Sexual Behaviour in Context: A Global Perspective: Sexual and Reproductive Health 2
Sexual Behaviour in Context: A Global Perspective: Sexual and Reproductive Health 2
Sexual Behaviour in Context: A Global Perspective: Sexual and Reproductive Health 2
Journal paper
Sexual behaviour in context:
a global perspective*
Kaye Wellings, Martine Collumbien, Emma Slaymaker, Susheela Singh, Zoé Hodges, Dhaval Patel,
Nathalie Bajos
Research aimed at investigating sexual behaviour and assessing interventions to improve sexual
health has increased in recent decades. The resulting data, despite regional differences in quantity
and quality, provide a historically unique opportunity to describe patterns of sexual behaviour and
their implications for attempts to protect sexual health at the beginning of the 21st century. In this
Journal paper
paper we present original analyses of sexual behaviour data from 59 countries for which they were
available. The data show substantial diversity in sexual behaviour by region and sex. No universal
trend towards earlier sexual intercourse has occurred, but the shift towards later marriage in most
countries has led to an increase in premarital sex, the prevalence of which is generally higher in
developed countries than in developing countries, and is higher in men than in women. Monogamy
is the dominant pattern everywhere, but having had two or more sexual partners in the past year is
more common in men than in women, and reported rates are higher in industrialised than in non-
industrialised countries. Condom use has increased in prevalence almost everywhere, but rates
remain low in many developing countries.
The huge regional variation indicates mainly social and economic determinants of sexual behaviour,
which have implications for intervention. Although individual behaviour change is central to improving
sexual health, efforts are also needed to address the broader determinants of sexual behaviour,
particularly those that relate to the social context. The evidence from behavioural interventions is
that no general approach to sexual-health promotion will work everywhere and no single-component
intervention will work anywhere. Comprehensive behavioural interventions are needed that take
account of the social context in mounting individual-level programmes, attempt to modify social
norms to support uptake and maintenance of behaviour change, and tackle the structural factors
that contribute to risky sexual behaviour.
* This is a pre-print copy of a paper published in the journal The Lancet : Kaye Wellings, Martine Collumbien,
Emma Slaymaker, Susheela Singh, Zoé Hodges, Dhaval Patel, Nathalie Bajos. Sexual behaviour in context: a global
perspective. The Lancet Sexual and Reproductive Health Series, October 2006.
so the evidence base is partial. Even so, a reasonable global impression health-care systems and public-health strategies have also wrought
can be gained from the large number of developing countries with changes;13 access to family-planning services has increased, and few
comparable data (those for which a Demographic and Health Survey areas have been unaffected by efforts to prevent HIV transmission. We
(DHS) has been undertaken8), and from other countries with comparable describe current trends and patterns in important variables of sexual
national surveys. Data are accumulating too, from assessments of the behaviour, and their implications for sexual-health status and the
effect of interventions to improve sexual health. design of sexual-health interventions. We present our key messages
in panel 1.
The resulting evidence provides a historically unique opportunity to take
stock of sexual behaviour and efforts to protect sexual health at the
beginning of the 21st century. Sexual behaviour changes in response to Cross-national comparisons
both secular and non-secular social forces. Recent decades have seen We compare data for entire countries (panel 2 and table 1) and in so
large socioeconomic changes in poverty, education, and employment. doing we pay less attention to the substantial variability within countries.
Demographic changes have taken place, in the age structure of Our choice of indicators shows a concern with sexual health and human
populations, in the timing of marriage, and in the scale of mobility rights, but is also determined by availability of comparable data. Age
and migration between and within countries, including seasonal at sexual initiation is of public health interest, since early initiation is
labour, rural-to-urban movement, and social disruption due to war more likely to be non-consensual and to be subsequently regretted,31–33
and political instabilities. Attitudes to sexual behaviour have altered in less likely to be protected against unplanned pregnancy and infection,
many countries.9,10 Worldwide communications, including the internet,11 and associated with larger lifetime numbers of sexual partners.34–36
have had a bearing on social norms, transporting sexual images from Risk behaviours for sexually transmitted infections, such as multiple
more liberal to more conservative societies, especially those in which partnerships and sex between men are included as are prevalence
advances in information technology have been rapid.12 estimates of transactional sex, since clients of sex workers are important
bridging groups in the transmission of sexually transmitted infections
Advances in contraception have increasingly freed sexual expression
and HIV to wider sexual networks. We include condom use as a marker
from its reproductive consequences. Policy and legislation that govern
of safer sex.
2
Panel 2: Methods
Search strategy
We reviewed medical, public health, and social science research by searching Embase, MEDLINE, Popline, PubMed, and Web of Science electronic
databases for articles published between 1996 and 2006 using the key words “sexual behaviour” and “sexual behaviour change”. We also contacted
experts in sexual behaviour studies, we hand-searched AIDS, The Lancet, and Social Science and Medicine, we found references cited by key papers,
and accessed web-based information through Internet sites that are commonly used in the context of sexual health.
We searched the Cochrane Database of Systematic Reviews and the Database of Abstracts of Reviews of Effects for systematic reviews published
in the past 10 years (from 1996 onwards). Terms used to identify reviews in the Cochrane Database of Systematic Reviews, and those that were
successful in identifying relevant reviews were used subsequently to search the Database of Abstracts of Reviews of Effects: (MESH terms) “sexual
behaviour”, “contraception”; (free text terms) “sex”*, “sex* near behavio*”, “sex* near risk*”, “sex* near safe*”, “behavio*”, “barrier*”. We included
reviews that reported sexual behaviour outcomes or outcomes that are closely associated with sexual behaviour, for example, sexual partnerships
(eg, reductions in partner numbers, monogamy, exclusivity), sexual practices (eg, abstinence, condom use, non-penetrative sex), or changes in
behavioural determinants associated with either of these (eg, empowerment, self-efficacy, resistance to peer pressure, alternative income generation
for commercial sex workers). The results of the reviews were summarised by population group, with emphasis given to reviews done most recently
and those providing useful information regarding the essential components of successful interventions.
3
Almost everywhere, sexual activity begins for most men and women
Existence Data Published Unable to
in the later teenage years (ages 15–19 years), but regional and sex of data obtained estimates include*
variations between men and women are substantial (table 2). For established obtained
women, median age at first intercourse is low in regions in which early Countries with ≥ 1 survey of:
marriage is the norm (for example, in south Asia, central, west, and east All women 76 48 7 20
Ever-married women 4 5† - -
Africa), and high in Latin America and in some countries of the middle
Men 65 44 4 17
east and southeast Asia. For men, age at first intercourse is, in general,
Countries with > survey of:
not linked to age at marriage. In most African and Asian countries, men
Women 39 27 1 8
start to have sex later than do women. Gender differences are most
Men 25 18 1 4
pronounced in the less industrialised countries. For three countries we were unable to obtain the most recent or most comprehensive data
and so used alternative data.
None of these data suggest a universal trend towards sex at a young
*Report or dataset unobtainable or published estimates unsuitable or both. †Data for all
age. The trends are complex and vary greatly with male or female sex. women in one country available but unobtainable; estimates from an ever-married sample
The proportion of women who report early intercourse (before age 15 presented in their place.
years) decreased significantly or remained stable between the late Table 1: Availability of survey data
1970s and the late 1990s in four of five countries for which comparisons
can be made (figure 1). For men, the proportion either remained
stable or increased during that period. The increase was significant in countries, in which the time is about the same for both (about 5 years).
several Latin American and African countries. In societies in which first Trends in the average time spent between first sexual experience and
intercourse still occurs mainly within marriage, the trend towards later settling with a marital or cohabiting partner result from changes in both
marriage has been accompanied by a trend towards later sex in young the proportion who have first sexual intercourse before marriage, and
women,37 and is particularly a feature of countries in Africa and south the changing interval between these two events. Although the number
Asia (figure 2). In some industrialised countries, sexual activity before of countries in which first sexual intercourse and marriage coincide is
age 15 years has become more common in recent decades (though diminishing39 a clear increase in the intervening period is taking place
the prevalence is lower than in other regions and the increase is not in only very few countries, again, notably in the more industrialised
generally significant). (figure 2).
The trend towards later marriage in many countries has also led to an Most people are married and married people have the most sex (figure 5).
increase in the prevalence of premarital sexual intercourse.38 Again, large Sexual activity in single people tends to be sporadic and, in most regions,
regional and sex differences in prevalence exist (figure 3). As shown in many fewer than half unmarried non-virgins report having had sex in
figure 4, on average, the time between first sexual intercourse and living the past month. Single men and women in many African countries are
with a partner is longer for men than for women (typically 3–6 years fairly sexually inactive, with two-thirds reporting recent sexual activity
compared with 0–2 years, respectively) apart from in industrialised compared with three-quarters of those in industrialised countries.
1995–99
40
30
20
10
0 USA ‡
Rwanda†
Italy ‡
Egypt ¶ †
Cameroon‡
Togo †
Mali †
Guinea‡
Namibia†
South Africa ‡
Mozambique†
Uzbekistan†
Kazakhstan†
Brazil †
Columbia ‡
Haiti †
Philippines†
Vietnam ¶ ‡
Nepal‡
Switzerland‡
France ‡
Australia ‡
Dominican Republic *
Nicaragua *
Norway*
Madagascar †
Ethiopia‡
Morocco §*
Gabon *
Chad *
CAR *
Ghana *
Burkina Faso *
Senegal*
Benin *
Cote d’Ivoire *
Nigeria *
Niger *
Zimbabwe*
Tanzania *
Kenya *
Malawi *
Comoros*
Uganda*
Zambia *
Peru *
Bolivia *
Guatemala*
Indonesia*
India ¶ *
Bangladesh ¶*
Britain *
Armenia *
Turkey § *
60 Men
50
Sex before 15 years (%)
40
30
20
10
0
Norway †
Britain †
Chad †
CAR †
Ghana †
Nigeria †
Togo †
Côte d’Ivoire†
Ethiopia †
Zimbabwe†
Uganda †
Kazakhstan †
Peru †
Dominican Republic†
Indonesia †
Philippines†
Bangladesh†
France ‡
Italy ‡
Switzerland‡
Australia‡
USA ‡
Zambia†
Armenia †
Cameroon*
Gabon *
Senegal *
Burkina Faso *
Mali *
Niger *
Guinea*
Benin *
Rwanda *
Tanzania *
Malawi *
Mozambique*
Namibia*
Kenya*
Turkey *
Bolivia*
Haiti*
Brazil*
Nepal *
North and West Africa East and South Africa Countries in Latin America Asia Industrialised countries
central Africa transition and Caribbean
Figure 1: Percentage who had sex before age 15 years: comparison between two time points
Difference between groups: *p<0·05, †p≥0·5, ‡Test not possible. §Age at first marriage used as a proxy for age at first sex. ¶Ever-married sample and age at first marriage used as a proxy for
age at first sex.
4
Men Women
Median age at first % had SI before age Median age at first Median age at first % had SI before age Median age at first
SI (years, IQR) 15 years (95% CI) marriage (years, IQR) SI (years, IQR) 15 years (95% CI) marriage (years, IQR)
Central Africa
Central African Republic 17·5 (15·5–18·5) 23 (19·0–27·0) 15·5 (14·5–17·5) 30·3 (27·6–33·1) 17 (15·0–20·0)
Cameroon 18·5 (16·5–20·5) 8·8 (6·6–10·9) 25 (21·0–30·0) 15·5 (14·5–17·5) 30·7 (27·8–33·5) 17 (15·0–20·0)
Chad 18·5 (16·5–20·5) 5·2 (2·6–7·8) 22·01 (19·0–25·0) 15·5 (14·5–17·5) 39 (36·3–41·8) 15·01 (14·0–18·0)
Egypt .. .. .. .. .. 19 (16·0–22·0)
Gabon 17·5 (14·5–18·5) .. 24 (20·0–29·0) 15·5 (14·5–17·5) .. 19 (16·0–25·0)
Morocco 22 (17·0–30·0)
West Africa
Benin 17·5 (15·5–19·5) 16·9 (13·6–20·2) 24 (20·0–28·0) 17·5 (15·5–18·5) 15·1 (13·2–17·0) 18·01 (16·0–21·0)
Burkina Faso 20·5 (17·5–22·5) 2·5 (1·2–3·8) 25 (21·0–28·0) 17·5 (15·5–18·5) 9·8 (8·3–11·4) 17·01 (16·0–19·0)
Côte d’Ivoire 18·5 (16·5–20·5) 7·7 (3·2–12·1) 26 (22·0–32·0) 15·5 (14·5–17·5) 29·6 (25·0–34·3) 18 (15·0–22·0)
Ghana 19·5 (17·5–22·5) 4·3 (2·5–6·1) 24 (21·0–28·0) 17·5 (16·5–19·5) 10·6 (8·4–12·9) 19 (16·0–22·0)
Guinea 17·5 (16·5–20·5) .. 26 (21·0–30·0) 15·5 (14·5–17·5) .. 16 (14·0–18·0)
Mali 19·5 (17·5–21·5) 5·5 (3·4–7·6) 25(22·0–29·0) 15·5 (14·5–17·5) 28·5 (26·4–30·6) 16 (15·0–18·0)
Niger 20·5 (17·5–23·5) 4 (1·8–6·2) 22·01 (19·0–26·0) 15·5 (14·5–16·5) 46·6 (43·0–50·2) 15·01 (14·0–16·0)
Nigeria 20·5 (18·5–24·5) 5·5 (2·3–8·8) 27 (21·0–30·0) 15·5 (14·5–18·5) 39·4 (35·4–43·4) 15·01 (14·0–20·0)
Senegal 20·5 (18·5–24·5) .. 29·01 (25·0–32·0) 17·5 (15·5–19·5) 16·9 (14·4–19·3) 17·01 (15·0–21·0)
Togo 18·5 (16·5–20·5) 7·7 (4·9–10·4) 24 (21·0–29·0) 17·5 (15·5–18·5) 18·7 (16·4–21·0) 18·01 (16·0–21·0)
East and South Africa
Comoros 18·5 (15·5–23·5) 18 (14·1–21·9) 19·01 (16·0–25·0)
Ethiopia 18·5 (15·5–20·5) 4·2 (0·7–7·8) 24 (20·0–27·0) 15·5 (14·5–17·5) 31·5 (27·2–35·7) 15·01 (14·0–18·0)
Kenya 16·5 (14·5–18·5) 26 (23·0–29·0) 25 (22·0–28·0) 17·5 (15·5–19·5) 19 (17·3–20·7) 19 (17·0–23·0)
Madagascar .. .. .. 16·5 (15·5–19·5) 20·9 (18·3–23·5) 19 (16·0–22·0)
Malawi 17·5 (15·5–20·5) 13·5 (10·2–16·8) 23 (20·0–25·0) 16·5 (15·5–18·5) 20·6 (18·0–23·2) 17·01 (15·0–19·0)
Mozambique 18·5 (16·5–19·5) 13·5 (9·5–17·5) 21·01 (18·0–25·0) 15·5 (14·5–17·5) 32·8 (29·3–36·3) 17 (15·0–20·0)
Namibia 18·5 (16·5–20·5) 6·9 (3·9–9·9) 29·01 (25·0–35·0) 18·5 (16·5–20·5) 5·4 (3·8–7·0) 26·01 (20·0– )
Rwanda 18·5 (17·5–21·5) 3·2 (0·8–5·7) 25 (22·0–28·0) 20·5 (18·5–22·5) 3 (2·0–4·0) 20·01 (18·0–23·0)
South Africa .. .. .. 17·5 (16·5–19·5) .. 24·01 (19·0– )
Tanzania 18·5 (16·5–20·5) 9·3 (7·6–11·0) 24 (21·0–28·0) 16·5 (15·5–18·5) 15·6 (14·0–17·1) 18·01 (16·0–21·0)
Uganda 17·5 (15·5–20·5) 15·5 (12·2–18·8) 22 (20·0–25·0) 16·5 (14·5–18·5) 27·8 (25·3–30·2) 17·01 (15·0–20·0)
Zambia 16·5 (15·5–20·5) 22·6 (18·9–26·4) 23 (21·0–27·0) 16·5 (15·5–18·5) 22·9 (20·7–25·2) 18 (16·0–21·0)
Zimbabwe 19·5 (17·5–21·5) 6 (4·0–8·1) 24 (22·0–28·0) 18·5 (16·5–20·5) 10·1 (8·4–11·7) 19 (17·0–22·0)
Countries in transition
Armenia 20·5 (18·5–22·5) 1·6 (0–3·4) 24 (22·0–26·0) 20·5 (18·5–22·5) 0·2 (0–0·4) 20·01 (18·0–22·0)
Kazakhstan 20·5 (17·5–22·5) 3·6 (0·1–7·0) 23·01 (21·0–27·0) 20·5 (18·5–22·5) 0·4 (0·0–0·8) 21 (19·0–23·0)
Turkey 18·5 (16·5–20·5) .. 24 (22·0–26·0) .. .. 20 (17·0–23·0)
Uzbekistan .. .. .. 20·5 (18·5–21·5) .. 20·01 (18·0–21·0)
Latin America and Caribbean
Bolivia 17·5 (15·5–18·5) 19 (16·0–21·9) 23 (20·0–27·0) 18·5 (16·5–20·5) 9·2 (8·1–10·3) 20 (17·0–24·0)
Brazil 16·5 (14·5–18·5) 29·6 (23·7–35·4) 24 (21·0–28·0) 18·5 (16·5–21·5) 8·8 (7·5–10·2) 20·01 (18·0–25·0)
Chile* 17 .. .. 20·2 .. ..
Colombia .. .. .. 18·5 (16·5–22·5) 9·7 (8·9–10·5) 21·01 (18·0–27·0)
Dominican Republic 16·5 (14·5–18·5) 29·9 (25·5–34·4) 23 (20·0–27·0) 18·5 (16·5–22·5) 14·1 (12·8–15·4) 19·01 (16·0–23·0)
Guatemala .. .. .. 18·5 (15·5–20·5) .. 19 (16·0–22·0)
Haiti 17·5 (15·5–19·5) 16·5 (12·2–20·8) 26 (22·0–32·0) 18·5 (16·5–20·5) 10·9 (8·7–13·1) 20 (17·0–24·0)
Nicaragua .. .. .. 17·5 (15·5–20·5) 17 (15·6–18·5) 18 (15·0–21·0)
Peru 16·5 (15·5–18·5) 20·9 (15·8–26·0) 23 (21·0–30·0) 18·5 (16·5–22·5) 8·3 (7·6–9·0) 21 (18·0–26·0)
Asia
Bangladesh 22·5 (18·5–26·5) .. 23 (20·0–26·0) .. .. 14 (13·0–16·0)
India .. .. .. .. .. 16 (15·0–19·0)
Indonesia 24·5 (21·5–27·5) .. 24·01 (21·0–27·0) 18·5 (16·5–21·5) .. 18·01 (16·0–21·0)
Nepal 18·5 (16·5–21·5) 9·1 (5·6–12·5) 20 (17·0–22·0) 16·5 (15·5–18·5) 21·9 (19·6–24·1) 16·01 (15·0–18·0)
Philippines 20·5 (18·5–23·5) 2·6 (1·3–3·8) 24 (21·0–29·0) 21·5 (18·5–26·5) 3·3 (2·6–4·0) 22 (18·0–26·0)
Vietnam .. .. .. .. .. 20 (18·0–22·0)
Industrialised countries
Australia 17·5 (16·5–18·5) 13 (9·5–17·6) 29 (24·0– ) 17·5 (16·5–19·5) 6·5 (3·9–10·8) 24 (22·0–30·0)
Britain 16·5 (15·5–18·5) 12·5 (10·3–14·7) 24 (21·0–28·0) 17·5 (16·5–18·5) 6·9 (5·4–8·4) 22 (20·0–25·0)
France 17·5 (16·5–19·5) 7·2 (4·5–11·1) .. 18·5 (17·5–19·5) 5·9 (3·5–9·9) ..
Italy 17·5 (17·5–18·5) 4 (2·0–11·8) .. 18·5 (18·5–21·5) 2·8 (1·3–6·2) ..
Norway 18·5 (16·5–20·5) 5·5 (3·3–8·9) .. 17·5 (16·5–19·5) 9·6 (7·0–13·1) ..
Switzerland 18·5 (16·5–20·5) 6·8 (4·1–10·8) .. 18·5 (17·5–20·5) 3·4 (1·8–6·2) ..
USA 17·3 (15·7–18·8) 17·8 (17·0–18·8) 27·9 (23·3– ) 17·5 (15·9–19·6) 12·6 (12·0–13·2) 24·8 (21·3–30·9)
*Age at first sexual intercourse (SI) refers to the group born between 1964 and 1968. Some data for South Africa, Namibia, USA, and Australia cannot be
calculated because less than three quarters are married. Published data used for Chile only, original analyses for all other countries.
Table 2: Age at first sexual intercourse and marriage, for men and women born between 1965 and 1969, by country
5
Industrialised countries Countries in transition Latin America and Caribbean
Australia Armenia Brazil
Men Women Men Women Men Women
25 25 25
24 24 24
22 22 22
20 20 20
18 18 18
16 16 16
14 14 14
12 12 12
10 10 10
1995–99
1995–99
1995–99
1995–99
2000–04
2000–04
1970–74
1975–79
1980–84
1985–89
1990–94
1975–79
1980–84
1985–89
1990–94
1975–79
1980–84
1985–89
1990–94
1975–79
1980–84
1985–89
1990–94
1975–79
1980–84
1985–89
1990–94
1975–79
1980–84
1985–89
1990–94
1970–74
1970–74
1970–74
1970–74
1970–74
1995–99
1995–99
1980–84
1985–89
1990–94
1980–84
1985–89
1990–94
1980–84
1985–89
1990–94
1995–99
1995–99
2000–04
1995–99
1995–99
2000–04
1970–74
1975–79
1970–74
1975–79
1995–99
1975–79
1980–84
1985–89
1990–94
2000–04
1980–84
1985–89
1990–94
2000–04
1995–99
1975–79
1970–74
1970–74
Haiti
Proportion of person-years Men Women
spent: 25
24
Single—never had sex 22
20
Single—had sex at least once 18
16
Cohabiting
14
12
10
2000–04
2000–04
1995–99
1995–99
1975–79
1980–84
1985–89
1990–94
1980–84
1985–89
1990–94
1970–74
1970–74
1975–79
Peru
Men Women
25
24
22
20
18
16
14
12
10
1980–84
1985–89
1990–94
1995–99
2000–04
1980–84
1985–89
1990–94
1995–99
2000–04
1975–79
1970–74
1970–74
1975–79
Figure 2: Sexual experience and relationships between ages 10 and 25 years, by successive age-groups in selected countries.
Vertical axes show person-years of age, and horizontal axes show people who reached the age of 25 years in years shown.
*Data not available for women. †Data not available for men
6
10
12
14
16
18
20
22
24
25
10
12
14
16
18
20
22
24
25
10
12
14
16
18
20
22
24
25
10
12
14
16
18
20
22
24
25
Asia
1970–74 1970–74 1970–74
Nepal
Men
Men
Men
Men
1975–79 1975–79
Ethiopia
1970–74 1975–79
Indonesia
1975–79 1980–84 1980–84 1980–84
Bangladesh*
Women
Women
Women
1980–84 1980–84 1980–84
1985–89 1985–89 1985–89
1990–94 1990–94 1990–94
1995–99 1995–99 1995–99
2000–04 2000–04 2000–04
10
12
14
16
18
20
22
24
25
10
12
14
16
18
20
22
24
25
10
12
14
16
18
20
22
24
25
1970–74 1970–74 1970–74
Kenya
Men
Men
Men
1975–79
Uganda
1975–79 1975–79
Rwanda
South Africa†
1985–89 1985–89
1990–94 1990–94 1990–94
1995–99 1995–99 1995–99
2000–04 2000–04 2000–04
10
12
14
16
18
20
22
24
25
1970–74 1970–74 1970–74 1970–74
1975–79 1975–79 1975–79 1975–79
Women
Women
Women
Women
1985–89 1985–89 1985–89 1985–89
1990–94 1990–94 1990–94 1990–94
1995–99 1995–99 1995–99 1995–99
2000–04 2000–04 2000–04 2000–04
10
12
14
16
18
20
22
24
25
10
12
14
16
18
20
22
24
25
1970–74 1970–74
Men
Men
Nigeria
1975–79 1975–79
Cameroon
1980–84 1980–84
1985–89 1985–89
1990–94 1990–94
1995–99 1995–99
2000–04 2000–04
Central and West Africa
1970–74
1975–79 1970–74
1980–84 1975–79
Women
Women
1985–89 1980–84
1990–94 1985–89
1995–99 1990–94
2000–04 1995–99
2000–04
7
Men higher in women; more men than women report having had no partner in
the past year. In older age-groups (those aged ≥ 20 years), the reverse
is true; women are predominantly monogamous, but large proportions
of men report having more than one sexual partner. The partners of
these men are likely to be the monogamous women from the younger
age-group. Median age differences between spouses in Africa are high;
9·2 years in west Africa, 7·4 years in north and central Africa, and 5·5
years in east and southern Africa (age of non-spousal partners is not
recorded) (table 3).
By contrast, in South America (especially in Brazil) more men than
women report having one or more recent sexual partners in all age-
groups. These findings beg the question of who the men are having sex
with. With a median age difference between partners of 4 years, patterns
of age mixing and age structure do not account for the difference, and
less than 1% of Brazilian men reported sex workers amongst their past
three partners. The Latin macho culture might encourage men to over-
Women
report, and women to under-report, sexual activity.
<25%
Mali 2001 DHS 13·2 (12·9-13·5)
25–49% Nigeria 2003 DHS 12·1 (11·5–12·7)
50–74%
≥75% East and South Africa
Not available from survey Ethiopia 2000 DHS 9·3 (9·0–9·5)
Kenya 2003 DHS 7·7 (7·4–8·0)
Figure 3: Percentage of ever-married men and women born in 1965–1966 who
had sex before marriage Malawi 2000 DHS 6·3 (6·1–6·5)
Mozambique 2003 DHS 7·3 (7·0–7·6)
Namibia 2000 DHS 9·4 (8·6–0·2)
Much is made of the effect on sexual health of the growing prevalence of
Rwanda 2000 DHS 6·7 (6·3–7·1)
sex outside marriage; yet marriage does not reliably safeguard sexual-
Tanzania 2005 DHS 7·5 (7·0–8·0)
health status. In Uganda, married women constitute the population
Uganda 2001 DHS 7·7 (7·3–8·1)
group in whom HIV transmission is increasing most rapidly. A study in
Zambia 2002 DHS 7·2 (7·0–7·4)
Kenya and Zambia showed that the sexual-health benefits of marriage Zimbabwe 1999 DHS 9·1 (8·6–9·6)
for women were offset by higher frequency of intercourse, lower rates of Countries in transition
condom use, and their husbands’ risk behaviour.40 Married women find Armenia 2000 DHS 4·3 (4·2–4·4)
negotiation of safer sex more difficult than do single women41,42 and few Kazakhstan 1999 DHS 2·7 (2·5–2·9)
married women use condoms for family planning.43 Nor can marriage Latin America and Caribbean
be relied on to ensure safer early sexual experience. In Asian countries Bolivia 2003 DHS 3·2 (3·1–3·3)
where early marriage is encouraged to protect young women’s honour, Colombia 2005 DHS 4·5 (4·4–4·6)
early sexual experiences can be coercive and traumatic and, with Dominican Republic 2002 DHS 5·8 (5·8–5·8)
respect to early pregnancy, dangerous for mother and child.44 Haiti 2000 DHS 6·3 (5·9–6·7)
Monogamy is the dominant pattern in most regions of the world Nicaragua 2001 DHS 5·3 (5·0–5·4)
(figure 6). Despite substantial regional variation in the prevalence Peru 2004 DHS 3·9 (3·6–4·2)
countries, most people report having only one recent sexual partner. Nepal 2001 DHS* 4·6 (4·5–4·8)
The data show pronounced asymmetry between men and women Philippines 2003 DHS 3·2 (3·1–3·3)
(figure 6). Worldwide, men report more multiple partnerships than do Industrialised countries
women; only in some industrialised countries are the proportions of Australia 2002 ASHR 1·9
men and women who report multiple partnerships more-or-less equal. USA 2002 NSFG 2·2
Reporting bias45 doubtless accounts for some of the difference between DHS=Demographic and Health Surveys. ASHR=Australian Study
of health and Relationships. NSFG=National survey of Family Growth.
the sexes. However, in Africa, where young people vastly outnumber *Indicates ever-married sample.
older people (figure 6), the difference between men and women can
be largely explained by the age structure and patterns of age mixing Table 3:Age difference between married men and women, by country
(i.e., older men having sex with younger women). In African countries,
in the younger age-groups (those aged 15–19 years), sexual activity is
8
% who have had sex in past mont h % who have had sex in past mont h Person-years of ag e
Ca Ca
m m
60
60
80
90
80
90
40
40
0
0
20
20
16
16
18
18
20
50
20
50
30
30
14
14
10
22
25
22
25
70
70
100
100
10
10
10
12
12
er er
o oo
Ga on Ch n 2
bo 20
n 0 Ga ad 004 Ca Gab Ca Gab
CA 20 4 bo 200
R 0 m on
er
m on
er
Ch 19 0 CA n20 4 o o
Single
ad 94 R 00 C h on
Married
19 Ch on
To 200 Cô 94 ad ad
Central Africa
te
Cohabiting
Bu go 4
Central Africa
rk Be 19 dI’ Se Se
in ni 9 vo
a n 8 ir ne ne
To e 1 g
Cô G Faso 200 g 99 Gh al Gh ga l
te han 2 1 Be o 1 8 Bu B ana Bu B ana
West Africa
West Africa
in 19 1 Ni ia 2 99
Ivo vI o
ge 00 ire ire
M Nig ea 1 98
oz er 9 M r19 3 M M
am ia 99 al 98 Gu al i Gu ali
i2 in in
bi 200 00 e
Za que 3 1 Ni a Ni e a
m 2 ge
r
ge
r
M bi 003 Na
al a 2
m
Za aw 00 Ta ib R
Ta mb i 20 2 nz ia
nz ia 00 2 Cowan
a 2 M ani 00 m da Rw
a a 0 N or
K ni 00 M Za law 200 Na and
C en a 2 2 oz m i 2 5 Zi am os
So om ya 005 So m b ibi Zi m a
u or 20 Zi amb bia 000 m ib
m iq 20 ut ab a b ia
Zi th A os 1 03 ba u 0 h w Ta abw
m fr 9 bw e 2 2 A e nz e
ba ic 96 Ta fric a
a Ke e 1 003 nz a Ke nia
Et bwe 19
hi 1 98 Et nya 999 M K ania n
o h 2 ad e M ya
a
M Ug pia 999
Men
Men
ag ny
Women
ag d 00 a a 3 as a hi i
a 0 Rw nd 20 M car
Rw asca 20 an a 2 00 a Ug opi
a
an r 1 01 da 00 Et law M Z and
hi i
Az da 997 20 1 o oz am a
am b
er 2 00 East and South Africa Ug pia
b 00 Ar M Z an bi i a
qu
Ge aija 0 m oz a da e
o n Al eni
am mb
Al rgi 20
b a 0 b a bi ia
qu
M an 20 1 Ro an 20 e
o ia 0 Ka m i a 2 00 Ka Ka
Ar ldo 20 5 za an 00 za za
m v 0 kh ia 2 kh kh
Figure 5: Percentage of 15–24 year olds who had sex in the 4 weeks preceding survey, by marital status
Ro en a 19 2 st 19 Ar sta Ar sta
Uz m ia 9 an 96
m n m n
b an 20 7 19 en e
Ka eki s ia 1 00 99 ia Tu nia
za ta 99 rk
transition
ey
Countries in
kh n 1 6
st 99 Do H
an 6 m Bo ait Pe
in liv i 20
i
Countries in transition
Ha 19 ca i 0 Bo ru Do
Bo iti 99 n Pe a 20 0 liv m
liv 20 Re ru 0 Ni Br ia in Pe
Ni P ia2 00 pu 19 3 c a ica
Do
c
m G ar eru 003 b 9 Do Gu raazil n Bo ru
in ua ag 2 Br lic 2 6 m at gu Re liv
ica te u 0 az 0 in C e m a pu ia
n m a 2 04 il 1 02 ica ol ala bl
Re ala 00 Ph 99 n R om
ilip 6 ep bia
Ha ic
iti
Co pub 19 1
Latin America
pi ub
and Caribbean
lo lic 99 ne
m 2
Figure 4: Sexual experience and cohabitation between ages 10 and 25 years for people who reached the age of 25 years in 1995–1999
b 0 Ne s 2 Ha lic Ph
Latin America
Br ia 2 02 it i
and Caribbean
pa 00 Ph ilip
az 0 l2 3 ili In pi
Ph Ne il 1 05 00 do ne
il p 99 Au 1 In ppin
do e ne s
In ippi al * 6
do ne 20 st ne s s
ne s 2 03
ra
li a s Ba N e i a
† Ne i a ng pa
Asia
Married refers to people currently married. *Only ever-married women surveyed. †Respondents aged 16–24 years. ‡Respondents aged 18–24 years
sia * 00 pa la l
2 3 U 2 l de
Asia
Au U 003 Br SA 002 sh
ita 2
CAR=Central African Republic. Denominator for single people excludes those who have never had sex. Single refers to people who have never married.
st SA Fr in † 002 B
ra 2 Au rita
Br lia † 00 an 20 Br
ce‡ 00 st in ita
Fr itain 200 2 ra in
an † 2 20 lia
ce 20 04 US US
‡ 00 A A
countries
20
9
countries
Industrialised
04
Industrialised
Industrialised countries r
More than one sexual partne One sexual partner No sexual partners
12 10 8 6 4 2 0 2 4 6 8 10 12
1 0 00
0
20 00
3 0 00
20 00
3 0 00
0
25 0
10 0
00
10 0
00
0
00
25 0
15 0
00
15 0
25 0
00
15 0
00
15 0
20 00
25 0
1 000
20 0
0
75
75
25
25
50
50
50
50
50
50
0
0
0
0
0
10
30
30
10
Thousands Thousands Thousands Millions
Countries intransition Central and west Africa
0
00
0
00
0
75
75
25
25
50
50
10
10
Thousands Thousands Thousands Millions
East Africa
00
00
00
00
20 00
3000
00
00
25 0
10 0
00
15 00
00
15 00
25 0
10 0
20 0
0
50
50
0
50
50
0
00
00
15
15
10
20
20
10
30
10 8 6 4 2 0 2 4 6 8 10
00
00
00
0
00
00
0
00
0
0
0
0
0
0
20
20
50
50
40
40
60
60
20
20
40
40
60
60
15
15
10
10
20
20
Millions Thousands Thousands Thousands
Figure 6: Population distribution by age and sex and number of sexual partners in the past year
These charts present results obtained by applying survey-based proportions to UN population estimates for the year 2005. 27
Population numbers are shown in thousands, or in millions, depending on the country
These data do not capture whether partnerships exist concurrently or The proportion of sexually active young people who report using a
serially. Concurrent sexual partnerships (those that overlap in time) allow condom for their most recent sexual intercourse is higher for men than
more rapid spread of sexually transmitted infections than do the same for women (figure 7). Much of this difference is because young women’s
rate of new sequential partnerships. Evidence is available that, although partners are more likely to be their husbands, with whom condoms
lifetime numbers of partners might be lower, concurrent relationships are less often used.48,55 In most countries for which two estimates
in men in some African countries might have been more common and are available, condom use at last sexual intercourse is increasing—in
of longer duration than in other regions.46–50 Many men who have sex some cases, for example in Uganda, strikingly so. Use of condoms to
with men have also had sex with women51 and, apart from in developed prevent pregnancy increased from 5% to 19% of single women in 19
countries, are unlikely to identify as being homosexual52 The socially African countries between 1993 and 2001, and use of a condom at last
censored nature of same-sex activity could lead to under-reporting and intercourse increased from 19% to 28% during the same period.43 In
might also account for the absence of such activity from the research industrialised countries, rates of condom use are generally higher than
agenda. A recent review of the prevalence of same-sex activity in men those in non-industrialised countries, especially in women. The increase
identified 67 studies,53 none of which were from Africa, the middle east, in condom use in recent years has also been substantial in industrialised
or the English-speaking Caribbean. Estimates of lifetime prevalence of countries. No comparative data are available for the consistency with
sexual intercourse with men ranged from 3–5% of men in east Asia which condoms are used.
6–12% in south and southeast Asia, 6–15% for eastern Europe, 20% for
In countries with wide differences between men and women in the
Latin America.53 The prevalence is 6% in the UK15 and 5% in France.54
prevalence of premarital sexual inter-course, young men are more
About half those with lifetime homosexual experience report having had
male partners in the past year.53
10
likely to report having intercourse with sex workers.44 Estimates of the Explanations for the variation
proportion of men who are clients of sex workers range from 1% to The regional variation in sexual behaviour underlines the powerful role
14% dependent on region.30 Varying operational definitions of sex work of environmental factors in shaping behaviour and its consequences
frustrate efforts to interpret prevalence at different times and in different for sexual health. Through the interplay between demographic and
places. The continuum of sexual exchange ranges from expectation of structural factors, social norms, and public policies, spatial differences
gifts or favours within personal relationships, to more formal trading of can be properly understood. One of the most notable features of the
sex for money.56 The proportion who reported having “sex in exchange data, for example, is the striking gender difference in sexual behaviour.
for money, gifts, or favours”30 in the past year is highest in countries in Women might be disadvantaged in protecting their sexual health if their
central and southern Africa (medians 13·6% and 11·3%, respectively), partner is older than them, of higher status than them,58,59 or if they are
followed by eastern and west Africa (9·8% and 8·9%, respectively).30 beholden to a man for favours, goods, or money in return for sex.60,61
More recent African surveys that used the more restricted definition of Women’s power to maintain monogamous relationships might be
paying for sex30 have reported reduced prevalence. In the Caribbean, diminished in locations in which they outnumber men with whom they
6–7% of men reported having paid for sex, and estimates for Latin might have sex. Women might outnumber men in this way as a result of
America, eastern and western Europe, and central Asia were less than the age structures of populations and patterns of age mixing, or where
3%. Cautious estimates are available for northern Africa and the middle cultural practices such as polygyny are prevalent10,62,63 and where high
east (1–3%), south Asia (3–5%), southeast Asia (3–10%), and China and levels of imprisonment of black men distort sex ratios in predominantly
Hong Kong (11%).30 African-American communities—as in the USA for example.64
Estimates from the WHO study on gender-based violence of lifetime Poverty, deprivation, and unemployment work with gender relations
prevalence of sexual violence by an intimate partner range between to promote change of partner, con-currency, and unprotected sex.64,65
10% and 50%.57 DHS data lend support to this finding; between 11% Economic adversity restricts the power of men and women to take
and 49% of women say they cannot always refuse sex. In more than half control of their health;66 deprivation and unemployment might drive
the WHO settings, more than 30% of women who reported first sexual men and women to sell sex10,12,67,68 or travel greater distances to work.
intercourse before age 15 years described having been forced into it, Being away from home is associated in both developed and developing
and three-quarters of women who had been abused since the age of 15 countries with concurrency of partnerships69 and an increase in risk
years identified the perpetrator as their intimate partner. behaviours.46,70–73 Possibly the most powerful influences on human
70
Men
60
50
Used a condom (%)
40
30
20
10
0
ic
Za e
e
o
n
a
da
*
nia
ad
ya
n
ia
†
n
i
bl
qu
bw
al
ce
as
oo
bi
A
ni
ai
iv
pu
n
an
Ch
US
za
m
aF
an
it
bi
Be
l
Ke
er
ba
Bo
Br
Ug
Re
n
am
m
Fr
kin
m
Ta
Ca
n
oz
Zi
r
ica
Bu
in
m
Do
80
70
Women
60
50
Used a condom (%)
40
30
20
10
0
o
n
ia
a
da
*
e
ad
a
n
ic
in
ru
ia
A†
a
i
gu
al
as
bi
ce
oo
bi
qu
ny
bw
ni
an
bl
liv
ita
Pe
an
Ch
om
US
m
aF
an
Be
ra
Ke
er
bi
pu
nz
Bo
ba
Br
Ug
Za
ca
m
Fr
in
am
l
Ta
Re
Co
m
Ca
Ni
rk
Zi
oz
n
Bu
ica
M
in
m
Do
CentralAfrica WestAfrica East and South Africa LatinAmerica and Caribbean Industrialised countries
Figure 7: Percentage of 15–24 year-olds who reported condom use at last sex in successive surveys
*18–24 year-olds. †15–19 year-olds only; comparable data are available for 15–24 year-old men and women only for 2002: 58% for
men, 44% for women. Data shown are for respondents who reported being sexually active in past year
11
sexuality are the social norms that govern its expression. Morals, taboos, exclusively on expectations of individual behaviour change are unlikely
laws, and religious beliefs used by societies worldwide circumscribe to produce substantial improvements in sexual-health status. They are
and radically determine the sexual behaviour of their citizens. The scale especially inappropriate to poor country circumstances, where sex is
of the regional diversity in sexual behaviour is matched only by the more likely to be tied to livelihoods, duty, and survival,78 and where
range of cultural constraints on practice. In some societies, for example, individual agency is restricted. In wealthier countries, personal choice
homosexual behaviour is celebrated in public parades of pride; in others is greater than in poorer countries, yet power inequalities persist.79
it carries the death penalty. In some countries, such as Brazil, condoms Numerous calls have been made for public-health interventions to
are available to young people in schools; in others, for example in parts pay greater heed to the social context within which sexual intercourse
of Indonesia, their possession is a criminal offence.12 Such strictures occurs.78,80–83 Interventions encouraging adoption of risk reduction
hinder attempts by men and women to protect their sexual health. practices remain a cornerstone of sexual-health promotion but the
The sexual double standard, whereby restraint is expected of women, evidence shows that they need to go beyond mere provision of
whereas excesses are tolerated for men, compounds the problems for information to be effective (table 4). Systematic reviews have focused
both men and women.44,74–77 mainly on assessment of interventions to change individual behaviour
and show increased effectiveness where information is supplemented
Interventions to improve sexual health by skill building and counselling, such as use of condoms and safe sex
With the diversity of sexual behaviour, a range of preventive strategies negotiation86,91,93,96,97 where theory guides design,92,94 where several
are needed to protect sexual health. And in view of the importance of delivery methods are used,85,96,97 and where context and the need for
the broader determinants on sexual behaviour, approaches focusing sustainability are taken into account.94–97
12
(Continued fr om previous page)
Drug users ||
94
Coyle (1999) Intravenous drug USA* Street based outreach Hidden populations should be targeted Limited details of search strategy
users out of risk reduction on the street or in other settings frequented by them Reviewing processes not reported
treatment programmes No quality assessment
No consideration for heterogeneity
95
Gibson (1998) Intravenous drug USA,* Australia Individual counselling, Characteristics associated with success: Limited details of search strategy
users in or out of group, voluntary HIV Intense and sustained interventions English language reports only
treatment counselling and testing, More stable and motivated subgroups Statistical significance not reported
street outreach, social No detail for assessment of outcomes
in primary studies
96
Semaan (2002) Intravenous drug USA* Social and behavioural* Intervention recipients more likely to reduce sexual- Pooling despite heterogeneity
users and interventions risk behaviours than comparison groups
non-injecting
drug users
97
Van Empelen Intravenous drug USA,* Australia Psychosocial and Characteristics associated with success: No language or date restrictions
(2003) users and behavioural Use of multiple theories and methods reported
non-injecting interventions (individual, Inclusion of peers
drug users in or group and community) Rehearsal of skills
out of treatment Element of sustainability
Social context must be considered
Young people
Kirby 98(2006) Young people USA,* Canada, Curriculum and Features relating to development, content, and No language restrictions reported
aged 9–24 years Netherlands, group-based education implementation of curriculum associated with Limited quality assessment
Norway, Spain, UK, programmes typically success Vote-counting method used for data
Belize, Brazil, Chile, focused on pregnancy or synthesis
Jamaica, Kenya, HIV and sexually No formal consideration of
Mexico, Namibia, transmitted infection heterogeneity
Nigeria, South Africa, prevention behaviours
Tanzania, Thailand,
Zambia
99
Mullen (2002) Adolescents aged USA* Behavioural* or social Interventions targeting single ethnic groups outside of Inclusion criteria not strictly adhered
13–19 years programmes done in the classroom showed larger effects than those with to. Limited reporting of reviewing
*
(sexually and out of the classroom mixed ethnic groups in either setting processes
experienced) Pooling despite heterogeneity
*Half or more than half the primary studies were undertaken in this setting. †More than half the primary studies examined interventions of this type. ‡All were direct contact
interventions according to CHAPS framework. §Most participants defined themselves as heterosexual, but homosexual participants were also included. ¶More than half the
primary studies examined only these participants. ||All included systematic reviews reported outcomes related to sexual behaviour change.
Individual-based interventions also need to be targeted to be successful.88 take account of the diverse reasons for having sexual intercourse109 and
Men have been successfully targeted in occupational contexts with for changing sexual behaviour.
consequent reduction in sexual-risk taking.87 Young people are most
Reviews of individual-based interventions also emphasise the
commonly targeted in schools and the evidence is that curriculum-
importance of interventions to address social norms that act against
based sex education does not increase risky sexual behaviour as many
safer sex.88,89,91 The effects of behaviour change interventions will be
fear.98 Systematic reviews have shown school-based sex education to
transient if participants return to an unsupportive environment. Actions
lead to improved awareness of risk and knowledge of risk reduction
such as deferring or denying sex are not intrinsically rewarding and need
strategies, increased self-effectiveness and intention to adopt safer
to be supported by group norms. Community interventions have been
sex behaviours,98 and to delay, rather than hasten, the onset of sexual
effective in mobilising local groups in support of preventive strategies.
activity. Broad-spectrum strategies to achieve behaviour change,
The prompt response from homosexual communities to the prevention
with mass-media communication, have proved effective in increasing
of HIV/AIDS in industrialised countries in the early 1980s owed much to
awareness and knowledge, and in reducing high-risk behaviour.100
the preexistence of non-governmental organisation (NGO) infrastructures
In this context, techniques used in social marketing, which target
and to the visibility and mobilisation of homosexual men. The evidence
individuals according to their lifestyles, values, and risk status are an
is that information gained through social networks is more salient, and
improvement on conventional targeting approaches with demographic
more likely to lead to behaviour change, than that conveyed by more
characteristics alone.101–103
impersonal agencies.110
Tailoring of behaviour-change interventions to individual needs and
Where population groups are marginalised and NGOs and community-
circumstances is also essential. A range of messages is needed that
based service programmes are weak, the starting point might be the
respect diversity and preserve choice. Enough studies have shown that
informal groups in which norms are maintained. Preventive programmes
first intercourse is retrospectively regretted by many women, and some
that use naturally occurring social networks have reduced risky
men,104 for efforts to help young people to achieve the best timing of
behaviour in homosexual men in Russia,111 increased contraceptive
first sex to be justified. Yet, abstinence might not be an option where first
use in married women in Bangladesh,112 increased condom use in
sexual relations are forced,31,44 where the sexual abuse of adolescents
sex workers in India,113 and have proved more effective in changing
is common,105 and where financial circumstances force young people
norms than more orthodox approaches using conventional health care
to sell sex. Moreover, people do not adhere to only one type of sexual
and field workers. Strong social prohibitions and sanctions, especially
behaviour. Many men who have sex with men also have sex with
those underpinned by legislation, might present greater challenges. In
women, and different preventive strategies might be used for these two
general, laws protect the young and those vulnerable to coercion and
behaviours.106 Broader sexual repertoires need to be taken into account.
exploitation, but they might also impede safer sex practices. If practices
The issue of recommending non-penetrative sex (such as mutual
are illegal, they are more likely to be engaged in a furtive or clandestine
masturbation) is seldom tackled.107,108 Risk reduction messages need to
13
Panel 3: 100% condom-use programme in Thailand
The 100% condom-use programme in Thailand, implemented nationally in 1991, has been widely recognised and documented as an impressive
success in achievement of behaviour change. The programme aimed to bring about 100% condom use in commercial sex. Brothel owners, brothel-
based sex workers, and their clients were targeted with simultaneous National AIDS Education and Condom Promotion Campaigns125 that used the
mass media and workplace programmes. Condoms were made freely available at all sex establishments.
The programme was linked to a longstanding government venereal-disease programme in each province. Sex workers receive regular physical
check-ups for sexually transmitted infections, and free condoms. Men who presented with a sexually transmitted infection were asked to name the
sex establishment that was the source of infection as evidence of non-compliance with the policy. Public-health officials encouraged sex workers
to be tested and treated, venereal-disease units provided extensive contact tracing, and law-enforcement authorities took any necessary legal
action against brothel owners.125 Co-ordination between national and local government, public-health officials, and brothel owners was a strong
characteristic of the programme. Although prostitution is illegal in Thailand, there has long been tacit cultural acceptance of patronage of sex workers.
No attempt was made to eliminate commercial sex or change public morality. With a high level of political commitment and no fear of religious
sensitivities, a pragmatic approach to condom promotion was possible.
The success of the campaign was evidenced by an increase in condom use and a decrease in commercial sex and rates of sexually transmitted
infection in sex workers and their clients.125 The most convincing evidence is the rate of new HIV infections in 21-year-old conscripts in northern
Thailand, which fell from 3·3 per 100 person-years to 0·3 per 100 person-years between 1991 and 1995.124 Condom use in these men rose from
61·0% to 92·5% in the same period.126 The proportion of indirect sex workers (those working from bars, restaurants, etc) increased from half to
almost two-thirds, in the first half of the 1990s.124 Although the focus was on making sexual intercourse with sex workers safer, a substantial
reduction took place in both supply and demand for sex work. Fear of AIDS might have stopped women entering the trade, and more women started
working as indirect sex workers charging higher rates.127 Casual sex with both men and women with low levels of protection increased.126 The
proportion of northern Thai army conscripts who report visits to sex workers fell, and decreased even more sharply in older men126 from 81% in
1991 to 64% in 1995. The success of the Thai programme is attributable to efforts made on several fronts. Strategies limited to voluntary individual
behaviour change would have been unlikely to achieve such radical change without sociostructural changes.
manner, and opportunities for protection are constrained. The success of social interventions. The range of people to be engaged in partnership
preventive strategies is heavily dependent on acceptance of the reality of is broad and includes economists, politicians, industry, the judiciary, and
sexual practices that are socially censured. Condom use is uncommon in NGOs. A way of ensuring that joint action takes place is to make it not
sex workers in India113,114 for example, where commercial sex is heavily merely a generalised goal of interventions, but an explicit component of
socially proscribed. By contrast, condom use is near-universal in those the programme (as in the case of the UK Teenage Pregnancy Strategy123).
in Kampala115 and in Mexico,116 where public-health agencies have Intervention at a structural level needs political will and commitment,124
actively and openly engaged in co-operation with female sex workers. not least to dismantle legal and other obstacles to pursue strategies to
The mass media have been powerful in shifting social norms and, in protect sexual health. The successes of Thailand and Uganda in reducing
some instances, achieving legislative reform. In the Lebanon, where HIV rates show the importance of political leadership (panels 3 and 4).
homosexuality is illegal, a popular weekly television programme includes
The evidence shows that, where improvements in sexual health have
a homosexual voice.117 In South Africa, the strategic use of entertainment
been achieved, a combination of factors has contributed to the success.
as education, media advocacy, and social mobilisation to change public
Behaviour change in Thailand and Uganda, for example, has been
opinion and influence decision-makers has led to implementation of the
attributed to an array of preventive policies and strategies, mounted
Domestic Violence Act.118
by different agencies, with strong partnerships between the media,
To address the broader structural determinants of sexual behaviour is government, NGOs, sex workers, people living with HIV/AIDS, and
daunting. To do so demands a broader definition of public health than international and local public-health agencies, endorsed at the highest
many might feel comfortable working within. Social determinants are political level. Yet, the polarity of views on abstinence, be faithful, and
the least amenable to intervention. Structural factors such as poverty, use condoms (ABC) strategies,131 about which of these three elements
unemployment, and gender are difficult to modify, and social and has contributed most to reductions in rates of HIV in Uganda or teenage
political norms are slow to shift.63,81,83,119 Yet, efforts are being made to pregnancy in the USA (panels 4 and 5), stems from a search for single-
address forces such as gender and poverty in innovative ways. These factor explanations to support particular ideological positions. The
efforts include mainstreaming of HIV and sexual health services in preoccupation with ABC strategies has the negative effect of distracting
development; empowerment of sex workers to avoid sex work through attention from the need for broader, integrated programmes in which all
business and information-technology training; and integration of sexual- components are mutually reinforcing.
health education into microfinance schemes.120 Success has also been
The mix of components in national programmes needs to be
achieved in tackling of social attitudes, for example, those of young men
tailored to the local context.148 Comparisons of HIV policy in different
towards fatherhood, relationships, and contraception.121,122
countries104,132,149,150 show the importance of ensuring that public health
Addressing of structural determinants, particularly poverty, demands approaches are culturally appropriate and timely. In Brazil, for example,
the involvement of social as well as health sectors, and so requires co- the adoption of sex positive approaches and explicit condom promotion
ordination and collaboration across sectors and agencies, and with other has been well managed and has ignited little political controversy.149 In
14
Panel 4: HIV in Uganda
Uganda’s success in reducing HIV prevalence115,128 and improving reproductive-health status129 compared with neighbouring countries has been
attributed to the selective emphasis on the abstinence and being faithful strands of the ABC strategy in the country’s HIV programmes.34 The
suggestion has been made that later onset of sexual activity and a reduction in non-regular sexual partners (a 65% reduction from 1991 to 1998)
have been more important than condom use in curbing the HIV epidemic.130 These conclusions have been interpreted as providing evidence of the
merit of abstinence-based approaches to HIV prevention generally.130,131
Several features of Uganda’s epidemiological situation and social-context suggest that Uganda’s success should not be attributed to a few specific
interventions. The first feature relates to the timing of events. The fall in prevalence of HIV corresponds to a drop in incidence from the beginning of
1985, when Uganda did not have public national HIV-prevention programmes in place. Furthermore, as our data show, the trend towards older age
at first intercourse occurred gradually for women from the 1970s to the present—i.e., before the start of HIV-prevention programmes—for men
remarkably little change in age at first intercourse has taken place in recent decades. However, evidence shows that other changes in behaviour
have taken place. Condoms were cautiously and gradually introduced in Uganda132 and were largely unavailable to the general population during the
1980s, but rates of condom use were high in high-risk groups, such as sex workers.115
These behaviour changes have been attributed to successful public education campaigns in Uganda. Public-health agencies gained the confidence of
communities by galvanising support from local leaders and the church for the Love Carefully and Zero Grazing messages,133 and only later introduced
condom-promotion messages to the programme. Yet awareness of the severity of AIDS, its effect on family and friends,134 and the willingness
of those affected to publicly endorse behaviour change also played their part.133 Moreover, many have pointed to the broader factors that have
contributed to the reduction in HIV prevalence in Uganda, many of which preceded current HIV-prevention programmes.128 These include successful
assimilation of scientific knowledge about modes of transmission,135 re-establishment of community life with restored civil stability at the end of
the Ugandan civil war in 1986, which substantially reduced prevalence of sexually transmitted infections,129,136 and two decades of strong policy
support.128
Uganda was the first African nation to establish a national AIDS programme137 and the response to the epidemic went beyond individual-level factors.
Factors included concerted and integrated efforts made at community level, including mobilisation of local groups, widespread participation of several
NGOs in HIV prevention, collaborative partnerships with religious groups and community activists, ample funding, and openness about the scale of
the problem and commitment to tackling it at the highest political level. It is the lessons from this joint approach that are most valuably transferred to
other settings.137
15
Botswana, by contrast, the premature introduction of condom messages Conclusions
to promote condom use without attention to cultural norms served to This is the first time an attempt has been made to bring together
undermine public confidence in public-health agencies.132 Attention comprehensive survey data for sexual behaviour from around the world.
needs to be given to the feasibility and acceptability of interventions that The data show perhaps less change over time than might have been
have worked in one setting before introducing them into another. Too supposed. People who fear a tide of youthful promiscuity might take
often, locally successful models are scaled up without communication heart from the fact that trends towards early and premarital sex are
and engagement with wider structures and forces.151 Transferability of neither as pronounced nor as prevalent as is sometimes assumed.
interventions is now being tested across areas of India with the roll- Similarly, the apparent absence of an association between regional
out of the Sonagachi project, 152 an acclaimed community-mobilisation variations in sexual behaviour and in sexual-health status might also
strategy that involves the active participation of Calcutta’s female be counterintuitive. In particular, the comparatively high prevalence of
brothel-based sex workers. Since most commercial sex in India is street multiple partnerships in developed countries, compared with parts of
based,113 generalisability of this strategy will be put to the test. the world with far higher rates of sexually transmitted infections and
An assessment of interventions that adapt elements of the Sonagachi HIV, such as African countries, might hold some surprises. Only rates of
project and the Thai 100% condom programme (panel 3) to the condom use are predictably lower in countries with lower sexual-health
Dominican context,152 compared the effects of community solidarity status, and this is likely to be attributable to factors relating to access
with and without changes in government policy and showed them to and service provision. The data make a powerful case for an intervention
be greater and more sustained when changes in government policy focus on the broader determinants of sexual health, such as poverty and
were included in the model. More intervention studies of this kind are mobility, but especially gender inequality.
needed to assess the effect, not of a single-component intervention in The comparative data are important in countering misinformation and
one setting, but of more than one approach, drawing on components quelling fears relating to sexual behaviour. The selection of public-
from different models in different combinations and in settings other health messages needs to be guided by epidemiological evidence rather
than those in which they originated. than by myths and moral stances. The greatest challenge to sexual-
The adoption of multifaceted interventions with a balanced emphasis on health promotion in almost all countries comes from opposition from
changes to individual behaviour and the social context has implications conservative forces to harm-reduction strategies. Governments tend to
for what counts as evidence of effectiveness. First, the broader the be wary of controversy and, faced with resistance from groups with a
scope of the intervention, the less well it lends itself to assessment by strong moral agenda, shy away from supporting interventions other than
experimental methods that are widely regarded as essential to prove those with orthodox approaches. Policy-makers and programme planners
effect.153,154 Yet a broad scope is what is needed. Interaction and synergy need strong evidence of beneficial effect to make the case to address
between components need to be seen as valued goals of behavioural stigmatised groups and adopt messages that do not support the dominant
interventions rather than obstacles to experimental research. Controlled ethos of monogamous, procreative, and heterosexual sex. Policy-makers
trials might be the assessment method of choice for individual-based and programme planners need to be able to show that the effect on sexual-
interventions, but in the case of those that tackle social-contextual health status of providing services to unmarried young women, supplying
factors, more can be learned from country case studies that document condoms, decriminalising commercial sex and homosexual activity, and
the experience of implementation of programmes in specific settings. prosecuting people who commit sexual violence is likely to be beneficial
Second, strategic decisions with respect to sexual-health promotion rather than detrimental, and that to do otherwise will force stigmatised
need to be based on a strong understanding of process.155 The emphasis behaviours underground, leaving the most vulnerable people unprotected.
must be not be only on which approaches work, but also on why and Scientific evidence of effectiveness will counter misconceptions (for
how they do so in particular social contexts. example, that sex education encourages promiscuity).
Third, policy-makers and programme designers might need to abandon Sexuality is an essential part of human nature and its expression needs
their preoccupation with seeing progress towards biomedical endpoints to be affirmed rather than denied if public-health messages are to
as a measure of success.156 The link between sexual behaviour and be heeded. As we have seen, men and women have sex for different
rates of HIV and other sexually transmitted diseases and unplanned reasons and in different ways in different settings. This diversity needs
pregnancy is anyway not always easily drawn. Comparative studies have to be respected in a range of approaches tailored to whole societies, and
failed to explain the heterogeneity of HIV transmission in Africa157 and to particular groups and individuals within them. Public-health strategies
developed countries54,158 in terms of differences in sexual behaviour, and include health promotion, social marketing, media advocacy, legislative
assessment studies have not always shown a link between behaviour activities, and community empowerment. Strategies used should enable
change and health outcomes.159 Where the goals of interventions are people to make their own choices, rather than have them imposed on
predominantly social, a reappraisal of the choice of endpoints is essential. them. Goals related to the improvement of sexual health need to be
When strong evidence exists of a link between changes in the social linked with development goals, and the lynchpin here is partnership
context and an improvement in sexual health status, achievements that between statutory agencies, between sectors, and between national and
relate to domestic violence,118 gender empowerment,160 and changes local agencies, underpinned by political leadership.
in social norms121 can be seen as valid interim endpoints. Failure to Evidence from behavioural interventions shows that no general approach
recognise them as such might prevent potentially useful interventions to sexual-health promotion will work everywhere and no single-
from being scaled up,161 which in turn has implications for the timescale component intervention is likely to work anywhere. Comprehensive
of assessments, the length of which has to take account of the slower multilevel, multipartner behavioural interventions are needed that take
process of social change. account of the social context in mounting individual-level programmes,
attempt to modify social norms to support uptake and maintenance of
behaviour change, and tackle the structural factors that contribute to
risky sexual behaviour.
16
Conflict of interest statement 15 Johnson A, Mercer C, Erens B, et al. Sexual behaviour in Britain:
partnerships, practices, and HIV risk behaviours. Lancet 2001; 358:
We declare that we have no conflict of interest.
1835–42.
Acknowledgments
16 Beltzer N, Lagarde M, Wu Zhou X, Vongmany N, Gremy I. Les
We thank Joanna Busza, John Cleland, Roger Ingham, Shireen Jejeebhoy, Connaissances, Attitudes, Croyances et Comportements Face au
Simon Lewin, Cicely Marston, Sandra McDonagh, Justin Parkhurst, Nono VIH/sida en Ile de France - Evolutions 1992, 1994, 1998, 2001, 2004.
Simelela, Lindsey Gilroy, Alfred Spira, James Trussell and Ravi Verma Paris: Observatoire Régional de Santé d’Ile de France, 2005.
for commenting on earlier drafts of this paper; Ahmet Gulmezoglu and
17 Groves RM, Benson G, Mosher WD, et al, for the US Centers for
Anna Glasier for guidance with this paper, Lawrence Finer and Suzette
Disease Control. Plan and operation of cycle 6 of the National Survey of
Audam for their assistance with data analysis; and the UK Department
for International Development for funding ES and MC during the time Family Growth. Vital Health Stat 2005; 42: 1–86.
this paper was written. 18 Facultes Universitaires Saint-Louis à Bruxelles Centre d’études
References sociologiques. EU New Encounter Module Project. May 13 2003: http://
centres.fusl.ac.be/CES/document/SIDA_AC2/EN/rechsida.AC21.html
1 Bernstein S, CJ Hansen, for the UN Millennium Project. Public (accessed Sept 22, 2006).
choices, private decisions: sexual and reproductive health and the
Millennium Development Goals. 2006. http://www.unmillennium 19 Herold J, Seither R, Ylli A, et al. Reproductive health survey Albania
project.org/documents/MP_Sexual_Health_screen-final.pdf. (accessed 2002: preliminary report. Atlanta US Centers for Disease Control and
Sept 22, 2006). Prevention, 2003.
2 Ezzati M, Vander Hoorn S, Rodgers A, Lopez AD, Mathers CD, Murray 20 Serbanescu F, Imnadze P, Bokhua Z, Nutsubidze N, Jackson D,
CJL. Estimates of global and regional potential health gains from Morris L. Reproductive health survey Georgia 2005: preliminary report.
reducing multiple major risk factors. Lancet 2003; 362: 271–80. Atlanta: US Centers for Disease Control and Prevention, 2005.
3 Slaymaker E. A critique of international indicators of sexual risk 21 Serbanescu F, Morris L. Young adult reproductive health survey
behaviour. Sex Transm Infect 2004; 80 (suppl II): 13–21. Romania 1996: final report. Atlanta: US Centers for Disease Control and
Prevention, 1998.
4 Collumbien M, Gerressu M, Cleland J. Non use and use of ineffective
methods of contraception. In: Ezzati M, Lopez A, Rodgers A, Murray 22 Serbanescu F, Morris L, Stratila M, Bivol O. Reproductive health
C, eds. Comparative quantification of health risks: global and regional survey Moldova 1997: final report. Atlanta: US Centers for Disease
burden of disease attributable to selected major risk factors. Geneva: Control and Prevention, 1998.
World Health Organization, 2004: 1255–320. 23 Serbanescu F, Morris L, Rahimova S, Stupp P. Reproductive health
5 Miller AM, Vance CS. Sexuality, Human Rights, and Health. Health survey Azerbaijan 2001: final report. Atlanta: US Centers for Disease
Hum Rights 2004; 7: 5–16. Control and Prevention, 2003.
6 Hubert M, Bajos N, Sandfort T, eds. Sexual behaviour and HIV/AIDS in 24 Arredondo A, Goldstein E, Pia Olivera M, Bozon M, Giraud M,
Europe. London: University College London Press, 1998. Messiah A. Estudio nacional de compotamiento sexual: primeros
analisis. Chile: Gobierno de Chile Ministerio de Salud, 2000.
7 Cleland J, Ferry B, eds. Sexual behaviour and AIDS in the developing
world. Geneva: World Health Organization, 1995. 25 Liu D, Ng M, Zhou L, Haeberle E. Sexual behaviour in modern China:
report on nationwide survey of 20 000 men and women 1997. New
8 Demographic and Health Surveys. MEASURE DHS, 1984–present. York: The Continuum Publishing Company, 1997.
http://measuredhs.com (accessed Oct 6, 2006).
26 Biddlecom A. Trends in sexual behaviours and infections among
9 Zheng Z, Zhou Y, Zheng L, et al. Sexual behaviour and contraceptive young people in the United States. Sex Transm Infect 2004; 80 (suppl
use among unmarried young women migrant workers in five cities in 2): 74–79.
China. Repro Health Matters 2001; 9: 118–27.
27 UN Population Division. World population prospects: The 2004
10 Mufune P. Changing patterns of sexuality in northern Namibia: revision population database, 2004.
implications for the transmission of HIV/AIDS. 2003; 5: 425–38.
28 Curtis SL, Sutherland EG. Measuring sexual behaviour in the era
11 Cameron KA, Salazar LF, Bernhardt JM, Burgess-Whitman N, of HIV/AIDS: the experience of Demographic and Health Surveys and
Wingood GM. Adolescents’ experience with sex on the web: results similar enquiries. Sex Transm Infect 2004; 80 (Suppl 2): 22–7.
from online focus groups. J Adolesc 2005; 28: 535–40.
29 Nnko S, Boerma JT, Urassa M, Mwaluko G, Zaba B. Secretive
12 Simon S, Paxton SJ. Sexual risk attitudes and behaviours among females or swaggering males? An assessment of the quality of
young adult Indonesians. Cult Health Sex 2004; 6: 393–409. sexual partnership reporting in rural Tanzania. Soc Sci Med 2004; 59:
13 Parker RG, Easton D, Klein CH. Structural barriers and facilitators 299–310.
in HIV prevention: a review of international research. AIDS 2000; 14: 30 Carael M, Slaymaker E, Lyerla R, Sarkar S. Clients of sex workers in
S22–32. different regions of the world: hard to count. Sex Transm Infect 2006;
14 Smith MA, Rissel CE, Richters J, Grulich AE, De Visser R. The 82 (suppl 3): 26–33.
rationale and methods of the Australian Study of Health and
Relationships. Aust N Z J Public Health 2003; 27: 106–16.
17
31 Ayiemba EH. The effect of health education programmes on 48 Ferguson A, Pere M, Morris C, Ngugi E, Moses S. Sexual patterning
adolescent sexual behaviour: a case study of Nairobi city adolescents. and condom use among a group of HIV vulnerable men in Thika, Kenya.
2001; 16: 87–103. Sex Transm Infect 2004; 80: 435–39.
32 Wight D, Henderson M, Raab G, et al. Extent of regretted first 49 Williams BG, Taljaard D, Campbell CM, Gouws E, Ndhlovu L.
intercourse among young teenagers in Scotland: a cross sectional Changing patterns of knowledge, reported behaviour and sexually
survey. BMJ 2000; 44: 1245–44. transmitted infections in a South African gold mining community. AIDS
2003; 17: 2099–107.
33 Dickson N, Paul C, Herbison P, Silva P. First intercourse: age,
coercion and later regrets reported by a birth cohort. BMJ 1998; 316: 50 Leridon H, van Zessen G, Hubert M. The Europeans and their sexual
29–33. partners. In Hubert M, Bajos N, Sanfordt T. Sexual behaviour and HIV/
AIDS in Europe. London: UCL Press, 1998.
34 Genuis SJ, Genuis SK. Adolescent behaviour should be priority
[commentary]. BMJ 2004; 328: 894. 51 Verma RK, Collumbien M. Homosexual activity among rural Indian
men: implications for HIV interventions. AIDS 2004; 18: 1845–47.
35 Harrison A, Cleland J, Gouws E, Frohlich J. Early sexual debut among
young men in rural South Africa: heightened vulnerability to sexual risk? 52 Asthana S, Oostvogels R. The social construction of male
Sex Transm Infect 2005; 81: 259–61. ‘homosexuality’ in India: implications for HIV transmission and
prevention. Soc Sci Med 2001; 52: 707–21.
36 Giesecke J, Scalia Tomba G, Göthberg M. Sexual behaviour related
to the spread of sexually transmitted diseases: a population based 53 Caceres C, Konda K, Pecheny M, Chatterjee A, Lyerla R. Estimating
survey. Int J STD AIDS 1992; 3: 255–60. the number of men who have sex with men in low and middle income
countries. 2006; 82 (suppl 3): 3–9.
37 National Research Council, Institute of Medicine. Panel on transitions
to adulthood in developing countries. In CB Lloyd, ed. Growing up 54 Bajos N, Wadsworth J, Dudcot B, et al. Sexual behaviour in HIV
global: the changing transitions to adulthood in developing countries. epidemiology: Comparative analysis in France and Britain. AIDS 1995;
Washington, DC: The National Academies Press, 2005. 9: 735–43.
38 Mensch BS, Grant MJ, Blanc AK. The changing context of sexual 55 Adrien A, Leaune V, Dassa C, Perron M. Sexual behaviour, condom
initiation in sub-Saharan Africa. Policy Research Division working paper. use and HIV risk situations in the general population of Quebec. Int J
New York: Population Council, 2005. STD AIDS 2001; 12: 108–15.
39 Population Reference Bureau. The world’s youth 2000. Washington 56 Busza J. For love or money: the role of exchange in young people’s
DC: Measure Communication, 2000. sexual relationships. In: Ingham R, Aggleton P, eds. Promoting young
people’s sexual health: international perspectives. London: Routledge,
40 Clark S. Early marriage and HIV risks in sub-Saharan Africa. Stud
2006: 134–52.
Fam Plann 2004; 35: 149–60.
57 Garcia-Moreno C, Jansen HA, Ellsberg M, Herse L, Watts C. WHO
41 Lagarde E, Pison G, Enel C. Knowledge, attitudes and perception of
Multi-country study on women’s health and domestic violence against
AIDS in rural Senegal: relationship to sexual behaviour and behaviour
women. Geneva: World Health Organization, 2005.
change. AIDS 1996; 10: 327–34.
58 Bajos N, Marquet J. Research on HIV sexual risk: social relations-
42 Sangi Haghpeykar H, Poindexter AN, Young A, Levesque JE, Horth F.
based approach in a cross-cultural perspective. Soc Sci Med 2000; 50:
Extra relational sex among Hispanic women and their condom related
1533–46.
behaviours and attitudes. AIDS Care 2003; 15: 505–12.
59 Kordoutis PS, Loumakou M, Sarafidou JO. Heterosexual relationship
43 Cleland J, Ali MM. Sexual abstinence, contraception, and condom
characteristics, condom use and safe sex practices. AIDS Care 2000;
use by young African women: a secondary analysis of survey data.
12: 767–82.
Lancet (in press).
60 Soler H, Quadagno D, Sly DF, Riehman KS, Eberstein IW, Harrison DF.
44 Brown A, Jejeebhoy SJ, Shah I, Yount KM. Sexual relations among
Relationship dynamics, ethnicity and condom use among low-income
young people in developing countries: evidence from WHO case studies:
women. F Plann Perspect 2000; 32: 82–8, 101.
special programme of research, development and research training in
human reproduction. Geneva: World Health Organization, 2001. 61 Pulerwitz J, Amaro H, De Jong W, Gortmaker SL, Rudd R.
Relationship power, condom use and HIV risk among women in the
45 Konings E, Bantebya G, Carael M, Bagenda D, Mertens T. Validating
USA. AIDS Care 2002; 14: 789–800.
population surveys for the measurement of HIV/STD prevention
indicators. AIDS 1995; 9: 375–82. 62 Halton K, Ratcliffe AA, Morison L, West B, Shaw M. Herpes simplex
2 risk among women in a polygynous setting in rural West Africa. AIDS
46 Mwaluko G, Urassa M, Isingo R, Zaba B, Boerma JT. Trends in HIV
2003; 17: 97–103.
and sexual behaviour in a longitudinal study in a rural population in
Tanzania, 1994–2000. AIDS 2003; 17: 2645–51. 63 Slap GB, Lot L, Huang B, Daniyam CA, Zink TM. Sexual behaviour
of adolescents in Nigeria: cross sectional survey of secondary school
47 Kapiga SH, Lugalla JL. Sexual behaviour patterns and condom use
students. BMJ 2003; 326: 15–20.
in Tanzania: results from the 1996 Demographic and Health Survey.
AIDS Care 2002; 14: 455–69. 64 Adimora AA, Schoenbach VJ. Contextual factors and the black-white
disparity in heterosexual HIV transmission. Epidemiology 2002; 13:
707–12.
18
65 Buvé A, Bishikwabo-Nsarhaza K. The spread and effect of HIV-1 in 84 Herbst JH, Sherba RT, Crepaz N, et al. A meta-analytic review of HIV
sub-Saharan Africa. Lancet 2002; 395: 2011–17. behavioral interventions for reducing sexual risk behavior of men who
have sex with men. J Acquir Immune Defic Syndr 2005; 39: 228–41.
66 Blankenship KM, Friedman SR, Dworkin S, Mantell JE. Structural
interventions: concepts, challenges and opportunities for research. J 85 Rees R, Kavanagh J, Burchett H, et al. HIV health promotion and
Urban Health 2006; 83: 59–72. men who have sex with men (MSM): a systematic review of research
relevant to the development and implementation of effective and
67 Dunkle Kl, Jewkes RK, Brown HC, Gray GE, McIntryre JA.
appropriate interventions. London: Research Unit EPPI-Centre, Social
Transactional sex among women in Soweto, South Africa: prevalence,
Science Research Unit, Institute of Education, University of London,
risk factors and association with HIV infection. Soc Sci Med 2004; 59:
2004.
1581–92.
86 Albarracin D, McNatt PS, Klein CT, Ho RM, Mitchell AL, Kumkale
68 Gregson S, Zhuwau T, Anderson RM, Chandiwana SKSSaM. Is there
GT. Persuasive communications to change actions: an analysis of
evidence for behaviour change in response to AIDS in rural Zimbabwe?
behavioral and cognitive impact in HIV prevention. Health Psychol
Soc Sci Med 1998; 46: 321–30.
2003; 22: 166–77.
69 Coffee MP, Garnett GP, Mlilo M, Voeten HA, Chandiwana S, Gregson
87 Elwy AR, Hart GJ, Hawkes S, Petticrew M. Effectiveness of
S. Patterns of movement and risk of HIV infection in rural Zimbabwe. J
interventions to prevent sexually transmitted infections and human
Infect Dis 2005; 191 (suppl 1): S159–67.
immunodeficiency virus in heterosexual men: a systematic review.
70 Furber AS, Newell JN, Lubben MM. A systematic review of current Arch Intern Med 2002; 162: 1818–30.
knowledge of HIV epidemiology and of sexual behaviour in Nepal. Trop
88 Logan TK, Cole J, Leukefeld C. Women, sex, and HIV: social and
Med Int Health 2002; 7: 140–48.
contextual factors, meta-analysis of published interventions, and
71 Agha S. Sexual behaviour among truck drivers in Pakistan. Cult implications for practice and research. Psychol Bull 2002; 128:
Health Sex 2002; 4: 191–206. 851–85.
72 Nyanzi S, Nyanzi B, Kalina B, Pool R. Mobility, sexual networks and 89 Mize SJ, Robinson BE, Bockting WO, Scheltema KE. Meta-analysis
exchange among bodabodamen in southwest Uganda. Cult Health Sex of the effectiveness of HIV prevention interventions for women. AIDS
2004; 6: 239–54. Care 2002; 14: 163–80.
73 Puri MC, Busza J. In forests and factories: sexual behaviour among 90 Neumann MS, Johnson WD, Semaan S, et al. Review and meta-
young migrant workers in Nepal. Cult Health Sex 2004; 6: 145–58. analysis of HIV prevention intervention research for heterosexual adult
populations in the United States. J Acquir Immune Defic Syndr 2002;
74 MacPhail C. Challenging dominant norms of masculinity for HIV
30 (suppl 1): S106–17.
prevention. 2003; 2: 141–49.
91 Rotheram-Borus MJ, Cantwell S, Newman PA. HIV prevention
75 Global Coalition on Women and AIDS: Joint United Nations
programs with heterosexuals. AIDS 2000; 14 (suppl 2):S59–67.
programme on HIV/AIDS. HIV prevention and protection efforts are
failing women and girls. Geneva: The Joint United Nations Programme 92 Shepherd J, Weston R, Peersman G, Napuli IZ. Interventions for
on HIV/AIDS, 2004. encouraging sexual lifestyles and behaviours intended to prevent
cervical cancer. Cochrane Database Syst Rev 2000; 2: CD001035.
76 Seeley J, Grellier R, Barnett T. Gender and HIV AIDS impact
mitigation in sub-Saharan Africa: recognising the constraints. Sahara J 93 Weinhardt LS, Carey MP, Johnson BT, Bickham NL. Effects of HIV
2004; 1: 87–98. counseling and testing on sexual risk behavior: a meta-analytic review
of published research, 1985–1997. Am J Public Health 1999; 89:
77 Tangmunkongvorakul A, Kane R, Wellings K. Gender double
1397–405.
standards in young people attending sexual health services in Northern
Thailand. Cult Health Sex 2005; 7: 361–73. 94 Coyle SL, Needle RH, Normand J. Outreach-based HIV prevention
for injecting drug users: a review of published outcome data. Public
78 Barnett T, Parkhurst J. HIV AIDS: sex, abstinence, and behaviour
Health 1999; 113 (suppl 1): 19–30.
change. Lancet Infect Dis 2005; 5: 590–93.
95 Gibson DR, McCusker J, Chesney M. Effectiveness of psychosocial
79 Marie A. L’Afrique des individus. Paris: Edition Karthala, 1997.
interventions in preventing HIV risk behaviour in injecting drug users.
80 Richens J, Imrie J, Weiss H. Human immunodeficiency virus risk: is AIDS 1998; 12: 919–29.
it possible to dissuade people from having unsafe sex? J R Stat Soc
96 Semaan S, Des Jarlais DC, Sogolow E, et al. A meta-analysis of
Ser A Stat Soc 2003; 166 (part 2): 207–15.
the effect of HIV prevention interventions on the sex behaviors of drug
81 Lesch E, Kruger L. Mothers, daughters and sexual agency in one users in the United States. J Acquir Immune Defic Syndr 2002; 30
low-income South African community. Soc Sci Med 2005; 61: 1072–82. (suppl 1): S73–93.
82 Roberts AB, Oyun C, Batnasan E, Laing L. Exploring the social 97 van Empelen P, Kok G, van Kesteren NM, van den Borne B, Bos
and cultural context of sexual health for young people in Mongolia: AE, Schaalma HP. Effective methods to change sex-risk among drug
implications for health promotion. Soc Sci Med 2005; 60: 1487–98. users: a review of psychosocial interventions. Soc Sci Med 2003; 57:
83 Adimora AA, Schoenbach VJ. Social context, sexual networks, and 1593–608.
racial disparities in rates of sexually transmitted infections. J Infect Dis
2005; 191 (suppl 1): S115–22.
19
98 Kirby D, Laris BA, Rolleri L. Impact of sex and HIV education 115 Bowley DM. Action is still needed to stem heterosexual
programs on sexual behaviors of youth in developing and developed transmission of HIV in Africa letter. BMJ 2002; 324: 1586.
countries: FHI youth research working paper no 2. North Carolina:
116 del Rio C, Sepúlveda J. AIDS in Mexico: lessons learned and
Family Health International (Youth Net Program), 2006.
implications for developing countries. AIDS 2002; 16: 1445–57.
99 Mullen PD, Ramirez G, Strouse D, Hedges LV, Sogolow E. Meta-
117 Hélie A. Holy Hatred. Reprod Health Matters 2004; 12: 120–24.
analysis of the effects of behavioral HIV prevention interventions on the
sexual risk behavior of sexually experienced adolescents in controlled 118 Usdin S, Scheepers E, Goldstein S, Japhet G. Achieving social
studies in the United States. J Acquir Immune Defic Syndr 2002; 30 change on gender-based violence: A report on the impact evaluation of
(suppl 1): S94–105. Soul City’s fourth series. Soc Sci Med 2005; 61: 2434–45.
100 Bertrand JT, O’Reilly K, Denison J, Anhang R, Sweat M. Systematic 119 Halperin DT, Steiner MJ, Cassell MM, Green EC, Hearst N. The time
review of the effectiveness of mass communication programs to has come for common ground on preventing sexual transmission of HIV.
change HIV/AIDS-related behaviors in developing countries. Health Lancet 2004; 364: 1913–15.
Educ Res 2006; 21: 567–97.
120 Schuler S, Hashemi S, Riley A. The influence of women’s changing
101 Grier S, Bryant CA. Social marketing in public health. Annu Rev roles and status in Bangladesh’s fertility transition: evidence from a
Public Health 2005; 26: 319–39. study of credit programmes and contraceptive use. World Development
1997; 25: 563–75.
102 Slater MD. Theory and method in health audience segmentation.
1996; 1: 267–83. 121 Barker G, Nascimento M, Segundo M, Pulerwitz J. How do we
know if men have changed? promoting and measuring attitude change
103 Chapman S. Evaluating social marketing interventions. In:
with young men: lessons from program H in Latin America. Washington:
Thorogood M, Coombes Y, eds. Evaluating health promotion: practice
Oxfam, 2000.
and methods, 2nd edn. Oxford: Oxford University Press, 2004: 93–109.
122 Verma RK, Pulerwitz J, Mahendra VS, Khandekar S, Barker G.
104 Wellings K, Nanchahal K, Macdowall W, et al. Sexual behaviour in
Promoting gender equity among young men to reduce HIV and violence
Britain: early heterosexual experience. Lancet 2001; 358: 1843–50.
risk: positive experiences of Yari dosti in India. Sex Health Exch 2005;
105 Patel V, Andrews G, Pierre T, Kamat N. Gender, sexual abuse and 2: 5–6.
risk behaviours in adolescents: a cross-sectional survey in schools
123 Wilkinson P, French R, Kane R, et al. Teenage conceptions,
in Goa, India. In: Bott S, Jejeebhoy S, Shah I, Puri C, eds. Towards
abortions, and births in England, 1994–2003, and the national teenage
adulthood: exploring the sexual and reproductive health of adolescents
pregnancy strategy. Lancet (in press).
in south Asia. Geneva: World Health Organization, 2003: 99–102.
124 Joint UN Programme on HIV/AIDS. Relationships of HIV and STD
106 Amirkhanian YA, Kelly JA, Kukharsky A, et al. Patterns of HIV risk
declines in Thailand to behavioural change: a synthesis of existing
behavior in a large community sample of Russian men who have sex
studies. Geneva: Joint UN programme on HIV/AIDS, 1998.
with men: an emerging epidemic. AIDS 2001; 15: 407–12.
125 Joint UN Programme on HIV/AIDS. Evaluation of the 100%
107 Anonymous. Warning: teenagers may view non coital sex as a safe
Condom Programme in Thailand. UNAIDS case study. Geneva: Joint UN
option. Contracept Technol Update 2005; 26: 4.
programme on HIV/AIDS, 2000.
108 Donovan B, Ross MW. Preventing HIV: determinants of sexual
126 Nelson KE, Celentano DD, Eiumtrakol S, et al. Changes in sexual
behaviour. Lancet 2000; 355: 1897–901.
behavior and a decline in HIV infection among young men in Thailand. N
109 Ingham R. The importance of context in understanding and seeking Engl J Med 1996; 335: 297–303.
to promote sexual health. In: Ingham R, Aggleton P, eds. Sexual health
127 Hanenberg R, Rojanapithayakorn W. Changes in prostitution and
of young people in developing countries. London: Routledge, 2006:
the AIDS epidemic in Thailand. AIDS Care 1998; 10: 69–79.
155–73.
128 Low-Beer D, Stoneburner RL. Behaviour and communication
110 McIntyre JA. Sex, pregnancy, hormones, and HIV. Lancet 2005;
change in reducing HIV: is Uganda unique? Johannesburg: Centre for
366: 1141–42.
AIDS Development, Research and Evaluation [CADRE], 2004.
111 Amirkhanian YA, Kelly JA, McAuliffe TL. Identifying, recruiting,
129 Ahmed FH, Prada E, Bankole A, Singh S, Wulf D. Reducing
and assessing social networks at high risk for HIV AIDS: methodology,
unintended pregnancy and unsafe abortion in Uganda. Alan Guttmacher
practice, and a case study in St Petersburg, Russia. AIDS Care 2005;
Institute: New York, 2005.
17: 58–75.
130 Alonso A, de Irala J. Strategies in HIV prevention: the A-B-C
112 Kincaid DL. Social networks, ideation and contraceptive behaviour
approach. Lancet 2004; 364: 1033.
in Bangladesh: a longitudinal analysis. Soc Sci Med 2000; 50: 215–31.
131 Rifkin W. Strategies in HIV prevention: the A-B-C approach Lancet
113 Dandona R, Dandona L, Gutierrez JP, et al. High risk of HIV in non-
2004; 364: 1033.
brothel based female sex workers in India. BMC Public Health 2005; 5:
87. 132 Allen T, Heald S. HIV/AIDS policy in Africa: What has worked in
Uganda and what has failed in Botswana? J Int Dev 2004; 16: 1141–
114 Singh CM, Verma BL. Sexual behaviour and condom usage in some
54.
high risk groups of district Jhansi (U.P). Indian J Prev Soc Med 2004;
35: 179–83.
20
133 Parkhurst JO, Lush L. The political environment of HIV: lessons 152 Kerrigan D, Moreno L, Rosario S, et al. Environmental-structural
from a comparison of Uganda and South Africa. Soc Sci Med 2004; 59: interventions to reduce HIV/STI risk among female sex workers in the
1913–24. Dominican republic. J Public Health 2006; 96: 120–25.
134 Macintyre K, Brown L, Sosler S. It’s not what you know but who 153 Lambert H. Accounting for EBM: notions of evidence in medicine.
you knew: examining the relationship between behavior change and Soc Sci Med 2006; 62: 2633–45.
decline in HIV infection in Africa. AIDS Educ Prev 2001; 13: 160–74.
154 Lambert H, Gordon EJ, Bogdan-Lovis EA. Introduction: Gift horse
135 Powles J, Day N. Syndromic STI and behaviour-change or Trojan horse? Social science perspectives on evidence-based health
interventions in Uganda. Lancet 2003; 361: 2086. care. Soc Sci Med 2006; 62: 2613–20.
136 Korenromp EL, Bakker R, de Vlas SJ, Gray RH, Wawer MJ. HIV 155 Coombes Y, Thorogood M. Introduction. In: Thorogood M, Coombes
dynamics and behaviour change as determinants of the impact of Y, eds. Evaluating health promotion: practice and methods, 2nd ed.
sexually transmitted disease treatment on HIV transmission in the Oxford: Oxford University Press, 2004.
context of the Rakai trial. AIDS 2002; 16: 2209–18.
156 Ellis S, Grey A. Prevention of sexually transmitted infections: a
137 Parkhurst JO. The Ugandan success story? Evidence and claims of review of reviews into the effectiveness of non-clinical interventions.
HIV-1 prevention. Lancet 2002; 360: 78–80. London: Health Development Agency, 2004.
138 Darroch JE, Singh S. Why is teenage pregnancy declining? The 157 Ferry B, Carael M, Buve A, Auvert B, Laourou M. Comparison of
roles of abstinence, sexual activity, and contraceptive use. New York: key parameters of sexual behaviour in four African urban populations
Alan Guttmacher Institute, 1999. with different levels of HIV infection: The multicentre study of factors
determining the different prevalences of HIV in sub-Saharan Africa.
139 Finer LB, Henshaw SK. Disparities in rates of unintended
AIDS 2001; 15 (suppl 4): S41–50.
pregnancy in the United States 1994 and 2001. Persp on Sexual and
Repr Health 2006; 38: 90–96. 158 Michael RT, Wadsworth J, Feinleib J, Johnson AM, Laumann
E, Wellings K. Private sexual behaviour, public opinion, and public
140 Mohn JK, Tingle LR, Finger R. An analysis of the causes of the
health policy related to sexually transmitted diseases: A US British
decline in non-marital birth and pregnancy rates for teens from 1991
comparison. Am J Public Health 1998; 88: 749–54.
to 1995. Adolesc Fam Health 1995; 3: 39–47.
159 Quigley MA, Kamali A, Kinsman J, Kamulegeya I, Nakiyingi-Miiro J.
141 Jones JM, Toffler W, Reed B. The declines in adolescent pregnancy,
The impact of attending a behavioural intervention on HIV incidence in
birth and abortion rates in the 1990s: what factors are responsible? In:
Masaka, Uganda. AIDS 2004; 18: 2055–63.
Fanwood, NJ: Consortium of state Physicians Resource Councils, 1999.
160 Kim J, Watts C, Hargreaves J, et al. Women’s economic
142 Saul R. Teen pregnancy: progress meets politics: The Guttmacher
empowerment can reduce gender-based violence: results from
Review 1999; 2: 6–9.
the IMAGE study, a cluster randomized trial in South Africa. XVI
143 Kauffman RB, Spitz AM, Strauss LT, et al. The decline in US teen International AIDS Conference, Toronto, 13–18 Aug, 2006 (abstr
pregnancy rates: 1990–1995. Pediatrics 1998; 102: 1141–47. THPE0683).
144 US National Institutes of Health. Interventions to prevent HIV risk 161 Bonell C, Hargreaves J, Strange V, Pronyk P, Porter J. Should
behaviors: NIH consensus statement. 1997; 15: 15–16. structural interventions be evaluated using RCTs? The case of HIV
prevention. Soc Sci Med 2006; 63: 1135–42.
145 Santelli JS, Abma J, Ventura S, et al. Can changes in sexual
behaviors among high school students explain the decline in teen
pregnancy rates in the 1990s? J Adolesc Health 2004; 35: 80–90.
146 Santelli JS, Lindberg LD, Singh S, Finer LB. Recent declines in
adolescent pregnancy in the United States: more abstinence or better
contraceptive use? Am J Public Health (in press).
147 Donovan P. Falling teen pregnancy, birth rates: what’s behind the
decline? The Guttmacher Review 1998; 1: 6–9.
148 Agyei-Mensah S. The HIV/AIDS epidemic in sub Saharan Africa:
homogeneity or heterogeneity? Nor Geogr Tidsskr 2005; 59: 14–25.
149 Okie S. Fighting HIV: lessons from Brazil. N Engl J Med 2006; 354:
1977–81.
For more information, please contact:
150 Wellings K. Evaluating AIDS public education in Europe: a cross- Department of Reproductive Health and Research
national comparison. In: Hornik R, ed. Public health communication: World Health Organization
evidence for behavior change. London: Lawrence Erlbaum Associates, Avenue Appia 20, CH-1211 Geneva 27
2002: 131–46. Switzerland
Fax: +41 22 791 4171
151 Kim JC, Watts CH. Gaining a foothold: tackling poverty, gender E-mail: reproductivehealth@who.int
inequality, and HIV in Africa. BMJ 2005; 331: 769–72. www.who.int/reproductive-health
21