diseases
1994 Obesity Richard West for the OHE
1995 Obesity: Reversing the Increasing Department of Health
Problem of Obesity in England
1995 Physical Status: The Use and WHO (technical report 854)
Interpretation of Anthropometry
1996 Obesity in Scotland: Integrating Scottish Intercollegiate
Prevention with Weight Management Guidelines Network (SIGN)
1999 Obesity British Nutrition Foundation
2000 Obesity: Preventing and Managing the WHO (technical report 894)
Global Epidemic

326
 APPENDIX 2

$ ! %( 1$%$ $/% * ! %& !- 6' ' 3 % / ! -#

6 -/ ) 7, )!/8 ' 5! / &/%' 96 %& ' )$%($*$(6 ! / 4 * % $/% &
/ , %$' $/% 5! ' / -$,- $%& '$%,. )'-/ !/%, !$' . %*$ /%3 % . &/' '
. )-$,- !/8.
1969 Obesity: Ponderal Not known Diabetes, Individual Treatment Not
Medical and index + coronary artery discussed
Scientific MLC disease,
Aspects: weight hypertension,
Proceeding of tables osteoarthritis
the First /relative
Symposiums of weight
the Obesity
Association
(eds I.M. Baird
& A. Howard)
1969 Obesity and Relative 20% Diabetes, Individual Treatment Low costs
Disease /ideal overweight, possibly CHD, largely to
(Office of weight 5% obese osteoarthritis individuals
Health
Economics)
1969 Obesity and its Relative 50% Coronary artery Individual Treatment Not
Management /ideal overweight disease, (with brief discussed
(Denis weight hypertension, discussion
Craddock) cardio- of social
respiratory factors)
disorders,
diabetes
1977 Research on MLC Small increase Diabetes, Individual Discusses Not
Obesity weight in prevalence possibly CHD, (but also prevention discussed
(Department tables/ since 1943 hypertension, related to but focus is
of Health and relative surveys osteoarthritis social class) on treatment
Social weight
Security/Medi plus
cal Research W/H²
Council)
1978 Obesity and its MLC 50% of men CVD, respiratory Individual Treatment Not
Management weight and women insufficiency, (with brief discussed
(3rd edition) tables/ overweight diabetes, discussion
(Denis relative gallstones, liver of social and
Craddock) weight damage, environment
osteoarthritis al factors)
1978 Energy MLC Not discussed Not discussed Individual Discusses Not
Balance and weight prevention discussed
Man (John tables/ but focus is
Garrow) relative on treatment
weight
1983 Obesity (Royal MLC 31% of men CHD, Individual Treatment Not
College of weight and 27% of hypertension, (with brief discussed
Physicians) tables/ women diabetes, discussion
relative overweight gallbladder of social
weight 5.4% of men disease, cancer, factors)
and breathing
6.4 % of difficulties
women obese
1984 Obesity BMI + Not discussed CHD, diabetes, Individual Treatment Not
(Clinics in Garrow’s joint disorders, discussed
Endocrinology cut-off hypertension,

327
 and points Respiratory
Metabolism disorders
vol 13 no.3)
(W Philip
James)
1985 Body Weight MLC 24.2 % of men CVD, diabetes, Individual Treatment Not
Control (eds weight and 27.1% of hyper-tension, (also related discussed
AE Bender tables women gout, to ethnicity)
and LJ /relative overweight hyperinsulinaemi
Brookes) weight (from a,
plus ref to NHANES II) raised blood
BMI cholesterol,
locomotor and
psychiatric
disease
1988 Obesity and BMI + Men: 34 % Heart disease, Individual Treatment Not
Related Garrow’s overweight, diabetes, (brief (plus brief discussed
Disease (John cut-off 6% obese hypertension, discussion discussion
Garrow) points Women: 24% gall- bladder of social of
overweight, disease, factors) prevention)
8% obese respiratory
insufficiency,
osteoarthritis
1994 Obesity BMI + Men: CHD, stroke, Individual Treatment First details
(Richard West Garrow’s overweight cancer, diabetes, (brief (plus brief of costs to
for the Office cut-off 40%, obese, hypertension, discussion discussion healthcare
of Health points 13% hypercholesterol of socio- of services:
Economics) Women: aemia, economic prevention) £29 million
overweight, gall bladder factors) in UK, $39
29%, obese, disease, osteo- billion in US
15% arthritis, gout,
ovulatory failure,
menstrual
irregularities and
PCOS**
1995 Obesity: BMI + UK Men: CVD, certain Individual Gives equal Not
Reversing the Garrow’s obese, 13% cancers, (but weight to discussed
Increasing cut-off UK Women: gall bladder discusses treatment
Problem of points obese, 15% disease, relation to and
Obesity in US Men: hypertension, social class prevention
England obese 31% and a range of and
(Department US Women: bone skin and ethnicity)
of Health) obese: 35% joint disorders
1996 Obesity in BMI + Scottish men: Diabetes, Individual Gives Not
Scotland: Garrow’s obese, 14%, hypertension, roughly discussed
Integrating cut-off overweight, stroke, equal weight
Prevention points 44% hyperlidiaemia, to treatment
with Weight Scottish CHD, gallstones, and
Management women: obese, cancers, prevention
(Scottish 17% and breathlessness,
Intercollegiate overweight, respiratory
Guidelines 32% disease and sleep
Network) apnoea,
menstrual
abnormalities,
pregnancy
complications,
musculo-skeletal
disorders and
arthritis, stress

328
 incontinence
1999 Obesity BMI + Between 1980 Diabetes, Individual Treatment Not
(British Garrow’s and 1996 coronary heart (plus one discussed
Nutrition cut-off prevalence in disease, cancer, chapter on
Foundation) points UK has more osteoarthritis, prevention)
than doubled gallstones, sleep
from 6 to 16% apnoea,
in men and 8 reproductive
to 18% in disorders,
women pregnancy
complications,
psychological
disorders
2000 Obesity: BMI + Prevalence of Greatly Individual Gives equal Contains
Preventing Garrow’s overweight increased risk: and weight to detailed
and Managing cut-off and obesity is diabetes, environment treatment discussion
the Global points increasing gallbladder and of economic
Epidemic (with globally at an disease, prevention costs in
(WHO discussion alarming rate dyslididaemia, developed
Technical about insulin and
Report 894) problems resistance, developing
of breathlessness, countries
applying sleep apnoea
to Moderately
different increased risk:
population CHD,
s) hypertension,
osteoarthritis,
hyperuricaemia
and gout
Slightly
increased risk:
some cancers,
reproductive
hormone
abnormalities,
PCOS, impaired
fertility, low
back pain,
anaesthetic
complications,
foetal defects
(due to maternal
obesity)

* MLC is an abbreviation for Metropolitan Life Company (see section 3.4.1 for a
discussion of their ideal weight tables)
** PCOS is an abbreviation for polycystic ovary syndrome

329
APPENDIX 3

Fletcher, Isabel (2010) Will the entire population be overweight by 2230? Common
sense, scientific consensus and the ‘obesity epidemic’ in Mathar, Thomas and Jansen,
Yvonne J.F.M (eds.) (2010) Health Promotion and Prevention Programmes in
Practice: How Patients’ health Practices and Rationalised, Reconceptualised and
Reorganised, transcript Verlap, Bielefield

!
" # $ % !
& ' & '" ( ) * + ,

"
"
)! -../0 123-+ *
"

4(
"
3
"5 )# ( 62220 -+

7 8 # ( )# (+
9 "(
9
" ,
"

7 9
3
"7 8
) 9 %:+
; -../
0

4( 3
" 7 <== 3

330
 "> -.?2
-..- 66=2 -22@ %!
* 61"A 61"= BCD
7
" =E A.E -.?? E1A/
"$ 3
"5 )F G , -../0 -=E+

7 ,
,
! " !
& 'H ,
" *
& '

" 7

"(
"

7
! 7 !
" ( *

"

" , 7 :
! ;
" F *
0
"

% ">
& '
" , , # ( A.E
6222
7 "! =
, ,
"

> -.?.
( $ )( $ -.?.+ ,3 ,"

331
7
" 7 6222 # ( *
0 ( $
%: 3
# ( -=--=6" 7 E
"I *
# ( 6222
& ' & ' "

7 "7 H
-==" 7
J J ;
"J ; ( 7 ,
-.12
, # ( 9 "

9
"F
, "

"

* ,
# ( # ( %:
) + -..= 6226" *

"

"

I ! 7 ! )!7!+

" 7 !7!

-=- %: F
! 6221
-=6 7 K L J ! G
M ,
-== ( $
> $ * "
> M F -..1

332
 " "
( 9 * G
)G -..1+"
!7! * ) +
"# G # )622/+
, !7!
, "

7 "(
& ' "
7 , 3
*
)L 622E G # 622/ ( 622?+
"7

, 7 : ): -.12+
" : '
H ,
"7
-.?2

, " : '
!7!" *
* ,
& ' ): 3L -...+"

# 9
3
3 ); 622/+" 7 ,
,
" 3 3
, " *
, ! ; * ); 622/
; -..2+" ! )6222+ ,
M ! ,

* %:-=E"

-=E *
) + )
+ K : ' ):
622/+"

333
7 8
" # !7!
" , I F M, K
)K 6221+" J
8
)J -../+H N ; ,"
7 , K "(
,
)> , ! -... , -...+" ,
* * 4,
5) , 622? + "

I $ F F *'
! )F
F * -..6+" F * F

) F F * -..60 .+" 7
!7!"

>
"#
)!
-... ! -.A? ! 6226+" 7 *
*
3
"; ! )-../+
-.12
3
" -=/

*
* 9 , ): 622. + 3
) ! 622A+"

7 62
)I -..A , 622- , 622/+" # K )622=+

, "
*
*

-=/ %: 8
3
& '

334
 )K 622=0 -.6362.+"
&
' ,
* * " 7
, K ) -..A+
& , ' L F !
" ,

, %! -./2 -.?2 ) -..A0

--=+"

" 8
3 & '" K
"
* *
H

)K -..6 K -..A+" 7 4 * 5

"-=?

>
3 )L + *
" ,
, 3 )ML +
"
9 %: $ %! -.
62 "!

3
& ' "7 & '
*

-=? * M,> I, I )> I 622=+

*
9

335
 )I I ! 622-0 .+"-=1 7
ML ,
& '
)K -..A+"

,
"
)K 622=+" -./2
& '
"> , 9
$ "L
3
)K 622=0 /+" !
* * " L
,
3 "
F ! L
) -..A+" F

, )K 622=0
61.3A/+"-=A

) K
622=0 =E=+ & '
* "
& ' )# ( 6222+" 7

3 "

L "
, ,

137
7 & '
"L I : )-.1.+
)
"
138

L
&7 ') -..A
K , 6222+"

336
 "
) -..A0 --?+"-=.

"
) +
, "
" 3 ,
,

) E+

F -.E2 -.?.0 ,

, F !
! L ! )-.EA -./A+ > > J
! )-./.+ "7 9 %!
# (" 7
I L "
8
3 O--2@ O-62@ 0
"

7 %! %:
* > "% ,
F ) -..A+ ,

) * +"
7 ,
"
3 ) + "

! -.12 -..2 0

7 # ( I(M L )I I(M
L + 9 -.1." 7
$ "(
* %: ) !! IKL -.1? KLJ -.A=+H
, ,

139
7 *

337
 " >I ) *+-E2
)>I O 6/ O =2 + -.1=
G >
0 "

8
)" " +
F ! L !
" 7

* " K
8 3 )
+"

7 -../ 62260

!
G M , )! GM+ -../
( 7 , ) (7F+ -..? M ( F )M(F+
6222" 9 # ( I(M L
6222" 7 # ( , -..1 7 K A.E
6222H 8 & ' 8
"-E- # (
&
' )I -..?0 .+" 7

"

( 9 *
)
+" !

, " K

9 *
3 "

"7 >I

-E2 7 * # P" F
Q -. , Q *"
, # P >I " , ' "
-E- ( & ' )I
" 622E+ & '

338
 # (
" " !"$" "7

3 "

F 622= 0

7
%: F 9 "
! 6221" $ J
G$M$! ) ( 3
9 + J(KGK(# 9
"

7
7 "(

3 & ' )$ ! -..1+

) +" $

"
, *
3 " 7
"

*
"

9 "

, 62 )# (
62220 -2-+"

!
9 "
7
,
"7 *
-..2 )! 622=+
"

"

339
7 8
H 8 H
, H
" ,

"7
*

"> * )" "

+ H
3 * " 7 ' *
): -.12+
,
*"

-.A2
9 0&
' )# ( 62220 -+"

*
" 7 - ) +
) 3-..2 +
) 3-..2 + "

340
 Environmental/
Individual paradigm population paradigm
Incidence/prevalence of obesity
Low and stable High and increasing
prevalence so an incidence so a population
individual health problem /public health problem
Measurement and classification of bodies
Measured by weight and Measured by BMI
frame size

Accepted consequences of obesity to health

Knowledge of health Knowledge of health
risks derived from case risks comes from large
histories and insurance scale epidemiological
company tables studies
Assumed causes of obesity
Due to overeating and Due to inappropriate diet
lack of exercise and sedentary lifestyle

Treatment/prevention
Corrected by dietary Solutions involve
restrictions changing food supply and
restructuring urban
environment
Cost of obesity
Cost of obesity limited Cost of obesity becomes
and mostly born by significant problem for
individuals national healthcare
systems

Table 1: The two paradigms of obesity

-"

7 *

" %:
& ' 62@ )( $ -.?.0 /+"
%: >I
?@ A@ =E@ 6E@
)KLJ -.A=0 6-+"

9
0 9 ) +
, " -..E
( $ %: &
9
' )# -..E0 /+" 7 -../

341
-=@ -?@
) -../0 6+

# ( # ( I(M L 9
$ M M$! ) %!+"
8 "
"
$
-/@ -?"/@ H L 6@ -@ H
%! 62@ 6/@ # ! /A@
11@ )# ( 62220 66+" K "

6"

7
0 &7 ,
" %! 6?

' )( $ -.?.0 E+" K

"7 I
L 0& 3 3
-62 4 5' ) !! IKL -.1?0 E+"
--2 --.@
-62@ )KLJ -.A=0 E+" 7 I L
-E6" 7
"

7 >I -E= "

7 H
"7 P
! ! ) !!+ I K L )IKL+ -.1?
> '
-EE " (
$ -..E -..? ! G M , )! GM
-..?+ > ' # (

H , 3
8
)# ( 62220 A+"

-E6 F / . -=? -E/ -/-3-12

H / / ---3-. -6/3-E6 )( $
-.?.0 /+"
-E= >I # P " " 8
-EE >I 6236E". H >I 6/36.". 0 >I =23E2
>I OE2 )# ( 62220 .+"

342
=" 7

(
G )G # 622/0 123-+"
3
" " )( $ -.?.0 --+"
* "
&
'
) !! IKL -.1?0 =2+"

,
"-E/ 7
K L J ' 0 &>

' )KLJ -.A=0 6E+"

"7 # (

"7
* &7
8

' )# ( 62220 -26+"

E"

F ,
8 , , I
L & '
& ' )( $ -.?.0 E+" 7
> > J ! )! -./.+
-2@ * * 62@
62@ )( $ -.?.0 -/
!! IKL -.1?0 6-+" 8 " " 9
" "
"

> K L J K -.A= 8
F
L ! "7

-E/ 7 , 3
,
, )I # ,
-...+

343
 , )KLJ -.A=0 .+"
> # (
8 & '
, & ' )# (
62220 E1+" 7
"

/" 7

7
"
H 9
"
,
& ' )( $ -.?.0 6E+-E?"
* 0 &

8 ' )( $ -.?.0 66+

, & ' )KLJ -.A=0 /2+

"7 -../
) -../+"
*
>I , M !
, ) -../0 -2+" 7
# (
&
' )# ( 62220 -/E+"

?" L

, GJ
" 7 R="/
R= 3 )( $
-.?.0 6/+"-E1 7 ( $
M ! R6."=/ ) + 3
3
)# -..E0 =.+"

-E? 7 -.?2 H -.A2

)KLJ -.A=0 1?+
-E1 7
"

344
> # ( 8
6 1@
-6 222 F -..6 </? =22 %! -.A?" 7
' &
* ' )# ( 62220 1.+"

"

"
: )-.12+ 3 *

"
)> -.A60 A?31+"

"7
,
# ("
& 3 ' ); -..2+" %
& ' ,
7 : ! ; ) " ; 622/+

"
, "

7 * )
+
"
J J
;
" > *
"

K
3 , , & 3
,'
)K 6221+"
"

345
 8 "7
3

" "
"7

* "

"

"
ML
*
# # 6" 7 9
, "
L ML
F ! ,
"(
,
I "
( ,
"

8 "

"

$ M "

*
,
* , " 7

# (
9
" ,
"
, " M

346
 "J
, ,
, " -EA

I
!7!
)> -.1?+
8 "!
)F F * -..6+ ,
"
7

"

-EA 7 * ,
) +
"

347
BIBLIOGRAPHY

ACS (2010) Cancer Prevention Study Overviews. American Cancer Society

www.cancer.org/docroot/res/content/res_6_2_study_overviews.asp.

AMSTERDAMSKA, O. (2001) Demarcating Epidemiology. Science, Technology and

Human Values, 30, 17-51.

AMSTERDAMSKA, O. (2005) Demarcating Epidemiology. Science Technology and

Human Values, 30, 17 - 51.

ANON (1997) WHO: where there is no vision the people perish (editorial). The
Lancet, 350, 749.

APFELBAUM, M. (1987) The mechanism of weight control. IN BENDER, A. E. &

BROOKES, L. J. (Eds.) Body Weight Control: The Physiology, Clinical Treatment
and Prevention of Obesity. Edinburgh, London, Melbourne and New York, Churchill
Livingstone.

APHRAMOR, L. (2005) Is a Weight-Centred Framework Salutogenic? Some

Thoughts on Unhinging Certain Dietary ideologies. Social Theory and Health, 3, 315-
40.

ARONOWITZ, R. A. (1998) Making Sense of Illness: Science, Society and Disease,

Cambridge, Cambridge University Press.

ARONSON, N. (1982) Nutrition as a social problem: a case study of entrepreneurial

strategy in science. Social Problems, 29, 474-87.

ASHLEY, F. W. & KANNEL, W. B. (1974) Relation of weight change to changes in

atherogenic traits: the Framingham Study Journal of Chronic Diseases, 27, 103-114.

ASHMORE, M., MULKAY, M. & PINCH, T. (1989) Health and Efficiency: A

Sociology of Health Economics Milton Keynes Philadelphia, Open University Press.

ASO (2011) A network for obesity researchers in the UK. Association for the Study
of Obesity http://www.aso.org.uk/.

ATKINS, P. & BOWLER, I. (2001) Food in Society: Economy, Culture, Geography

London, Arnold.

ATRENS, D. M. (1994) The Questionable Wisdom of a Low-fat Diet and Cholesterol

Reduction. Social Science and Medicine, 39, 433-47.

AUSTIN, S. B. (1999) Fat, Loathing and Public Health: The Complicity of Science in
a Culture of Disordered Eating. Culture, Medicine and Psychiatry, 23, 245-268.

348
BACON, L. (2008) Health at Every Size: The Surprising Truth about Your Weight,
Dallas, TX, Benbella.

BAGGOTT, R. (2000) Public Health: Policy and Politics, Basingstoke and London,
Macmillan Press Ltd

BAIRD, I. M. (1969) Obesity in Adults: A Review of the Current Position. IN

BAIRD, I. M. & HOWARD, A. N. (Eds.) Obesity Medical and Scientific Aspects:
Proceedings of the First Symposium of The Obesity Association of Great Britain
Edinburgh and London, E&S Livingstone Ltd.

BARKER, D. J. P. (2007) Symposium: The origins of the developmental origins

theory. Journal of Internal Medicine, 261, 412 - 7.

BARNES, B. (1982) T S Kuhn and Social Science, New York, Columbia University
Press.

BARNES, B., BLOOR, D. & HENRY, J. (1996) Scientific Knowledge: A

Sociological Analysis, London, Athlone.

BARTLEY, M. (1985) Coronary Heart Disease and the Public Health 1850 -1983.
Sociology of Health and Illness, 7, 289-313.

BARTLEY, M. (1992) Authorities and Partisans: The Debate on Unemployment and

Health Edinburgh, Edinburgh University Press.

BARTLEY, M. (1994) The relationship between research and policy: the case of
unemployment and health. IN OAKLEY, A. & WILLIAMS, A. S. (Eds.) The politics
of the welfare state. London, UCL Press.

BARTLEY, M. (1996) 'Probably, Minister...': the 'Strong Programme' Approach to

the Relationship between Research and Policy. IN SAMSON, C. & SOUTH, N.
(Eds.) The Social Construction of Social Policy: Methodologies, Racism, Citizenship
and the Environment. Basingstoke and London, Macmillan Press.

BARTLEY, M. (2004) Health Inequality: An Introduction to Theories, Concepts and

Methods, Cambridge, Polity Press.

BDA (1970) British Diabetic Association. Autumn Meeting Royal College of

Physicians, London, Diabetologia.

BEAGLEHOLE, R., BONITA, R. & KJELLSTROM, T. (1993) Basic Epidemiology,

Geneva, World Health Organization.

BENDER, A. E. & BROOKES, L. J. (Eds.) (1987) Body Weight Control: The

Physiology, Clinical Treatment and Prevention of Obesity, Edinburgh, London,
Melbourne and New York, Churchill Livingstone.

BERLIVET, L. (2003) “Association or causation?”The debate on the scientific status

of risk factor epidemiology 1947 – c.1965. IN BERRIDGE, V. (Ed.) Making health
policy; networks in research and policy after 1945. Amsterdam, Rodolphi.

349
BERRIDGE, V. (1999a) Health and Society in Britain since 1939, Cambridge,
Cambridge University Press.

BERRIDGE, V. (1999b) Passive smoking and its pre-history in Britain: policy speaks
to science? Social Science and Medicine, 49, 1183-95.

BERRIDGE, V. (2000) History in public health: who needs it? The Lancet, 356,
1923-5.

BERRIDGE, V. (2003a) Introduction: Inequalities and Health. IN BERRIDGE, V. &

BLUME, S. (Eds.) Poor Health: Social Inequality before and after the Black Report.
London and Portland, OR, Frank Cass.

BERRIDGE, V. (2003b) Post-war Smoking Policy in the UK and the Redefinition of

Public Health. Twentieth Century British History, 14, 61-82.

BERRIDGE, V. (2005a) Issue Network vs Producer Network? ASH, the Tobacco

Products Research Trust and UK Smoking Policy IN BERRIDGE, V. (Ed.) Making
Health Policy: Networks in Research and Policy after 1945. Amsterdam/New York,
NY, Rodopi.

BERRIDGE, V. (2005b) Making Health Policy: Networks in Research and Policy

after 1945. IN BERRIDGE, V. (Ed.) Making Health Policy: Networks in Research
and Policy after 1945. Amsterdam and New York, Rodopi.

BERRIDGE, V. (2006) The Policy Response to the Smoking and Lung Cancer
Connection in the 1950s and 1960s. The Historical Journal, 49, 1185-1209.

BERRIDGE, V. (2007a) Marketing Health: Smoking and the Discourse of public

Health in Britain, 1945-2000, Oxford, New York, Oxford University Press.

BERRIDGE, V. (2007b) Medicine and the Public: The 1962 Report of the Royal
College of Physicians and the New Public Health. Bulletin of the History of Medicine,
81, 286-311.

BERRIDGE, V. (2008) History Matters? History’s Role in Health Policy Making.

Medical History, 52, 311-26.

BERRIDGE, V. & LOUGHLIN, K. (2006) Smoking and the New Health Education
in Britain 1950s to 1970s. American Journal of Public Health, 95, 956-64.

BHOPAL, R. (2002) Concepts of Epidemiology Oxford, New York, Oxford

University Press.

BLACKBURN, H. (n.d.) Overview: The Seven Countries Study in Brief. University

of Minnesota http://www.epi.umn.edu/research/7countries/overview.shtm.

BLAXTER, M. (1990) Health and Lifestyles, London and New York,

Tavistock/Routledge.

BLAXTER, M. (2010) Health (2nd edition), Cambridge and Malden MA, Polity.

350
BLOOR, D. (1976) Knowledge and Social Imagery, Routledge and Kegan Paul,
London, Henley and Boston.

BNF (1999) Obesity: The Report of the British Nutrition Foundation Task Force.
Oxford, London and Edinburgh, Blackwell Science.

BNF (2004a) About BNF/What is BNF? , British Nutrition Foundation

http://www.nutrition.org.uk/home.asp?siteId=43&sectionId=345&parentSection=305
&which=7.

BNF (2004b) Publications/BNF Taskforce Reports

http://www.nutrition.org.uk/home.asp?siteId=43&sectionId=342&parentSection=303
&which=6. British Nutrition Foundation.

BOERO, N. (2007) All the New that's Fat to Print: The American "Obesity Epidemic"
and the Media. Qualitative Sociology, 30, 41-60.

BORDO, S. (1993) Unbearable Weight: Feminism, Western Culture and the Body,
Berkeley, LA and Londn, University of California Press.

BOROWY, I. (2009) Coming to Terms with World Health: The League of Nations
Health Organisation 1921-46, Frankfurt am Main, Berlin, Bern, Bruxelles, New
York, Oxford, Vienna, Peter Lang.

BOWKER, G. C. & LEIGH STAR, S. (1999) Sorting Things Out: Classification and
Its Consequences, Massachusetts, MIT Press.

BRANCA, F., NIKOGOSIAN, H. & LOBSTEIN, T. (2007a) The Challenge of

obesity in the WHO European Region and the Strategies for Response. Copenhagen,
WHO Regional Office for Europe.

BRANCA, F., NIKOGOSIAN, H. & LOBSTEIN, T. (Eds.) (2007b) The challenges of

obesity in the WHO European region and the strategies for response (summary),
Copenhagen, World Health Organization Regional Office for Europe.

BRAY, G. A. (Ed.) (1979a) Obesity in America, Bethesda, Maryland, National

Institutes of Health.

BRAY, G. A. (1979b) Overview. IN BRAY, G. A. (Ed.) Obesity in America.

Bethseda, Maryland, National Institutes of Health.

BROWN, N. & MICHAEL, M. (2003) A Sociology of Expectations: Retrospecting

prospects and prospecting retrospects. Technology Analysis and Strategic
Management, 15, 3-18.

BROWN, T. M., CUETO, M. & FEE, E. (2006) The World Health Organization and
the Transition from "International" to "Global" Public Health. American Journal of
Public Health, 96, 62-72.

BROWNELL, K. D. & HORGEN, K. B. (2004) Food Fight: The Inside Story of the
Food Industry, America's Obeisty Crisis and What We Can Do About It, New York,
McGraw-Hill

351
BRUCH, H. (1974) Eating disorders: obesity, anorexia nervosa and the person
within, London, Routledge and Kegan Paul.

BUFTON, M. W. (2005) British Expert Advice on Diet and Heart Disease c.1945-
2000 IN BERRIDGE, V. (Ed.) Making Health Policy: Networks in Research and
Policy after 1945. Amsterdam and New York, Rodopi.

BUFTON, M. W. & BERRIDGE, V. (2000) Post-war nutrition science and policy

making in Britain c.1945-1994: the case of diet and heart disease. IN SMITH, D. F. &
PHILLIPS, J. (Eds.) Food, Science, Policy and Regulation in the Twentieth Century.
London and New York, Routledge.

BUFTON, M. W., SMITH, D. F. & BERRIDGE, V. (2003) Professional Ambitions,

Political Inclinations, and Protein Problems: Conflict and Compromise in the BMA
Nutrition Committee 1947-50. Medical History, 47, 473-92.

BUSE, K., MAYS, N. & WALT, G. (2005) Making Health Policy, Maidenhead and
New York, Open University Press.

BUTTERFIELD, W. J. H. (1969) Introduction. IN BAIRD, I. M. & HOWARD, A. N.

(Eds.) Obesity Medical and Scientific Aspects: Proceedings of the First Symposium of
The Obesity Association of Great Britain. Edinburgh and London, E&S Livingstone
Ltd.

BYNUM, W. F. (2008) The McKeown thesis. The Lancet, 371, 644-5.

CAMPOS, P. (2004) The Diet Myth: Why America's Obsession with Weight is
Hazardous to Your Health, New York, Gotham Books.

CAMPOS, P., SAGUY, A. C., ERNSBERGER, P., OLIVER, J. E. & GAESSER, G.

A. (2006) The epidemiology of overweight and obesity: public health crisis or moral
panic? . International Journal of Epidemiology, 35, 55-60.

CANNON, G. (1987) The Politics of Food, London, Century.

CAPLAN, P. (Ed.) (1996) Food, Health and Identity, London and New York,
Routledge.

CDC (2010) Overweight and Obesity. Data and Statistics. U.S. Obesity Trends.
Trends by State 1985-2009. Centers for Disease Control
http://www.cdc.gov/obesity/data/trends.html#State.

CHANG, V. W. & CHRISTAKIS, N. A. (2002) Medical modelling of obesity: a

transition from action to experience in a 20th century American medical textbook.
Sociology of Health & Illness, 24, 151-77.

CHERNIN, K. (1981) Womansize: the tyranny of slenderness, London, The Women's

Press.

CHEUNG, M. H. (2010) Asia-Pacific faces diabetes challenge. The Lancet, 375,

2207-10.

352
COGAN, J. C. (1999) Re-evaluating the Weight-Centred Approach towards Health:
The Need for a Paradigm Shift. IN SOBAL, J. & MAURER, D. (Eds.) Interpreting
Weight: The Social Management of Fatness and Thinness. New York, Aldine de
Gruyter.

COHEN, S. (2002 [1972] ) Folk Devils and Moral Panics, London, Routledge.

COLDITZ, G. (1992) Economic costs of obesity. American Journal of Clinical

Nutrition, 55, 503s-507s.

CONNELLY, J. & WORTH, C. (1997) Public Health Medicine, Abingdon, Oxon,

Radcliffe Medical Press.

CONRAD, P. (1992) Medicalizaton and Social Control. Annual Review of Sociology,

18, 209-32.

CRADDOCK, D. (1969) Obesity and its Management, Edinburgh, E & S Livingston.

CRADDOCK, D. (1978) Obesity and Its Management Edinburgh London and New
York, Churchill Livingstone.

CRAWFORD, R. (1980) Healthism and the medicalization of everyday life.

International Journal of Health Services, 10, 365-88.

CRITSER, G. (2003) Fat Land; How Americans Became the Fattest People in the
World, Harmondsworth, Penguin Books.

CROTTY, P. (1995) Good Nutrition? Fact and fashion in dietary advice, St

Leonards, NSW, Allen and Unwin.

DAVIS, A. B. (1981) Life Insurance and the Physical Examination: A Chapter in the
Rise of American Medical Technology. Bulletin of the History of Medicine, 55, 392-
406.

DAVISON, C., FRANKEL, S. & DAVEY-SMITH, G. (1992) The Limits of

Lifestyle: Re-Assessing "Fatalism" in the Popular Culture of Illness Prevention.
Social Science and Medicine, 34, 675-85.

DAWBER, T. R. (1960) Summary of recent literature regarding cigarette smoking

and coronary heart disease. Circulation 22, 164-6.

DAWBER, T. R. (1980) The Framingham Study: The Epidemiology of

Atheroschlerotic Disease, Cambridge MA and London, Harvard University Press and
the Commonwealth Fund.

DE GAURINE, I. (1993) Obesity in a multicultural perspective. IN GEISSLER, C. &

ODDY, D. J. (Eds.) Food, Diet and Economic Change Past and Present. Leicester,
London and New York, Leicester University Press.

DELORMIER, T., FROHLICH, K. L. & POTVIN, L. (2009) Food and eating as

social practice – understanding eating patterns as social phenomena and implications
for public health. Sociology of Health & Illness, 31, 215-28.

353
DESROSIERES, A. (1998) The Politics of Large Numbers: A History of Statistical
Reasoning, Cambridge MA and London, Harvard University Press.

DHSS (1974) Diet and Coronary Heart Disease: Report of the Advisory Panel of the
Committee on Medical Aspects of Food Policy (Nutrition) on Diet in relation to
Cardiovascular and Cerebrovascular Disease. London, Department of Health and
Social Security.

DHSS/MRC (1976) Research on Obesity: A Report of the DHSS/MRC Group.

London, Department of Health and Social Security/Medical Research Council.

DOH (1992) The Health of the Nation: A Strategy for Health in England. London,
Department of Health.

DOH (1995) Obesity: Reversing the Increasing Problem of Obesity in England.

London, The Nutrition and Physical Activity Task Forces.

DOH (2004) Choosing Health: Making Healthy Choices Easier (Executive

Summary). London, Department of Health.

DOLL, R. & HILL, A. B. (1950) Smoking and Carcinoma of the Lung. British
Medical Journal, 212, 739-48.

DOWLER, E., CARAHER, M. & LINCOLN, P. (2007) Inequalities in food and

nutrition: challenging 'lifestyles'. IN DOWLER, E. & SPENCER, N. (Eds.)
Challenging Health Inequalities. Bristol, Policy Press.

DREWNOSKI, A. & POPKIN, B. M. (1997) The Nutrition Transition. Nutrition

Reviews, 55, 31-43.

EC (2005) Diet, Physical Activity and Health - A European Platform for Health.
Brussels, European Commission.

EGGER, G. & SWINBURN, B. (1997) An “ecological” approach to the obesity

pandemic. British Medical Journal 315, 477 – 80.

EPSTEIN, S. (2007) Inclusion: The Politics of Difference in Medical Research,

Chicago and London, The University of Chicago Press.

ERNSBERGER, P. & HASKEW, P. (1987) Health Implication of Obesity: An

Alternative View. The Journal of Obesity and Weight Regulation, 6, 58-137.

EYTON, A. (1987) Self-help lay groups. IN BENDER, A. E. & BROOKES, L. J.

(Eds.) Body Weight Control: The Physiology, Clinical Treatment and Prevention of
Obesity. Edinburgh, London, Melbourne and New York, Churchill Livingstone.

FAUSTO-STERLING, A. (2004) Refashioning Race: DNA and the Politics of Health

Care differences, 15, 1-37.

FEE, E. (1996) Public Health and the State: The United States. IN PORTER, D. (Ed.)
The History of Public Health and the Modern State. Amsterdam and Atlanta GA,
Rodopi.

354
FEE, E. & FOX, D. (1992) AIDS: The Making of a Chronic Disease, Berkelely,
University of California Press.

FEE, M. (2006) Racializing narratives: Obesity, diabetes and the “Aboriginal” thrifty
genotype. Social Science and Medicine, 62, 2988-97.

FLETCHER, I. (2007) Obesity: The Social Construction of a Modern Epidemic. The

University of Edinburgh

FORESIGHT (2007) Tackling Obesities: Future Choices - Project Report. London,

Government Office for Science.

FOREYT, J. & GOODRICK, K. (1995) The ultimate triumph of obesity (letter). The
Lancet, 346, 134-5.

FUJIMURA, J. H. (1988) The Molecular Biological Bandwagon in Cancer Research:

Where Social Worlds Meet. Social Problems, 35, 261-83.

FUJIMURA, J. H. (1992) Crafting Science: Standardized Packages, Boundary

Objects, and "Translation". IN PICKERING, A. (Ed.) Science as Practice and
Culture. London, University of Chicago Press.

FUMENTO, M. (1997) The fat of the land: our health crisis and how overweight
Americans can help themselves, Harmondsworth, Penguin Books.

FUNTOWITZ, S. O. & RAVETZ, J. R. (1993) Science for the Post-Normal Age.

Futures, 25, 739-55

FYE, W. B. (1996) American Cardiology: The History of a Speciality and Its College,
Baltimore and London, The Johns Hopkins University Press.

GAESSER, G. A. (2002) Big Fat Lies: The Truth about Your Weight and Your Health
(updated edition), Carlsbad, CA, Gurze Books.

GARD, M. (2008) The End of the "Obesity Epidemic". Holyrood magazine

Childhood Obesity conference. Balmoral Hotel, Edinburgh Charles Sturt University,
Bathhurst University.

GARD, M. & WRIGHT, J. (2005) The Obesity Epidemic: Science, morality and
ideology, Abingdon, Oxon and New York, Routledge.

GARRETY, K. (1997) Social Worlds, Actor-Networks and Controversy: The Case of

Cholesterol, Dietary Fat and Heart Disease. Social Studies of Science, 21, 727-773.

GARROW, J. S. (1978) Energy balance and obesity in man, Amsterdam, New York,
Oxford Elsevier/North-Holland Biomedical Press.

GARROW, J. S. (1981) Treat Obesity Seriously: A Clinical Manual, Edinburgh,

Churchill Livingstone.

355
GARROW, J. S. (1987) Dietary methods: an overview. IN BENDER, A. E. &
BROOKES, L. J. (Eds.) Body Weight Control: The Physiology, Clinical Treatment
and Prevention of Obesity. Edinburgh, London, Melbourne and New York, Churchill
Livingstone.

GARROW, J. S. (1988) Obesity and related diseases, Edinburgh, Churchill

Livingstone.

GARROW, J. S. & JAMES, W. P. (1993) Human nutrition and dietetics Edinburgh,

Churchill Livingstone.

GIERYN, T. (1983) Boundary-Work and the Demarcation of Science from Non-

Science: Strains and Interests in Professional Ideologies of Scientists. American
Sociological Review, 48, 781-95.

GIERYN, T. (1995) Boundaries of Science. The Handbook of Science and

Technology Studies. Thousand Oaks London and New Delhi, Sage Publications.

GIERYN, T. (1999) Cultural boundaries of science: credibility on the line, Chicago,

London, University of Chicago Press.

GODLEE, F. (1994) The World Health Organization: WHO in crisis. BMJ, 309,
1424-8.

GODLEE, F. (1997) WHO reform and global health (editorial). BMJ, 314, 1359.

GORSKY, M. (2008) The British National Health Service 1948-2008: A Review of

the Historiography. Social History of Medicine, 21, 437-60.

GOVERNMENT OFFICE FOR SCIENCE (2007) Foresight Tackling Obesities:

Future Choices - Project Report. London, Department of Innovation Universities and
Skills.

GRANT, I., FISCHBACHER, C. & WHYTE, B. (2007) Obesity in Scotland: An

Epidemiological Briefing. Edinburgh, Scottish Public Health Observatory.

GRIES, F. A. (1987) Problems with unusual methods of obesity control. IN

BENDER, A. E. & BROOKES, L. J. (Eds.) Body Weight Control: The Physiology,
Clinical Treatment and Prevention of Obesity. Edinburgh, London, Melbourne and
New York.

HACKING, I. (1990) The Taming of Chance, Cambridge, Cambridge University

Press.

HACKING, I. (1999) Making Up People. IN BIAGIOLI, M. (Ed.) The Science

Studies Reader. London and New York, Routledge.

HACKING, I. (2006a) Genetics, biosocial groups and the future of identity.

Daedalus, 135, 81 - 95.

HACKING, I. (2006b) Kinds of people: Moving targets (The Tenth British Academy
Lecture). British Academy http://www.britac.ac.uk/pubs/src/britacad06/index.cfm.

356
HACKING, I. (2006c) Where did the BMI come from? Bodies of Evidence: Fat
Across Disciplines. Centre for Research in the Arts, Social Sciences and Humanities,
University of Cambridge.

HANLEY, T. (1969) The Aetiology of Obesity IN BAIRD, I. M. & HOWARD, A. N.

(Eds.) Obesity Medical and Scientific Aspects: Proceedings of the First Symposium of
the Obesity Association of Great Britain. Edinburgh and London, E&S Livingstone
Ltd.

HANNIGAN, J. A. (1995) Environmental Sociology: A social constructivist

perspective, London and New York, Routledge.

HARDY, A. & TANSEY, E. M. (2006) Medical enterprise and global response. IN

BYNUM, W. F., HARDY, A., JAYCNA, S., LAWRENCE, C. & TANSEY, E. M.
(Eds.) The Western Medical Tradition 1800 to 2000. Cambridge, New York,
Cambridge University Press.

HAUTVAST, J. G. A. J. & DEURENBERG, P. (1987) The risks associated with

obesity; epidemiological studies. IN BENDER, A. E. & BROOKES, L. J. (Eds.) Body
Weight Control: The Physiology, Clinical Treatment and Prevention of Obesity.
Edinburgh, London, Melbourne and New York, Churchill Livingstone.

HEALTH COMMITTEE (2004) Obesity: Third Report of Session 2003-04. London,

House of Commons.

HILGARTNER, S. (2000) Science on stage: expert advice as public drama Stanford,

California, Stanford University Press.

HILGARTNER, S. & NELKIN, D. (1987) Communication Controversies over

Dietary Risks. Science, Technology and Human Values, 12 41-7.

HORROCKS, S. (1995) The Business of Vitamins: Nutrition Science and the Food
Industry in Inter-war Britain. IN KAMMINGA, H. & CUNNINGHAM, A. (Eds.) The
Science and Culture of Nutrition, 1840-1940 Amsterdam, Rodopi.

HORROCKS, S. (1997) Nutrition Science and the Food and Pharmaceutical

Industries in Inter-War Britain. IN SMITH, D. F. (Ed.) Nutrition in Britain: Science,
scientists and politics in the twentieth century. London and New York, Routledge.

HOWARD, A. N. (1969) Dietary Treatment of Obesity IN BAIRD, I. M. &

HOWARD, A. N. (Eds.) Obesity Medical and Scientific Aspects: Proceedings of the
First Symposium of The Obesity Association of Great Britain. Edinburgh and London
E&S Livingstone.

IASO (2011) IASO History. International Association for the Study of Obesity
http://www.iaso.org/about-iaso/history/.

IOTF (2011) Welcome to IOTF. International Association for the Study of Obesity
http://www.iaso.org/iotf/.

JAMES, W. P. (Ed.) (1984) Obesity, London, Philadelphia and Toronto, W B

Saunders Company.

357
JAMES, W. P. (1988) Being overweight: a working definition. IN BENDER, A. E. &
BROOKES, L. J. (Eds.) Body Weight Control. Edinburgh, London, Melbourne and
New York, Churchill Livingstone.

JAMES, W. P. (2006) Marabou 2005: Nutrition and Human Development. Nutrition

Reviews, 64, S1-S11.

JAMES, W. P., FERRO-LUZZI, A., ISAKSSON, B. & SZOSTAK, W. B. (1988)

Healthy Nutrition: Preventing nutrition-related diseases in Europe, Copenhagen,
World Health Organization.

JAMES, W. P., NELSON, M., RALPH, A. & LEATHER, S. (1997) Socioceconomic

determinants of health: The contribution of nutrition to inequalities in health. BMJ,
314, 1545-9.

JASANOFF, S. (1990) The Fifth Branch: Science Advisors as Policymakers,

Cambridge MA and London, Harvard University Press.

JASANOFF, S. (1997) Civilisation and Madness : The Great BSE Scare of 1996.
Public Understanding of Science, 6, 221-32.

JASANOFF, S. (2005) Designs on Nature: Science and Democracy in Europe and the
United States, Princeton and Oxford, Princeton University Press.

JEQUIER, E. (1987) Energy metabolism in the human body. IN BENDER, A. E. &

BROOKES, L. J. (Eds.) Body Weight Control: The Physiology, Clinical Treatment
and Prevention of Obesity. Edinburgh, London, Melbourne and New York, Churchill
Livingstone.

JUNG, J. (1984) Endocrinological Aspects of Obesity. IN JAMES, W. P. (Ed.)

Obesity. London, Philadelphia and Toronto, W.B. Saunders Company.

JUTEL, A. (2001) The emergence of overweight as a disease entity: measuring up

normality. Social Science and Medicine, 63, 2268-76.

JUTEL, A. (2006) Does Size Really Matter? Weight and values in public health.
Perspectives in Biology and Medicine, 44, 283-96.

KALM, L. M. & SEMBA, R. D. (2005) They Starved So That Others Be Better Fed:
Remembering Ancel Keys and the Minnesota Experiment. The Journal of Nutrition,
135, 1347-52.

KAMMINGA, H. (2000) 'Axes to Grind': popularising the science of vitamins, 1920s

and 1930s. IN SMITH, D. F. & PHILLIPS, J. (Eds.) Food, Science, Policy and
Regulation in the Twentieth Century: International and comparative perspectives.
London and New York, Routledge.

KAPRIO, L. A. (1991) Forty Years of WHO in Europe: The development of a

common health policy, Copenhagen, World Health Organization Regional
Publications.

358
KEYS, A. (1980a) Overweight, Obesity, Coronary Heart Disease, and Mortality.
Nutrition Today, 16 - 22.

KEYS, A. (1980b) Seven countries: a multivariate analysis of death and coronary

heart disease, Cambridge MA and London, Harvard University Press and the
Commonwealth Fund.

KEYS, A., FIDANZA, F., KARVONEN, M. J., KIMURA, N. & TAYLOR, H. L.

(1972) Indices of Relative Weight and Obesity. Journal of Chronic Diseases, 25, 329
- 343.

KICKBUSCH, I. (2000) The development of international health policies -

accountability intact? Social Science and Medicine, 51, 979-989.

KICKBUSCH, I. (2003) The Contribution of the World Health Organization to a New

Public Health and Health Promotion. American Journal of Public Health, 93, 383-8.

KLEIN, R. (1995) The New Politics of the NHS (4th edition), Harlow, Essex, Pearson
Education.

KNORR-CETINA, K. (1999) Epistemic Cultures: How the sciences make knowledge,

Cambridge MA and London, Harvard University Press.

KOHLER, R. (2005) Landscapes and labscapes: exploring the lab-field border in

biology, Chicago and London, University of Chicago Press.

KOPELMAN, P. (1984) Clinical Complications of Obesity. IN JAMES, W. P. (Ed.)

Obesity. London, Philadelphia and Toronto, W. B. Saunders Company.

KUCZMARSKI, R. J. & FLEGAL, K. M. (2000) Criteria for definition of overweight

in transition: background and recommendations for the United States. American
Journal of Clinical Nutrition, 72, 1074 - 81.

KUCZMARSKI, R. J., FLEGAL, K. M., CAMPBELL, S. M. & JOHNSON, C. L.

(1994) Increasing Prevalence of Overweight Among US Adults. JAMA, 272, 205-11.

KUHN, T. S. (1970) The Structure of Scientific Revolutions, Chicago and London,

The University of Chicago Press.

KVALE, S. (1996) InterViews: An Introduction to Qualitative Research Interviewing

Thousand Oaks CA, London and New Delhi, Sage Publications.

KWAN, S. (2009a) Framing the Fat Body: Contested Meanings between

Government, Activists and Industry. Sociological Inquiry, 79, 25-50.

KWAN, S. (2009b) Individual versus Corporate Responsibility: Market Choice, the

Food Industry, and the Pervasiveness of Moral Models of Fatness. Food, Culture and
Society, 12, 478-95.

LANDSBERG, L. & YOUNG, J. B. (1984) The Role of the Sympathoadrenal System

in Modulating Energy Expenditure. IN JAMES, W. P. (Ed.) Obesity. London,
Philadelphia and Toronto, W B Saunders Company.

359
LANG, T. & HEASMAN, M. (2004) Food Wars: The Global Battle for Mouths,
Minds and Markets, London, Sterling VA, Earthscan.

LAWRENCE, C. (1997) “Definite and material”: Coronary Thrombosis and

Cardiologists in the 1920s. IN ROSENBERG, C. E. & GOLDEN, J. (Eds.) Framing
Disease: Studies in Cultural History. New Jersey Rutgers University Press.

LAWRENCE, R. G. (2004) Framing Obesity: The Evolution of News Discourse on a

Public Health Issue. Press/Politics, 9, 56-75.

LEATHER, S. (1996) The making of modern malnutrition: An overview of food

poverty in the UK, London, The Caroline Walker Trust

LEE, K. (1998) Historical Dictionary of the World Health Organization, Lanham,

MD & London, The Scarecrow Press Inc.

LEE, K. (2009) The World Health Organization (WHO), London and New York,
Routledge.

LEVENSTEIN, H. (2003) Paradox of Plenty: A Social History of Eating in Modern

America, Berkeley, Los Angeles, London, University of California Press.

LEWIS, J. (1986) What Price Community Medicine: The Philosophy, Practice and
Politics of Public Health since 1919, Brighton, Wheatsheaf.

LEWIS, J. (1999) The concepts of community care and primary care in the UK: the
1960s to the 1990s. Health and Social Care in the Community, 7, 333-341.

LEWIS, J. & RITCHIE, J. (2003) Generalising from Qualitative Research. IN

RITCHIE, J. & LEWIS, J. (Eds.) Qualitative Research Practice: A Guide for Social
Science Students and Researchers. London, Thousand Oaks and New Delhi, SAGE
Publications.

LITSIOS, S. (2008) The third ten years of the World Health Organization, 1968-77,
Geneva, World Health Organization.

LOWY, I. (1990) Variances in Meaning in Discovery Accounts: The Case of Contemporary

Biology. Historical Studies in the Physical and Biological Sciences, 21, 87-121.

LUPTON, D. (1995) The Imperative of Health: Public Health and the Regulated
Body, London, Thousand Oaks CA and New Delhi, Sage.

MACGREGOR, A. (1994) WHO again under fire from external auditor. The Lancet,
343, 1278-9.

MACGREGOR, A. (1996) "Chronic" cash shortage hits WHO. The Lancet.

MACGREGOR, A. (1999) WHO reveals strategy for its "way ahead". The Lancet,
353, 477.

360
MACINTYRE, S. (2003) Before and After the Black Report; Four Fallacies. IN
BERRIDGE, V. & BLUME, S. (Eds.) Poor Health: Social Inequality before and after
the Black Report. London and Portland, OR, Frank Cass.

MACKENZIE, D. (1990) Inventing Accuracy: A Historical Sociology of Nuclear

Missile Guidance, Cambridge, MA and London, MIT Press.

MANSON, J. E., SKERRETT, P. J., GREENLAND, P. & VANITALLLIE, T. B.

(2004) The Escalating Pandemics of Obesity and Sedentary Lifestyle: A Call to
Action for Clinicians. Archives of Internal Medicine, 164, 249-58.

MARMOT, M. & WILKINSON, R. G. (Eds.) (1999) Social Determinants of Health,

Oxford and New York, Oxford University Press.

MASTERMAN, M. (1970) The Nature of a Paradigm. IN LASKATOS, I. &

MUSGRAVE, A. (Eds.) Criticism and the Growth of Knowledge. Cambridge,
Cambridge University Press.

MAYHEW, M. (1988) The 1930s Nutrition Controversy. Journal of Contemporary

History, 23, 445-64.

MCDERMOTT, R. (1998) Ethics, epidemiology and the thrifty gene: biological

determinism as a health hazard, , volume 47 number 9 pp. 1189-95. Social Science
and Medicine, 47, 1189-95.

MCKEE, M. (2003) MONICA: Monograph and Multimedia Sourcebook. Journal of

the Royal Society of Medicine, 96, 613-4.

MCKEOWN, T. (1976) The Modern Rise of Population, London, Edward Arnold.

MCKEOWN, T. (1979) The Role of Medicine: Dream, Mirage or Nemesis? , Oxford,

Basil Blackwell.

MCKEOWN, T. & BROWN, R. G. (1955) Medical evidence relating to English

population changes in the eighteenth century. Population Studies, 9, 119-41.

MCMURRAY, C. & SMITH, R. (2001) Diseases of Globalization: Socioeconomic

Transitions and Health, London and Sterling VA, Earthscan.

MENDELSOHN, J. A. (1998) From Eradication to Equilibrium: How Epidemics

Became Complex after World War 1. IN LAWRENCE, C. & WEISZ, G. (Eds.)
Greater Than The Parts: Holism in Biomedicine. Oxford and New York, Oxford
University Press.

MENNELL, S., MURCOTT, A. & VAN OTTERLOO, A. (1992) The sociology of

food: eating, dieting and culture, London, Sage.

MILLER, D. S. (1969) Energy Metabolism. IN BAIRD, I. M. & HOWARD, A. N.

(Eds.) Obesity Medical and Scientific Aspects: Proceedings of the First Symposium of
the Obesity Association of Great Britain. Edinburgh and London E&S Livingstone
Ltd.

361
MILLWARD, J. (2010) John Waterlow The Guardian. London.

MONAGHAN, L. F. (2005) Discussion Piece: A Critical Take on the Obesity Debate.

Social Theory and Health, 3, 302-14.

MONAGHAN, L. F., HOLLANDS, R. & PRITCHARD, G. (2010) Obesity Epidemic

Entrepreneurs: Types, Practices and Interests. Body and Society, 16, 37-71.

MORRIS, J. N. (1964) Uses of Epidemiology, Edinburgh, E. & S. Livingstone.

MORRIS, J. N., CHAVE, S. P. W., ADAM, C., SIREY, C., EPSTEIN, L. &
SHEEHAN, D. J. (1973) Vigorous Exercise in Leisure-Time and the Incidence of
Coronary Heart-Disease. The Lancet, 301, 333-9.

MORRIS, J. N. & HEADY, J. A. (1953) Mortality in Relation to the Physical

Activity of Work: a preliminary note on the experience in middle age. British Journal
of Industrial Medicine, 10, 245-54.

MORRIS, J. N., KAGAN, A., PATTISON, D. C. & GARDNER, M. J. (1966)

Incidence and Prediction of Ischaemic Heart-Disease in London Busmen. The Lancet,
290, 553-9.

MURCOTT, A. (Ed.) (1998) The nation's diet: the social science of food choice,
London, Longman.

MURRAY, C. & LOPEZ , A. D. (Eds.) (1996) The global burden of disease: a

comprehensive assessment of mortality and disability from diseases, injuries, and risk
factors in 1990 and projected to 2020 (summary booklet), Cambridge MA, Harvard
University Press.

NACNE (1983) A discussion paper on proposals for nutritional guidelines for health
education in Britain. London, The Health Education Council/ National Advisory
Committee on Nutrition Education.

NATIONAL PUBLIC HEALTH LABORATORY OF FINLAND (1981) Community

Control of Cardiovascular Diseases: Evaluation of a comprehensive community
programme for control of cardiovascular disease in North Karelia, Finland 1972 -
1977, Copenhagen, World Health Organization Regional Office for Europe.

NETTLETON, S. (2006) The Sociology of Health and Illness (second edition),

Cambridge and Malden MA, Polity.

NIH PANEL ON THE HEALTH IMPLICATIONS OF OBESITY (1985) The Health

Implications of Obesity. Annals of Internal Medicine, 103, 1073-7.

NOF (2011) About the National Obesity Forum

http://www.nationalobesityforum.org.uk/.

NORUM, K. R. (1987) How a government can influence nutrition behaviour:

experiences from the Norwegian nutrition and food policy. IN BENDER, A. E. &
BROOKES, L. J. (Eds.) Body Weight Control: The Physiology, Clinical Treatment

362
and Prevention of Obesity. Edinburgh, London, Melbourne and New York, Churchill
Livingstone.

NUFFIELD COUNCIL (2007) Public Health: Ethical Issues. London, Nuffield

Council on Bioethics.

O'CONNELL, J. (1993) Metrology: The Creation of Universality by the Circulation

of Particulars. Social Studies of Science, 23, 129 -173.

OAKLEY, A. (1997) Making Medicine Social: The Case of Two Dogs with Bent
Legs. IN PORTER, D. (Ed.) Social Medicine and Medical Sociology in the Twentieth
Century. London, Clio Medica/Wellcome Institute Series in the History of Medicine.

ODDY, D. J., ATKINS, P. J. & AMILIEN, V. (Eds.) (2009) The Rise of Obesity in
Europe: A Twentieth Century Food History, Farnham, Surrey and Burlington VT,
Ashgate.

OFFER, A. (2001) Body Weight and Self-Control in the United States and Britain
since the 1950s. Social History of Medicine, 14, 79-106.

OFFER, A. (2006) The Challenge of Affluence: Self-Control and Well-Being in the

United States and Britain since 1950, Oxford, Oxford University Press.

OHE (1969) Obesity and Disease. London, Office of Health Economics.

OLIVER, J. E. (2006) Fat Politics: The Real Story Behind America's Obesity
Epidemic, Oxford, New York, Oxford University Press.

OMRAM, A. R. (1971) The Epidemiologic Transition: A Theory of the Epidemiology

of Population Change. Milbank Memorial Fund Quarterly, 49, 509-38.

ORBACH, S. (1982) Fat is a Feminist Issue: How to lose weight permanently without
dieting, Feltham, Middlesex, Hamlyn.

PARASCANDOLA, M. (2004) Two approaches to etiology: the debate over smoking

and lung cancer in the 1950s. Endeavour, 28, 81-6.

PAYNE, P. (2010) Witness Seminar: Past and Current nutrition research at LSHTM
Nutrition and History in the Twentieth Century. London School of Hygiene and
Tropical Medicine.

PETERSEN, A. & LUPTON, D. (1996) The New Public Health: Health and Self in
the Age of Risk, London Thousand Oaks and New Delhi, Sage.

PLATT, J. (1981) Evidence and Proof in Documentary Research I: Some Specific

Problems of Documentary Research. Sociological Review, 29, 31-51.

POLLOCK, A. (2010) Reading Friedan: Towards a Feminist Articulation of Heart

Disease. Body and Society, 16, 77-97.

POOL, R. (2001) Fat: Fighting the Obesity Epidemic, Oxford and New York, Oxford
University Press.

363
POPKIN, B. M. (2004) The Nutrition Transition: An Overview of World Patterns of
Change. Nutrition Reviews, 62, S140-S143.

POPKIN, B. M. & DOAK, C. M. (1998) The Obesity Epidemic is a Worldwide

Phenomenon. Nutrition Reviews, 56, 106-114.

POPKIN, B. M. & MENDEZ, M. (2007) The Rapid Shifts in Stages of the Nutrition
Transition: The Global Obesity Epidemic. IN KAWACHI, I. & WAMALA, S. (Eds.)
Globalization and health. Oxford, Oxford University Press.

PORTER, D. (1997a) The Decline of Social Medicine in Britain in the 1960s. IN

PORTER, D. (Ed.) Social Medicine and Medical Sociology in the Twentieth Century.
London Clio Medica/Wellcome Institute Series in the History of Medicine.

PORTER, D. (1997b) Introduction. IN PORTER, D. (Ed.) Social Medicine and

Medical Sociology in the Twentieth Century. London, Clio Medica/Wellcome
Institute Series in the History of Medicine.

PORTER, D. (1999) Health Civilization and the State: A hisotry of public helath from
ancient to modern times, London, New York, Routledge.

PORTER, D. (2003) From Social Structure to Social Behaviour in Britain after the
Second World. IN BERRIDGE, V. & BLUME, S. (Eds.) Poor Health: Social
Inequality before and after the Black Report. London and Portland Or, Frank Cass.

PORTER, D. (2006) How Did Social Medicine Evolve and Where Is It Heading? .
PloS Medicine, 3, 1667-71.

PORTER, T. (1995) Trust in Numbers: The Pursuit of Objectivity in Science and

Public Life, Princeton, NJ, Princeton University Press.

PRINGLE, P. (1994) Washington counts the cost of fat America getting fatter. The
Independent. London.

PUSKA, P., PIETINEN, P., NISSINEN, A., TUOMILEHTO, J., SALONEN, J. T.,
MOISIO, S. & KOSKELA, K. (1987) Principles and potential of practical
community-based health education programmes: experiences from the North Karelia
project, Finland. IN BENDER, A. E. & BROOKES, L. J. (Eds.) Body Weight
Control: The Physiology, Clinical Treatment and Prevention of Obesity. Edinurgh,
London, Melbourne and New York.

QUIRKE, V. & GAUDILLIERE, J.-P. (2008) The Era of Biomedicine: Science,

Medicine and Public Health in Britain and France after the Second World War.
Medical History, 52, 453-70.

RABINBACH, A. (1990) The Human Motor: Energy, Fatigue and the Origns of
Modernity, New York, Basic Books.

RCP (1983) Obesity: A Report of the Royal College of Physicians. London.

364
RCP/BCS (1976) Prevention of Coronary Heart Disease: report of a joint working
party of the Royal College of Physicians of London and the British Cardiac Society.
London Royal College of Physicians.

RICH, E. & EVANS, J. (2005) ‘Fat Ethics’ – The Obesity Discourse and Body
Politics. Social Theory and Health, 3, 341-358.

RISKA, E. (2000) The rise and fall of Type A man. Social Science and Medicine, 51,
1665-74.

RITCHIE, J. (2003) The Applications of Qualitative Methods to Social Research. IN

RITCHIE, J. & LEWIS, J. (Eds.) Qualitative research Practice: A Guide for Social
Science Students and Researchers. London, Thousand Oaks and New Delhi, SAGE
Publications.

ROSE, G. (1987) Obesity: advise the individual or the whole population? IN

BENDER, A. E. & BROOKES, L. J. (Eds.) Body Weight Control: The Physiology,
Clinical Treatment and Prevention of Obesity. Edinburgh, London, Melbourne and
New York, Churchill Livingstone.

ROSE, G., BALL, K., CATFORD, J., JAMES, W. P., LAMBERT, D., MARYON-
DAVIES, A., OLIVER, M., PLAYER, D., ROBBINS, C. & SMITH, A. (1984)
Coronary Heart Disease Prevention Plans for Action: A report based on an
interdisciplinary workshop conference held at Canterbury on 28-30 September 1983.
London, Health Education Council.

ROSE, G. & MARMOT, M. (1981) Social class and CHD. British Heart Journal,
145, 13-19.

ROSE, N. (2007) The Politics of Life Itself: Biomedicine, Power and Subjectivity,
Princeton and Oxford, Princeton University Press.

ROSENBERG, C. E. (1992) Explaining Epidemics, Cambridge, Cambridge

University Press.

ROSENBERG, C. E. (1998) Pathologies as Progress: The Idea of Civilisation as Risk.

Bulletin of the History of Medicine, 72, 714-30.

ROTHSTEIN, W. G. (2003) Public Health and the Risk Factor: A History of an

Uneven Medical Revolution, Rochester, NY, University of Rochester Press.

RUXIN, J. (2000) The United Nations Protein Advisory Group IN SMITH, D. F. &
PHILLIPS, J. (Eds.) Food, Science, Policy and Regulation in the Twentieth Century:
International and comparative perspectives London and New York Routledge.

SAGUY, A. C. & ALMELING, R. (2008) Fat in the Fire? Science in the News
Media, and the "Obesity Epidemic". Sociological Forum, 23, 53-83.

SAGUY, A. C. & RILEY, K. W. (2005) Weighing Both Sides: Morality, Mortality

and Framing Contests over Obesity. Journal of Health Politics, Policy and Law, 30,
869-921.

365
SCHOENFIELDER, L. & WIESER, B. (Eds.) (1983) Shadow on a Tightrope:
Writings by Women on Fat Oppression, Glasgow, Rotunda Press.

SCHWARTZ, H. (1986) Never Satisfied: A Cultural History of Diets, Fantasies and

Fat, New York, London, Toronto, Sydney and Auckland, Anchor Books (Doubleday).

SCOTT, J. (1990) A Matter of Record: Documentary Sources in Social Research,

Cambridge, Polity Press.

SCOTTISH OFFICE (1993) The Scottish Diet. Scotland's health: a challenge to us

all. Report of a Working Party to the Chief Medical Officer for Scotland. Edinburgh,
Scottish Office Home and Health Department.

SHAPIN, S. (1982) History of Science and its Sociological Reconstructions. History

of Science, 20, 157-209.

SHAPIN, S. (2003) How to Eat Like a Gentleman: Dietetics and Ethics in Early
Modern England. IN ROSENBERG, C. E. (Ed.) Right Living: An Anglo-American
Tradition of Self-Help Medicine and Hygiene Baltimore, Maryland, The Johns
Hopkins University Press.

SHELL, E. R. (2003) Fat War: The Inside Story of the Obesity Industry, London,
Atlantic Books.

SHIM, J. K. (2002) Understanding the routinised inclusion of race, socioeconomic

status and sex in epidemiology: the utility of concepts from technoscience studies.
Sociology of Health and Illness, 24, 129-50.

SHOVE, E. (2010) Beyond ABC: climate change policy and theories of social
change. Environment and Planning A, 42, 1273-82.

SIDDIQI, J. (1995) World Health and World Politics: The World Health
Organization and the UN System, London, Hurst & Company.

SIGN (1996) Obesity in Scotland: Integrating Prevention with Weight Management

(A National Clinical Guideline recommended for use in Scotland). Edinburgh,
Scottish Intercollegiate Guidelines Network.

SIGN (2003) Management of obesity in children and young adults (A national clinical
guideline). Edinburgh, Scottish Intercollegiate Guidelines Network.

SILVERSTONE, J. T. (1969) Psychological Factors in Obesity. IN BAIRD, I. M. &

HOWARD, A. N. (Eds.) Obesity Medical and Scientific Aspects: Proceedings of the
First Symposium of The Obesity Association of Great Britain. Edinburgh and London,
E&S Livingstone.

SIMPSON, G. K., ABERNETHY, V. E. & MUNRO, J. F. (1987) Treatment with

drugs. IN BENDER, A. E. & BROOKES, L. J. (Eds.) Body Weight Control: The
Physiology, Clinical Treatment and Prevention of Obesity. Edinburgh, London,
Melbourne and New York

366
SIMS, E. H. (1979) Definitions, Criteria, and Prevalence of Obesity. IN BRAY, G. A.
(Ed.) Obesity in America. Bethesda, Maryland, National Institutes of Health.

SISMONDO, S. (2004) An introduction to Science and Technology Studies, Malden,

MA, Oxford and Carlton, Blackwell Publishing

SMITH, D. F. (1995) The Social Construction of Dietary Standards: The British

Medical Association-Ministry of Health Advisory Committee on Nutrition Report of
1934. IN MAURER, D. & SOBAL, J. (Eds.) Eating Agendas: Food and Nutrition as
Social Problems. New York, Aldine de Gruyter.

SMITH, D. F. (1997) Introduction. IN SMITH, D. F. (Ed.) Nutrition in Britain:

Science, scientists and politics in the twentieth century London and New York,
Routledge.

SMITH, D. F. & BUFTON, M. W. (2004) A Case of Parturiunt Montes, Nascetur

Ridiculus Mus? The BMA Nutrition Committee 1947-50 and the Political
Disengagement of Nutrition Science. Journal of the History of Medicine and Allied
Sciences, 59, 240-72.

SMITH, D. F. & NICOLSON, M. (1995) Nutrition, Education, Ignorance and

Income: A Twentieth Century Debate. IN MAURER, D. & SOBAL, J. (Eds.) Eating
Agendas: Food and Nutrition as Social Problems. New York, Aldine de Gruyter.

SOBAL, J. (1995) The Medicalization and Demedicalization of Obesity. IN SOBAL,

J. & MAURER, D. (Eds.) Eating Agendas: Food and Nutrition as Social Problems.
New York, Aldine de Gruyter.

SOBAL, J. (1999a) A Sociological Analysis of the Stigmatisation of Obesity. IN

GERMOV, J. & WILLIAMS, L. (Eds.) A Sociology of Food and Nutrition:
Introducing the Social Appetite. Melbourne, Oxford University Press.

SOBAL, J. (1999b) Sociological Analysis of the Stigmatization of Obesity. IN

GERMOV, J. & WILLIAMS, L. (Eds.) A Sociology of Food and Nutrition: The
Social Appetite: . New York, Oxford University Press.

SOBAL, J. & STUNKARD, A. J. (1989) Socioeconomic Status and Obesity: A

Review of the Literature. Psychological Bulletin, 105, 260 - 75.

SOCIETY OF ACTUARIES (1959) Build and Blood Pressure Study, Chicago,

Society of Actuaries.

STAPLES, A. L. (2006) The Birth of Development: How the World Bank, Food and
Agriculture Organization and World Health Organization Changed the World, 1945-
65, Kent, Ohio, The Kent State University Press.

STEARNS, P. (2002) Fat History: Bodies and Beauty in the Modern West, New York
and London, New York University Press.

STEWARD, F. (2010) If 'societal challenge' has re-entered the discourse of

innovation policy why is STIS still not seen as relevant for practice as science,

367
economic or psychology? ISSTI Retreat: Working at the margins?/working across the
divides? Alnwick.

STEWART, J. (2008) The Political Economy of the British National Health Service,
1945-75: Opportunities and Constraints? Medical History, 52.

STUNKARD, A. J. (1987) Behavioural therapy for obesity. IN BENDER, A. E. &

BROOKES, L. J. (Eds.) Body Weight Control: The Physiology, Clinical Treatment
and Prevention of Obesity Edinburgh, London, Melbourne and New York, Churchill
Livingstone.

STURM, R. (2002) The Health Risks of Obesity: Worse than Smoking, Drinking or
Poverty. Rand Corporation

STURM, R. (2007) Increases in Morbid Obesity in the USA: 2000-2005. Public

Health, 121.

STURM, R. (2008) Stemming the global obesity epidemic: What can we learn from
data about social and economic trends? Public Health, 122, 739-46.

STURM, R. & COHEN, D. (2009) Zoning for Health? The Year-Old Ban on New
Fast-Food Restaurants in South LA: The Ordinance Isn't a Promising Approach to
Attacking Obesity. Health Affairs, 28, w1088 - w1097.

STURM, R., RINGEL, J. S., LAKDAWALLA, D. N., BHATTACHARYA, J.,

GOLDMAN, D. P., HURD, M. D., JOYCE, G. F., PANIS, C. & ANDREYEVA, T.
(2007) Obesity and Disability: The Shape of Things to Come. RAND Corporation.

SUSSER, M. (1985a) Epidemiology in the United States after World War II: The
evolution of technique. Epidemiologic Reviews, 7, 147-177.

SUSSER, M. (1985b) Epidemiology in the United States after World War II: The
evolution of technique. Epidemiologic Reviews, 7, 147-77.

SWINBURN, B., EGGER, G. & RAZA, F. (1999) Dissecting Obesogenic

Environments: The Development and Application of a Framework for Identifying and
Prioritising Environmental Interventions for Obesity. Preventive Medicine, 29, 563 –
70.

SZRETER, S. (2007) Health and Wealth: Studies in History and Policy, Rochester,
NY, University of Rochester Press.

THROSBY, K. & ROBERTS, C. (2010) Getting bigger: children’s bodies, genes and
environments. The Sociological Review, 58, 73-92.

TIMMERMANN, C. (2006) A matter of degree: The normalisation of hypertension

c.1940-2000. IN ERNST, W. (Ed.) Histories of the Normal and the Abnormal.
Abingdon, Oxon and New York, Routledge.

TOWNSEND, P. & DAVIDSON, N. (1982) Inequalities in Health: The Black Report,

Harmondsworth, Penguin.

368
TSICHLIA, G. & JOHNSTONE, A. (Eds.) (2010) Fat Matters: From Sociology to
Science, Keswick, Cumbria, M&K Publishing.

TUNSTALL-PEDOE, H. (Ed.) (2003) MONICA Monograph and Multimedia

Sourcebook: World's largest study of heart disease, stroke, risk factors, and
population trends, 1979 - 2002, Geneva, World Health Organization.

TURNER, P. (1969) Drugs in the Treatment of Obesity. IN BAIRD, I. M. &

HOWARD, A. N. (Eds.) Obesity Medical and Scientific Aspects: Proceedings of the
First Symposium of The Obesity Association of Great Britain. Edinburgh and London,
E&S Livingstone.

VAGUE, J. (1968) Obesity, Isleworth, Boehringer Ingelheim.

VANDENBROUCKE, J. P. (2004) The history of confounding IN MORABIA, A.

(Ed.) A History of Epidemiologic Methods and Concepts. Basel Boston Berlin,
Birkhause Verlag.

VANITALLLIE, T. B. & WOTEKI, C. E. (1987) Who gets fat? IN BENDER, A. E.

& BROOKES, L. J. (Eds.) Body Weight Control. Edinburgh, London, Melbourne and
New York, Churchill Livingstone.

WALT, G. (1994) Health policy: an introduction to process and power, London, Zed
Books.

WANLESS, D., APPLEBY, J., HARRISON, A. & PATEL, D. (2007) Our Future
Health Secured? A Review of NHS funding and performance. London, The King's
Fund.

WARDE, A. (1997) Consumption, Food and Taste: Cultural Antinomies and

Commodity Culture, London, Thousand Oaks, New Delhi, Sage.

WEBSTER, C. (2002) The National Health Service: A Political History, Oxford, New
York, Oxford University Press.

WEBSTER, C. (2003) Investigating Inequalities in Health before Black. IN

BERRIDGE, V. & BLUME, S. (Eds.) Poor Health:Social Inequality before and after
the Black Report. London and Portland, OR, Frank Cass.

WEINBERG, A. M. (1972) Science and trans-science. Minerva, 10, 209-22.

WEINDLING, P. (1993) The Immunological Tradition. IN BRYNUM, W. F. &

PORTER, R. (Eds.) Companion Encyclopedia of the History of Medicine. London and
New York, Routledge.

WEINDLING, P. (1995) The Role of International Organizations in Setting

Nutritional Standards in the 1920s and 1930s. IN KAMMINGA, H. &
CUNNINGHAM, A. (Eds.) The Science and Culture of Nutrition, 1840-1940
Amsterdam, Rodopi.

WEST, R. (1994) Obesity. London, Office of Health Economics.

369
WHELAN, H. & SILVERSTONE, T. (Eds.) (1974) Obesity: a bibliography, London,
Information Retrieval Limited.

WHO (1950) Session of the Expert Committee on the Unification of Pharmacopoeias.

Geneva, World Health Organisation.

WHO (1957) Atheroschlerosis and Ischaemic Heart Disease. Geneva, World Health
Organization.

WHO (1958) The First Ten Years of the World Health Organization, Geneva, WHO.

WHO (1962) Arterial hypertension and ischaemic heart disease: preventitive aspects.
Geneva World Health Organization.

WHO (1968) The Second Ten Years of the World Health Organization, Geneva,
WHO.

WHO (1982) Prevention of Coronary Heart Disease. Geneva, World Health

Organization.

WHO (1985) Energy and Protein Requirements. Geneva, World Health Organization

WHO (1986) Community prevention and control of cardiovascular diseases. Geneva,

World Health Organization.

WHO (1990) Diet, nutrition, and the prevention of chronic diseases. Geneva, World
Health Organization.

WHO (1995) Physical Status: The Use and Interpretation of Anthropometry. Geneva,
World Health Organization.

WHO (2000) Obesity: Preventing and Managing the Global Epidemic. Geneva,
World Health Organization.

WHO (2003) Diet, Nutrition and the Prevention of Chronic Disease. Geneva, World
Health Organization.

WHO (2004) Global Strategy on Diet, Physical Activity and Health Geneva, World
Health Organization.

WHO (2006) WHO European Ministerial Conference on Counteracting Obesity.

Copenhagen, WHO Regional Office for Europe.

WHO (2010a) Forty-fourth report of the WHO Expert Committee on Specifications

for Pharmaceutical Preparations Geneva, World Health Organization.

WHO (2010b) WHO/Programmes and projects/Medicines/Publications/Technical

Report Series 957, 2010. World Health Organization
http://www.who.int/medicines/publications/44threport/en/index.html.

370
WHO REGIONAL OFFICE FOR EUROPE (2006) WHO European Ministerial
Conference on Counteracting Obesity: Conference Report. Copenhagen, WHO
Regional Office for Europe.

WILKINSON, R. G. (1996) Unhealthy Societies: The Afflictions of Inequality,

London and New York, Routledge.

WOLF, N. (1990) The Beauty Myth: How Images of Beauty are Used Against
Women, London, Vintage.

WORBOYS, M. (1988) The discovery of colonial malnutrition between the wars

IN ARNOLD, D. (Ed.) Imperial medicine and indigenous societies. Manchester and
New York, Manchester University Press.

WORBOYS, M. (2000) Spreading Germs: Disease Theories and Medical Practice in

Britain 1865-1900, Cambridge, Cambridge University Press.

WYNNE, B. (1993) Misunderstood misunderstandings: social identities and the

public uptake of science. Public Understanding of Science, 1, 281-304.

WYNNE, B. (1996) May the Sheep Safely Graze? A Reflexive View of the Expert-
Lay Knowledge Divide IN LASH, S., SZERSZYNSKI, S. & WYNNE, B. (Eds.)
Risk, Environment and Modernity. London, Sage.

YUDKIN, J. (1969) The Treatment of Obesity. IN BAIRD, I. M. & HOWARD, A. N.

(Eds.) Obesity Medical and Scientific Aspects: Proceedings of the First Symposium of
The Obesity Association of Great Britain. Edinburgh and London E&S Livingstone.

371