Hoddinott et al.

BMC Public Health 2010, 10:800

http://www.biomedcentral.com/1471-2458/10/800

CORRESPONDENCE Open Access

Group interventions to improve health outcomes:

a framework for their design and delivery
Pat Hoddinott1*, Karen Allan2†, Alison Avenell1†, Jane Britten3†

Abstract
Background: Delivering an intervention to a group of patients to improve health outcomes is increasingly popular
in public health and primary care, yet “group” is an umbrella term which encompasses a complex range of aims,
theories, implementation processes and evaluation methods. We propose a framework for the design and process
evaluation of health improvement interventions occurring in a group setting, which will assist practitioners,
researchers and policy makers.
Methods: We reviewed the wider literature on health improvement interventions delivered to patient groups and
identified a gap in the literature for designing, evaluating and reporting these interventions. We drew on our
experiences conducting systematic reviews, intervention, mixed method and ethnographic studies of groups for
breastfeeding and weight management. A framework for health improvement group design and delivery evolved
through an iterative process of primary research, reference to the literature and research team discussion.
Results: Although there is an extensive literature on group processes in education, work, politics and psychological
therapies, far less is known about groups where the aim is health improvement. Theories of behaviour change
which are validated for individual use are often assumed to be generalisable to group settings, without being
rigorously tested. Health improvement or behaviour change interventions delivered in a group setting are complex
adaptive social processes with interactions between the group leader, participants, and the wider community and
environment. Ecological models of health improvement, which embrace the complex relationship between
behaviour, systems and the environment may be more relevant than an individual approach to behaviour change.
Conclusion: The evidence for effectiveness and cost-effectiveness of group compared with one-to-one
interventions for many areas of health improvement in public health and primary care is weak or unknown. Our
proposed framework is the first step towards advocating a more systematic approach to designing, evaluating and
reporting interventions in group settings, which is necessary to improve this currently weak evidence base. This
framework will enable policy makers and practitioners to be better informed about what works, how it works and
in which contexts when aiming to improve health in a group setting.

Background cardiac rehabilitation groups [6], the expert patient pro-

Groups are an alternative to individual encounters for gramme led by trained patients with personal experience
health improvement, social support and changing beha- of a condition [7] and virtual internet self-help groups
viour, for example: smoking cessation [1]; weight loss [2]; [8]. However the evidence for health improvement inter-
parentcraft [3]; and self care for chronic conditions like ventions delivered in group settings is dispersed through
diabetes [4] and osteoarthritis [5]. Such groups are evol- several systematic reviews of specific lifestyle behaviours,
ving rapidly in response to cultural, epidemiological and most of which focus on individual behaviour change
environmental change, for example recent increases in interventions and theory. In our experience researching
breastfeeding support and weight management groups
* Correspondence: p.hoddinott@abdn.ac.uk since 2000 and 2002 respectively, we were surprised by
† Contributed equally the lack of guidelines for designing, evaluating or report-
1
Health Services Research Unit, University of Aberdeen, Health Sciences ing health improvement interventions in group settings.
Building, Foresterhill, Aberdeen AB25 2ZD, UK
Full list of author information is available at the end of the article

© 2010 Hoddinott et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.
Hoddinott et al. BMC Public Health 2010, 10:800 Page 2 of 9
http://www.biomedcentral.com/1471-2458/10/800

We identified this as a gap to which our research could Development of the framework
usefully contribute. The framework has developed through our research into
“Groups” feature in social, political, cultural, educa- breastfeeding support and weight management groups,
tional and work contexts, besides health, and have a which included systematic reviews, mixed method group
variety of meanings, underlying theories and definitions. interventions and ethnographic studies [2,16]. Studies
These are reviewed by Rupert Brown [9], a social psy- included a randomised controlled trial of a policy to
chologist, who proposes the following: provide breastfeeding groups across Scotland [17] with a
mixed method evaluation of implementation processes
“A group exists when two or more people define [18]; a controlled intervention study of individual and
themselves as members of it and when its existence is group peer support for breastfeeding [19,20] and an eth-
recognised by at least one other person or group of nographic study of participant and provider experiences
people who do not so define themselves” of weight management groups [21]. All of these studies
conducted in-depth qualitative interviews with partici-
The National Institute for Health and Clinical Excel- pants and providers and group observations.
lence (NICE) behaviour change guidelines distinguish In addition, our studies and this paper were informed
between interventions at the individual, community and by a literature review of health improvement interven-
population level [10]. Groups are included in the broad tions for patient groups. We searched Medline,
category of community interventions, defined as social CINAHL, Embase, The Cochrane Library, National
or family groups linked by networks, geographical loca- Institute for Health and Clinical Excellence (NICE), Psy-
tion or another common factor. cinfo and the International Bibliography of the Social
In this paper, we use a broad definition of health Sciences using group$, class$, club$, workshop$ and
improvement to include health promotion, disease pre- program$ as title words for papers published between
vention (primary and secondary), public health, commu- 1980 until 2009. This was a pragmatic search strategy to
nity development approaches and social support. We limit the number of papers identified and was necessary
refer to “interventions in group settings” to define what because of the widespread use of the word “group” in
happens to people within the group, the context in differing contexts in research papers. We identified sys-
which it happens and the relationship between the two. tematic reviews in relevant areas, for example smoking
Our definition of the setting refers to an observable and alcohol cessation; chronic diseases like diabetes,
health improvement group and the related activities and cancer support and heart disease and searched their
processes that occur. These include the place where the reference lists. We also searched our personal reference
group meets and we consider the wider aspects of the archives and hand searched references in key papers.
space occupied by people and things, including the We excluded educational groups for teaching students
attached meanings and relationships [11]. It also or staff; organisational and management literature on
includes the wider geographical, cultural, media, political work groups and teams, and treatment groups in mental
and organisational environment of health improvement health, where there is an extensive literature.
group settings. The framework evolved through an iterative process of
Our aim is to provide a framework for the design and mixed method data analysis, reference to the literature,
process evaluation of health improvement interventions reflection and research team discussion over a period of
in group settings, guided by the literature on designing 6 years. It builds on reviews of group processes [9],
complex interventions [12]. Detailed discussion of statis- small group work in education and work settings [22]
tical aspects are not covered but are important as they and self-help or support groups [23].
need to take account of interactions between both
group participants and group leaders [13]. Neither are What is the evidence for health improvement
specific methods of data collection for the process eva- interventions delivered in group settings?
luation of group interventions discussed. Instead our The evidence for health improvement interventions in
framework poses a series of questions which are impor- group settings is varied, reflecting their heterogeneity
tant to consider when designing and evaluating a group and complexity and some argue that policy for group
intervention. This is central to ecological theories of interventions to encourage self care has raced ahead of
behaviour, as groups are complex systems with multiple the evidence [24]. Systematic reviews often focus on an
interacting variables, at several levels which require a individual disease, the type of treatment, or behaviour
mixed method approach [14]. We recommend using a change theory and they often inappropriately combine
toolkit approach [15] to choosing the most appropriate the results of interventions delivered individually and in
methods (quantitative or qualitative) to answer each group settings in a meta-analysis [13]. For example, a
question, informed by existing evidence. Cochrane review of additional support for breastfeeding
Hoddinott et al. BMC Public Health 2010, 10:800 Page 3 of 9
http://www.biomedcentral.com/1471-2458/10/800

differentiates between whether the support is lay or pro- group self care interventions, there was insufficient
fessional, but not between interventions delivered to evidence to support the claim that group interventions
individuals or groups [25]. In a systematic review of are cheaper and there are often trade-offs between the
smoking cessation groups, theories of behaviour change numbers of patients treated and the quantity of inter-
validated for individual delivery, for example stages of vention each individual receives [24]. Smoking cessation
change or cognitive behavioural theory, have been groups are no more cost effective than intensive indivi-
assumed to be transferable to delivery in a group setting dual counseling [1] and breastfeeding groups provided
[1]. This systematic review of smoking cessation inter- as part of routine care would have similar costs to indi-
ventions [1] is one of few which has specifically analysed vidual health visitor home visits to group participants
outcomes for interventions in group settings and com- [17]. Group interventions for weight management in
pares them with self-help materials, individual counsel- obesity are potentially more resource saving in terms of
ling, nicotine replacement or no intervention. It total health professional-hours involved per participant
concludes that group programmes are more effective [2]. Few self care group interventions for arthritis have
than no intervention or self help interventions, but there measured service use and findings are conflicting
is insufficient evidence to evaluate whether groups are [29,30].
more effective than individual counselling. Another sys-
tematic review comparing group versus individually Why do we need a framework?
delivered interventions for weight management in adult In the literature discussed above, it is clear that the jury
obesity found that group interventions report signifi- is still out when it comes to deciding whether group or
cantly more weight loss at 1 year follow-up compared individual interventions are more effective and cost-
with the same intervention individually delivered [2]. A effective at improving health outcomes. Group processes
systematic review of group education programmes for and interactions have received less attention than dyadic
adults with type 2 diabetes compared with routine treat- or individual behaviour change mechanisms and little is
ment, waiting list control or no intervention, found known about which components of groups contribute to
group programmes were effective at lowering glycated effectiveness. As a result there are many unanswered
haemoglobin, fasting blood glucose, blood pressure, questions for practitioners and policy makers who aim
weight and medication use [4]. For cardiac rehabilita- to establish patient groups to improve health outcomes.
tion, interventions delivered individually at home were The NICE behaviour change guidelines recommend
as effective as those delivered in a centre to a group [6]. being as specific as possible about the content of the
The aims of a group intervention, the underlying intervention, spelling out what is done, to whom and in
behavioural change theories and group processes can what social and economic context [10]. A challenge with
determine who attends and both the group and the indi- groups is to unpick the extent to which outcomes are
vidual health outcomes. A one size fits all approach sel- determined by leadership style, personality, participants’
dom meets everyone’s needs and it has been argued that characteristics or more complex interactions. Having a
for self care in chronic disease, groups should only be framework for design and process evaluation is one step
considered for simple standardised messages, where towards producing this evidence and we describe this
peer support is beneficial or preferred and where a below.
group will save money [24]. Importantly, some groups
may increase health inequalities by attracting more edu- The place, setting and context of group interventions
cated and higher income participants [17,26,27]. It has The first and crucial stage in designing and evaluating a
been suggested that support groups are more likely to group intervention is to consider how aspects of the set-
be sought for diseases viewed as stigmatizing, like AIDS, ting will impact on all aspects of group processes, com-
alcoholism, breast and prostate cancer rather than less position and outcomes (Table 1). In a trial which
stigmatizing, yet important, diseases like heart disease randomised primary care organisations to deliver a pol-
[28]. There are inconsistencies in self care group defini- icy to provide breastfeeding support groups, the breast-
tions, for example in a survey of American self care feeding outcomes were explained by the characteristics
groups 60% of those identified had professional facilita- of the primary care organisation, including health
tors and the clinical potential for social support in com- inequalities and deprivation, the amount of organisa-
bination with professional guidance compared to peer tional change taking place and multi-disciplinary team-
only support is largely unknown [28]. work [18]. The breastfeeding outcomes could not be
The potential for cost savings with group compared to explained by the amount of intervention delivered or
individual health improvement interventions can appear the number of people attending the groups.
attractive; however the health economic evidence is Evaluating how different environmental settings influ-
mixed and often weak. In a review of individual or ence behaviour has been relatively neglected in research
Hoddinott et al. BMC Public Health 2010, 10:800 Page 4 of 9
http://www.biomedcentral.com/1471-2458/10/800

Table 1 The setting group leaders may behave differently outside health
A) Within the meeting place and venue - the activity setting service settings [18]. Group resources can vary between
B) The immediate surrounding environment including parallel purposely designed venues with state of the art props to
activity settings
C) The wider geographical area
“make do” multi-purpose clinically cluttered spaces.
For each of the above consider: Consideration should also be given to sensory percep-
○ Socio-demographic characteristics, facilities, human and tions including comfort, temperature, noise, smell and
structural resources (e.g. funds, time, people, physical objects)
○ Access issues
visual appearance [20].
○ Sensory aspects (comfort, temperature, noise, smell, visual)
○ Meanings attributed to the setting including opportunities and Designing a group intervention
threats
D) The wider policy, political, media, legal, cultural and
Table 2 focuses on the design of the intervention, the
environmental context and how it interacts with the group theory underlying the choice of intervention, the target
population and choosing the relevant behavioural out-
come to measure.
compared to the predominant approach of measuring What is the intervention and what quantity will be
the psychological variables of individuals receiving a delivered?
behaviour change intervention. Theorising interventions The interactions between group members may form the
as events in complex systems is particularly relevant to intervention as with peer support or self-care [7]. The
health improvement groups, where the context, the intervention might be something which is delivered by
change in relationships and resources over time are the group leader, like a particular diet or exercise pro-
important [31]. The importance of the behavioural set- gramme and there may be a range of intended and
ting [32] or the activity setting [33] as a unit of analysis actual interaction between group participants from
is a cross disciplinary feature of psychological, social and minimal, mainly non-verbal communication to high
biological ecology [32,34]. Behavioural or activity set- level engagement. The intervention may intend to have
tings refer to time and space bounded patterns of beha- an outreach beyond the group setting, for example sex
viour, and the concept originates from research which education delivered by peer educators [36]. It is particu-
characterised all behaviour settings in a small American larly important in trials of interventions in group set-
community [32]. This demonstrated that behavioural tings to consider the statistical aspects of whether the
settings have effects on individuals that extend beyond group attributes and processes detailed in Tables 1, 2
the specific behaviour demanded by the activity. and 3 are acting as mediators or moderators between
It can be helpful to consider the context in which the the intervention and the health outcomes. The group
group intervention takes place at macro, meso and size, frequency, duration and lifetime will all influence
micro levels, as proposed in ecological models of health group composition and processes. Decisions about the
improvement [35]. The macro level considers how wider components of the intervention are linked to the under-
policy, economic and socio-cultural factors interact with lying theory informing it and the outcomes of interest
the group, for example media scares or the cultural which are discussed below. There are detailed guidelines
values of participants. The meso level includes the inter- about how to report the attributes of complex interven-
relations between the group, the setting and the sur- tions which are relevant to groups [37,38]. However,
rounding environment, for example holding the group reducing complex social processes like groups into stan-
in an affluent or disadvantaged area can influence per- dardised, reproducible intervention components has its
ceptions and who attends a group [18]. Importantly, the critics [39] and interventions which are responsive to
impact of parallel activity settings [33] which may target local contexts and change over time may be more
the same population should be considered. For example, appropriate.
the impact that general parent craft groups open to all How does someone become a group member?
women can have on participation in breastfeeding only It is important to assess how entry rituals and gate-
groups [18]. Similarly, consideration should be given to keeper assumptions are influencing information dissemi-
assessing the impact of activities displaced by the group nation and recruitment to a group. Some groups have
intervention [31]. What providers and participants stop elaborate entry rituals, for example general practitioners
doing when a health improvement group starts is likely may be asked to complete forms prior to registering for
to be crucial to outcomes, yet this is seldom systemati- a cardiac rehabilitation group whereas other groups
cally described or evaluated. The micro level is the encourage open access. Convenience, minimising bar-
interaction between group participants and the space riers and ensuring that the participant benefits outweigh
where the group meets. Community settings like church the risks are crucial to a successful group and will be
halls or family centres convey different meanings for highly context dependent [20]. In a study comparing
people compared to health settings and health service commercial and health service groups for weight loss,
Hoddinott et al. BMC Public Health 2010, 10:800 Page 5 of 9
http://www.biomedcentral.com/1471-2458/10/800

Table 2 Designing interventions in group settings Table 3 What happens within a group
What is the intervention and what quantity will be delivered? System maintenance:
○ The group itself as the intervention ○ Who organises and leads the group? Is he/she internal or
○ The group leader delivers the intervention external to the group? How is he/she appointed or elected?
○ The group as a vehicle for delivering the intervention to a wider ○ The leader’s role in initiating, planning, setting up, organising
population and running the group
○ The group size, frequency, duration and lifetime ○ One or several group leaders/co-leaders? Similar or
complementary leader attributes; continuity or rotation of
How does someone become a group member? leadership; fixed or flexible?
○ Are there gatekeepers and how do they operate ○ Is group content: flexible; repetitive; different over time; leader or
○ Self or professional referral, with or without criteria participant determined?
○ Advertising: general or targeted ○ Is there group member feedback? Formal or informal? How does
○ Access to attend meetings: open (anyone can drop in and out of feedback change group processes?
attending meetings); closed (membership registration on Leader attributes:
attendance, in advance or for a fixed period) ○ Socio-demographic characteristics, professional qualifications,
○ Access during a group meeting: open (drop in and out); closed training, personal experience of the behaviour or problem,
(fixed start and finish) interpersonal communication skills
○ Barriers, facilitators and entry rituals ○ To what extent is the leader able to attend to both the group
○ Incentives and costs (financial and non financial) - joining, task and the socio-emotional aspects of the group?
recurring, optional, refundable ○ What is the leadership style: directive/nondirective; proactive/
What social and behaviour theories inform the intervention? reactive; led (hierarchical)/facilitated (co-operative)/present
(autonomous)?
○ Education: factual, tacit or experiential knowledge
○ How flexible is the leader and how does the leader change over
○ Support: for a specific behaviour, attitude or belief time?
○ Cognitive approaches: to change thinking about a behaviour ○ What are the benefits/rewards and costs/burdens of being a
○ Performing a behaviour or activity leader and how are they manifest?
○ Rewarding a behaviour or group attendance Relationship between the leader and group:
○ Competition between groups or group members ○ How does the leader have legitimacy in the eyes of the group: e.
How are the group influencing attitudes, beliefs and behaviours? g., expert knowledge; skills; competence; personal attributes;
For example: personal experience; conforming to group norms; acting fairly;
group identity; geographical residence?
○ Social comparison theory
○ What techniques does the leader use: education; persuasion;
○ Social support theory providing a practical task or service; advocacy; advising: supporting;
○ Social learning theory empowering; counselling; listening; providing vision; inspiration or
○ Social impact theory motivation; selling?
What are the outcomes? ○ What do the group initiators, leaders and group members view
as the purpose (aims and objectives) of the group? How similar or
○ Initiate or sustain a desired behaviour different are their perspectives?
○ Reduce, stop or prevent a relapse of an undesirable behaviour Attributes of the group participants
○ Substitute a desirable for an undesirable behaviour ○ To what extent are the group task/goal or socio-emotionally
○ Change how an existing behaviour is enacted orientated?
○ Change attitudes or beliefs which might predict or mediate a ○ To what extent are there shared goals?
behaviour e.g. self-efficacy ○ What does it mean to be a group member/non-member in
What is the target population for intervention delivery and terms of personal and social identity?
outcome measurement? ○ Do participants categorise themselves; adopt specific group roles
○ Who is targeted? People with specific behaviours, socio- or a hierarchical status?
demographic characteristics or diseases; from particular geographic ○ What is the level of anonymity or public performance within the
areas or organisations group? High with each individual speaking in turn or low as in a
○ Whose outcomes will be measured? Individual group attenders, crowd where anonymity can be maintained?
pooled group outcomes, wider population Group relationships
○ To what extent are socio-emotional interactions positive or
negative?
some participants valued the flexibility and autonomy ○ How do intra-, inter-group and non-group member relationships
offered by groups where you could drop in just to be change over time?
weighed and leave with minimal group interaction ○ Do group attitudes, beliefs and behaviours become more or less
whereas others preferred leader facilitated discussion extreme over time?
[21]. Relationships between professional gatekeepers and ○ How similar or different are the attitudes and behaviours of
group members?
“the group” can both facilitate or hinder attendance, for
○ What are the group norms, how are the limits of acceptable
example infrequent midwife participation led to limited behaviour defined and is difference tolerated?
attendance by pregnant women in a trial of breastfeed- ○ Do the group norms encourage or inhibit goal attainment and/
ing support groups [18]. or positive socio-emotional interactions?
Hoddinott et al. BMC Public Health 2010, 10:800 Page 6 of 9
http://www.biomedcentral.com/1471-2458/10/800

Table 3: What happens within a group (Continued) [43]. In groups, strong and weak social connections may
○ How cohesive is the group?
differ in their effects and either reinforce existing atti-
○ How much communication between group members occurs?
tudes and behaviours or mediate change. Influence often
Minimal, mostly non-verbal to in depth engagement? extends beyond the group to the family, local community
○ Is communication within the group channelled through the or population and this may be intentional, for example in
leader, within subgroups or free with multiple conversations? peer education interventions, or unintentional. This
poses challenges for outcome measurement and contami-
How will the group influence people? nation, as the cumulative outcome is the sum of the indi-
This is a contentious area as no single theory can cap- vidual outcomes and the collateral positive or negative
ture the complexity of intra and inter group behaviour. health outcomes of others [44]. There is a continuing
In Table 2, we suggest separating the social and beha- debate about the relative importance of social support
vioural theories that inform the group intervention from network size and the strength of connections. On-line
the theory of how the group itself is likely to influence support groups are becoming increasingly popular and
the attitudes, beliefs and behaviours of participants. provide large networks of relatively superficial support
Individual behaviour change theories and techniques [28]. Further comparisons between actual and virtual
have been reviewed [40], but not specifically for delivery health improvement and support groups are warranted.
in group settings and it cannot be assumed that they are Bandura’s Social Learning Theory with its emphasis
generalisable, as individuals can behave differently when on learning through observation and modelling beha-
in a group [9]. Group composition may have a causal viour [45] is particularly relevant to behaviours involving
effect on group outcomes; the group may be the social action or performing, like breastfeeding or parenting
context which allows other inter-personal psychological skills [20]. Minorities and majorities both influence
phenomena to unfold or group composition may be a group processes and Social Impact Theory proposes a
consequence of other external factors [41]. negatively accelerating continuum of influence based on
How intra and inter group processes influence peo- observations that the first social stimulus has the great-
ple’s attitudes, beliefs and behaviours are reviewed else- est effect, the second less effect and the third less still
where [9], although not specifically in the context of [46].
health improvement. In Table 2 we provide some exam- What are the health outcomes and target populations?
ples of theories which are particularly relevant to health Outcomes from interventions delivered in a group are
improvement group interactions. Festinger’s Social usually measured at the individual level [1-3], and most
Comparison Theory [42] proposes that conformity studies do not consider interactions between patients in
within a group is dependent on three main motivations: the same group which may lead to correlation of out-
dependence on others for information to self-evaluate; comes [13]. Cluster randomized controlled trials where
achieving group goals and the need for approval and a randomization and outcome measurement have
desire not to seem different. Festinger hypothesises that occurred at the group level are rare. Occasionally wider
group participants will try to improve their performance population level outcomes are measured, for example
and will differ in whether they compare upwards or the multifaceted STD/HIV Intervention Project (SHIP)
downwards, for similarities or dissimilarities. They will where peer educators delivered individual and group
select different attributes to compare, which may not interventions resulting in dramatic and sustained
always be the expected ones [20]. Downward compari- improvements in sexually transmitted infections at a
son where a person wants to know how dissimilar he/ population level [36]. As with cluster randomized trials,
she is from the most undesirable person is more com- individuals in a group cannot be considered to be inde-
mon in the expert patient programme and serves to pro- pendent of each other and variation in the outcome is
tect a threatened self-esteem [27]. Festinger hypothesises likely to be smaller for participants treated in the same
that high status members are motivated to try and group than for participants treated in different groups.
improve the performance of others less capable or they Similarly, if the same group leader delivers the interven-
might perform below their capability so that they would tion to different groups of participants, the outcome
not appear too different from the rest. Criticisms are may differ less than the outcome for participants treated
that social comparison theory focuses on intra-group by different group leaders. This design clustering is an
comparisons, whereas temporal comparisons with the additional consideration when deciding whether to ran-
past or the future, or comparisons with other groups or domise individuals or clusters to a group intervention.
non group participants may be of equal importance [9].
Social support theory proposes that information is dis- What happens within a group?
seminated more effectively between networks of people Table 3 proposes a series of questions examining the
with strong social ties and this confers health benefits micro-level of interactions between the group leader,
Hoddinott et al. BMC Public Health 2010, 10:800 Page 7 of 9
http://www.biomedcentral.com/1471-2458/10/800

new and existing group members and the group setting. exercise groups have standardised qualifications for
It is closely related to the ecological, social and beha- leaders, whereas in some primary care groups for weight
vioural theories described above. Group processes and loss, practice nurses are self-taught and learn on the job
their interactions with the group setting can determine [21].
group survival and the intended outcomes, but they The NICE behaviour change guidelines [10] recom-
have been under-researched in the context of health mend that community interventions should: improve
improvement groups and the relative importance of self-efficacy; develop and maintain supportive social net-
individual components is unknown. Group Environment works; promote resilience; build skills; promote partici-
Scales (GES) have been developed to systematically mea- pation in voluntary activities; promote involvement in
sure the norms, values and psychosocial characteristics planning and delivery of services and have access to the
of social environments [47]. The underlying theory is financial and material resources needed to facilitate
that environments like people have unique personalities, behaviour change. Of fundamental importance is having
climates or atmospheres which are important determi- a clearly stated purpose with aims and objectives that
nants of behaviour. GES have three basic dimensions: are agreed by everyone. However, even with clear aims
relationship dimensions which assess the extent to and objectives, differing interpretations will arise. For
which individuals are involved, for example cohesion, example, in a trial of breastfeeding support groups, lea-
support; personal development dimensions, for example ders interpreted the word “support” in different ways.
autonomy, goal attainment; system maintenance or sys- Despite a clear protocol stating that support should be
tem change dimensions for example order, organisation, woman-centred and based on informed choice, some
clarity, control. GES have been developed in the field of interpreted “support” as promoting and encouraging
mental health treatment, education and work and little exclusive breastfeeding as you might support a football
is known about their generalisablity to health improve- team [16]. Some experts classify groups into either task
ment groups. They have been used to compare self-help orientated groups, where individuals or the group have
groups [23] and these three dimensions seem salient in specified goals, or socio-emotional groups where feelings
a wide variety of social settings, although the character- and inter-personal relationships are paramount [9].
istics of each dimension may vary [34,48]. The dimen- However, our group observations suggest that this bin-
sions we use in Table 3 are adapted from GES to ary classification may not apply to the complexity of
provide a better fit with our data. In particular the per- health improvement interventions, which usually com-
sonal development dimension seems more relevant to bine task and socio-emotional objectives [16,21].
educational or treatment groups and we have reframed A study of cancer support groups illustrates how an
this as attributes of the leader and the group partici- analysis of group processes can provide important evi-
pants. As groups are complex systems there is some dence and highlights the difficulties with pre-selecting
overlap and interaction between dimensions. group composition [50]. Education and discussion
We have used the term “group leader” to embrace a groups combining distressed and undistressed cancer
range of observed styles which others have defined patients were compared with a group of distressed
[9,22]. Variation in the delivery skills of group leaders patients. The heterogeneous group increased the social
can theoretically determine the success of the interven- comparison opportunities for distressed patients and
tion and it is important to consider the statistical effects benefited them. However, undistressed women with
of such clustering in the design [13]. Education, support breast cancer who had high levels of social support
and behavioural interventions in a group setting are showed a slight deterioration in physical functioning
characterised by different communication styles. For when attending the heterogeneous group, which
example, a group leader may be an advocate of a raised ethical concerns. The authors suggest training
method, philosophy, activity, seller/deliverer of a “pro- undistressed participants to maximise their benefit for
duct” or a facilitator of person centred approaches, for distressed participants but to minimise negative conse-
example empowerment, counselling or support [49]. quences for themselves.
How leaders with personal experience of a condition, Individuals tend to conform to the attitudes and beha-
for example in commercial weight management groups viours of the majority within a group, and cliques can
[21] and the expert patient programme [7], influence develop. For example, some breastfeeding groups can be
group outcomes is largely unknown. Some groups are dominated by mothers breastfeeding older infants,
multifaceted with communication interspersed with which may be off-putting to new mothers joining, but
activities like relaxation, physical activity, weighing or an over-all “feel-good atmosphere” can override personal
physiological monitoring. Health service training in differences [20]. Studies suggest that conformity and
group leadership is variable depending on the type of uniformity in groups increases with group size and over
group and the nature of the intervention. For example, time and groups tend to exhibit more extreme attitudes,
Hoddinott et al. BMC Public Health 2010, 10:800 Page 8 of 9
http://www.biomedcentral.com/1471-2458/10/800

beliefs and behaviours than the individual group 6. Taylor RS, Dalal H, Jolly K, Moxham T, Zawada A: Home-based versus
centre-based cardiac rehabilitation. Cochrane Database of Systematic
participants [9]. Reviews 2010, , 1: CD007130.
7. Department of Health: The expert patient: a new approach to chronic
Conclusion disease management for the 21st century. London 2001.
8. Sandaunet AG: The challenge of fitting in: non-participation and
The evidence for effectiveness and cost-effectiveness of withdrawal from an online self-help group for breast cancer patients.
group compared with one-to-one interventions for health Sociology of Health & Illness 2008, 30(1):131-144.
improvement is weak or unknown. Health improvement 9. Brown R: Group Processes. Oxford: Blackwell;, 2 2000.
10. National Institute for Health and Clinical Excellence: Behavioural change at
or behaviour change interventions delivered in a group population, community and individual levels. 2007 [http://www.nice.org.
setting are complex adaptive social processes with inter- uk/PH6].
actions between the group leader, participants, and the 11. Cummins S, Curtis S, Diez-Roux AV, Macintyre S: Understanding and
representing ‘place’ in health research: A relational approach. Social
wider community and environment. As with any complex Science & Medicine 2007, 65(9):1825-1838.
intervention, there is a need to understand which aspects 12. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M:
of group processes work, for whom, how and in what cir- Developing and evaluating complex interventions: new guidance. 2008
[http://www.mrc.ac.uk/Utilities/Documentrecord/index.htm?d=MRC004871].
cumstances. Ecological models of behavioural change 13. Roberts C, Roberts SA: Design and analysis of clinical trials with clustering
may be more appropriate than applying theories of indi- effects due to treatment. Clinical Trials 2005, 2(2):152-162.
vidual behavioural change in a group setting. This frame- 14. Maton KI: Moving beyond the individual level of analysis in mutual help
group research: an ecological paradigm. The Journal of Applied Behavioral
work is the first step towards advocating a more Science 1993, 29(2):272-286.
systematic approach to designing, evaluating and report- 15. Ritchie J, Lewis J: Qualitative Research Practice London: Sage; 2003.
ing interventions in group settings, which is necessary to 16. Hoddinott P: Group based approaches to supporting breastfeeding in
primary care. PhD thesis, University of Aberdeen; 2008.
improve the currently weak evidence base. 17. Hoddinott P, Britten J, Prescott GJ, Tappin DM, Ludbrook A, Godden DJ:
Effectiveness of a policy to provide breastfeeding groups (BIG) for
pregnant and breastfeeding mothers in primary care: a cluster
Acknowledgements randomised controlled trial. British Medical Journal 2009, 338:a3026.
The Chief Scientist’s Office of the Scottish Government Health Directorate 18. Hoddinott P, Britten J, Pill R: Why do interventions work in some places
provided funding for PH through a Primary Care Research Career Award and and not others: A breastfeeding support group trial. Social Science &
AA through a Career Scientist Award. The Health Services Research Unit, Medicine 2010, 70(5):769-778.
University of Aberdeen, is funded by the Chief Scientist Office of the 19. Hoddinott P, Lee AJ, Pill R: Effectiveness of a breastfeeding peer coaching
Scottish Government Health Directorates. The views expressed are those of intervention in rural Scotland. Birth 2006, 33(1):27-36.
the authors. 20. Hoddinott P, Chalmers M, Pill R: One-to-one or group based peer support
for breastfeeding? Women’s perceptions of a breastfeeding peer
Author details coaching intervention. Birth 2006, 33(2):139-146.
1
Health Services Research Unit, University of Aberdeen, Health Sciences 21. Allan K, Hoddinott P, Avenell A: A qualitative study comparing
Building, Foresterhill, Aberdeen AB25 2ZD, UK. 2NHS Education for Scotland, commercial and health service weight loss groups, classes and clubs.
Thistle House, 5th Floor, 91 Haymarket Terrace, Edinburgh, EH12 5HD, UK. Journal of Human Nutrition and Dietetics 2010 [http://onlinelibrary.wiley.
3
Centre for Rural Health, University of Aberdeen, The Centre for Health com/doi/10.1111/j.1365-277X.2010.01110.x/full].
Science, Old Perth Road, Inverness, IV2 3JH, UK. 22. Elwyn G, Greenhalgh T, Macfarlane F: Groups. A Guide to Small Group
Work in Healthcare, Management, Education and Research. Abingdon,
Authors’ contributions Oxon: Radcliffe Medical Press Ltd; 2001.
PH had the idea for the paper and drafted it. All authors contributed to the 23. Kurtz LF: Self-help and Support Groups: A Handbook for Practitioners.
analysis of data in the ethnographic studies of interventions in group California, USA: Sage;, First 1997.
settings which informed this paper and to writing the paper. All authors 24. Greaves CJ, Campbell JL: Supporting self-care in general practice. British
read and approved the final manuscript. PH is guarantor. Journal of General Practice 2007, 57(543):814-821.
25. Britton C, McCormick FM, Renfrew MJ, Wade A, King SE: Support for
Competing interests breastfeeding mothers. Cochrane Database of Systematic Reviews 2007, , 1:
The authors declare that they have no competing interests. CD001141.
26. Shaw J, Baker M: “Expert patient"–dream or nightmare? British Medical
Received: 10 September 2010 Accepted: 31 December 2010 Journal 2004, 328(7442):723-724.
Published: 31 December 2010 27. Rogers A, Gately C, Kennedy A, Sanders C: Are some more equal than
others? Social comparison in self-management skills training for long-
References term conditions. Chronic Illness 2009, 5(4):305-317.
1. Stead LF, Lancaster T: Group behaviour therapy programmes for smoking 28. Davison KP, Pennebaker JW, Dickerson SS: Who talks? The social
cessation. Cochrane Database of Systematic Reviews 2005, , 2: CD001007. psychology of illness support groups. American Psychologist 2000,
2. Paul-Ebhohimhen V, Avenell A: A systematic review of the effectiveness 55(2):205-217.
of group versus individual treatments for adult obesity. Obesity Facts 29. Bodenheimer T, Lorig K, Holman H, Grumbach K: Patient self-management
2009, 2:17-24. of chronic disease in primary care. Journal of the American Medical
3. Gagnon AJ, Sandall J: Individual or group antenatal education for Association 2002, 288(19):2469-2475.
childbirth or parenthood or both. Cochrane Database of Systematic 30. Patel A, Buszewicz M, Beecham J, Griffin M, Rait G, Nazareth I, Atkinson A,
Reviews 2007, , 3: CD002869. Barlow J, Haines A: Economic evaluation of arthritis self management in
4. Deakin TA, McShane CE, Cade JE, Williams R: Group based training for self- primary care. British Medical Journal 2009, 339:b3532.
management strategies in people with type 2 diabetes mellitus. 31. Hawe P, Shiell A, Riley T: Theorising interventions as events in systems.
Cochrane Database of Systematic Reviews 2005, , 2: CD003417. American Journal of Community Psychology 2009, 43:267-276.
5. Buszewicz M, Rait G, Griffin M, Nazareth I, Patel A, Atkinson A, Barlow J, 32. Barker RG: Ecological Psychology: Concepts and Methods for Studying
Haines A: Self management of arthritis in primary care: randomised the Environment in Human Behavior. Stanford, California, USA: Stanford
controlled trial. British Medical Journal 2006, 333(7574):879. University Press; 1968.
Hoddinott et al. BMC Public Health 2010, 10:800 Page 9 of 9
http://www.biomedcentral.com/1471-2458/10/800

33. O’Donnell CR, Tharp RG, Wilson K: Activity settings as the unit of analysis:
a theoretical basis for community intervention and development.
American Journal of Community Psychology 1993, 21(4):501-520.
34. Moos RH, Insel PM: Issues in Social Ecology. California, USA: National Press
Books;, First 1974.
35. McLeroy KR, Bibeau D, Steckler A, Glanz K: An ecological perspective on
health promotion programs. Health Education & Behaviour 1988,
15(4):351-377.
36. Jana S, Basu I, Rotheram-Borus M, Newman P: The Sonagachi Project: A
sustainable community intervention program. Aids Education and
Prevention 2004, 16(5):405-414.
37. Equator Network: Enhancing the Quality and Transparency Of Health
Research:[http://www.equator-network.org/home/].
38. Armstrong R, Waters E, Moore L, Riggs E, Cuervo LG, Lumbiganon P,
Hawe P: Improving the reporting of public health intervention research:
advancing TREND and CONSORT. Journal of Public Health 2008,
30(1):103-109.
39. Hawe P, Shiell A, Riley T: Complex interventions: how “out of control” can
a randomised controlled trial be? British Medical Journal 2004,
328(7455):1561-1563.
40. Abraham C, Michie S: A taxonomy of behavior change techniques used
in interventions. Health Psychology 2008, 27(3):379-387.
41. Levine JM, Moreland RL: Progress in small group research. Annual Review
of Psychology 1990, 41(1):585-634.
42. Festinger F: A theory of social comparison. Human Relations 1954,
1:117-140.
43. Berkman LF: Assessing the physical health effects of social networks and
social support. Annual Review of Public Health 1984, 5:413-432.
44. Christakis NA: Social networks and collateral health effects. British Medical
Journal 2004, 329(7459):184-185.
45. Bandura A: Social Learning Theory. New York: General Learning Press;, First
1977.
46. Latané B: The psychology of social impact. American Psychologist 1981,
36:343-356.
47. Moos RH, Group Environment Scale Manual:[http://www.mindgarden.com/
products/gescs.htm].
48. Insel PM, Moos RH: Health and the Social Environment. Massachusetts,
USA: D.C. Heath and Company;, First 1974.
49. Kiiti N, Nielson E: Facilitator or advocate: what’s the difference? In The Art
of Facilitating Participation. Edited by: White SA. London: Sage; 1999:52-67.
50. Carmack Taylor CL, Kulik J, Badr H, Smith M, Basen-Engquist K, Penedo F,
Gritz ER: A social comparison theory analysis of group composition and
efficacy of cancer support group programs. Social Science & Medicine
2007/7 65(2):262-273.

Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1471-2458/10/800/prepub

doi:10.1186/1471-2458-10-800
Cite this article as: Hoddinott et al.: Group interventions to improve
health outcomes: a framework for their design and delivery. BMC Public
Health 2010 10:800.

Submit your next manuscript to BioMed Central

and take full advantage of:

• Convenient online submission

• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution

Submit your manuscript at

www.biomedcentral.com/submit
BioMed Central publishes under the Creative Commons Attribution License (CCAL). Under the CCAL, authors
retain copyright to the article but users are allowed to download, reprint, distribute and /or copy articles in
BioMed Central journals, as long as the original work is properly cited.