This paper assesses the extent to which HIV prevention interventions for young people in sub-Saharan Africa are grounded in theory and if theory-based interventions are more effective. Three databases were searched for evaluation studies of HIV prevention interventions for youth. Additional articles were identified on websites of international organisations and through searching references. 34 interventions were included; 25 mentioned the use of theory. Social Cognitive Theory was most prominent (
With an estimated 2.7 million new infections worldwide in 2010, HIV incidence remains at very high levels [
Theory is said to be an essential component of successful health promotion interventions [
In health promotion research, a large number of theories coexist that aim to understand health-related behaviour and provide tools for behaviour change. The Social Learning/Cognitive Theory (SCT), Theory of Reasoned Action/Planned Behaviour (TRA/TPB), and Health Belief Model (HBM) are the most dominant theories, more recently joined by the Stages of Change (SoC) and Social Ecological Model (SEM) [
The SCT posits that people acquire and maintain particular behavioural patterns through a constant interaction between three factors: environment, personal factors, and behaviour [
These dominant theories work at various levels and for different purposes. While the HBM and TRA/TPB are explanatory theories operating at the individual level, the SCT and SEM include the interpersonal and environmental levels, respectively. The SoC theory, in turn, is a change theory, not explaining a particular behaviour, but providing a framework for how people alter their behaviour.
With the overarching objective of improving effectiveness of HIV prevention interventions that target young people’s sexual behaviour in sub-Saharan Africa, this paper examines the extent to which these interventions are grounded in theory, how these theories are applied and assesses if theory-based interventions are more effective in modifying sexual behaviour than interventions not explicitly grounded in theory.
We performed a systematic review to locate evaluated interventions that aim to reduce sexual risk behaviour of young people in sub-Saharan Africa. Studies were considered eligible if they reported on the evaluation of an HIV prevention intervention for young people on the subcontinent, had a control group, and were published between January 1990 and March 2012. Further, to be included, studies had to report on the general population of young people (10–25 years) and the intervention needed to aim to prevent HIV transmission by reducing sexual risk taking. Searches were performed in the online databases Medline (PubMed interface), ISI Web of Science, and EBSCOhost. Additional articles were identified on websites of international organisations and through searching references of eligible articles. Data extraction was then done in duplicate by five investigators using a predesigned and pretested extraction sheet. Further details of the search terms, study eligibility, and data extraction are detailed elsewhere [
We extracted data on characteristics of the interventions and theory use. Firstly, whether any theory had been used and, if so, which. Secondly, for what purpose the theory was used. We extracted full-text descriptions of how the theory had been used, which was later recoded into three categories: theory used to inform the intervention (e.g., for curriculum development); theory used to guide evaluation (e.g., to develop indicators); or both. Thirdly, a binary variable was derived, capturing whether an explanation was provided about why this theory was chosen. For the studies not reporting the use of a theory, we looked at the topics dealt with in the interventions and the envisaged interventions’ outcomes. This gives us an indication of the underlying theoretical assumptions used in these interventions.
Data were also extracted on the behavioural outcomes of the interventions: condom use (at last sex; consistency and intention), sexual behaviour (primary abstinence; the proportion of sexually active youth; recent sexual intercourse; number of sexual partners and multiple partnerships), and biological outcomes (HIV/STI incidence).
1073 article titles and/or abstract were screened. After analysis of title and abstract, we reviewed 73 full-text publications. In total, evaluations of 34 studies met the inclusion criteria, reported on in 38 articles. Table
Characteristics of studies included in systematic review of use of behavioural theory in HIV prevention interventions in youth in sub-Saharan Africa.
Author, year | Country | Year of intervention | Study design | Sample size at baseline (males/females) | Main intervention activities (duration) | Intervention setting (urban/rural) | Theory or theories used | Role of theory in the study | Explanation provided about why theory used? |
---|---|---|---|---|---|---|---|---|---|
Central Africa | |||||||||
| |||||||||
Van Rossem and Meekers [ |
Cameroon | 1996-1997 | Repeat C/S, quasiexperimental | 1606 (753/757) | Behaviour change communication and promotion through peers and in media, condom distribution, youth-friendly services (13 months) | Community (urban) | Health Belief Model | Development of intervention and questionnaire/evaluation | No |
| |||||||||
Speizer et al. [ |
Cameroon | 1997-1998 | Repeat C/S, quasiexperimental | 802 (400/402) | Through discussion groups, one-on-one meetings, and health and sport association gatherings, peer educators informed their peers and referred them to services. Promotional materials were distributed in schools and community (18 months) | School + Community (urban) | NR, focus on knowledge | NA | NA |
| |||||||||
Meekers et al. [ |
Cameroon | 2000-2001 | Repeat C/S, pre post-controlling for exposure | 1956 (1056/900) | Media and interpersonal communication campaign. Peer education, magazine, radio drama, radio call-in show, media campaign, condom promotion (12 months) | Community (urban) | Health Belief Model, Social Learning Theory, Theory of Reasoned Action | Development of intervention and questionnaire/evaluation | No |
| |||||||||
Eastern Africa | |||||||||
| |||||||||
Klepp et al. [ |
Tanzania | 1990 | Cohort, randomized schools | 1063 (502/561) | Teachers provided information, students created posters and performed songs, poetry, drama and role-play, small-group discussions among students. Interviews and panel discussions with parents and community members (2-3 months) | Primary school (urban + rural) | Social Learning Theory and Theory of Reasoned Action | Development of intervention and questionnaire/evaluation | No |
| |||||||||
Shuey et al. [ |
Uganda | 1994–1996 | Repeat C/S, quasiexperimental | 800 (398/402) | Strengthen existing school health curriculum, meeting with parents and community leaders, formation of school health clubs with peer education, question boxes (2 years) | Primary school (urban + rural) | Social Cognitive Theory | NR | No |
| |||||||||
Kinsman et al. [ |
Uganda | 1997-1998 | Cohort, quasiexperimental | 2077 (920/1157) | Extracurricular classes by trained teachers (1 year) | Primary and secondary schools (rural) | Behaviour Changes for Interventions Model | Development of intervention and questionnaire/evaluation | No |
| |||||||||
Erulkar et al. [ |
Kenya | 1998–2000 | Repeat C/S, quasiexperimental | 1544 (792/752) | Adult counsellor in community educating youth, referral to youth-friendly services and encouraging parent-child communication (3 years) | Community (urban + rural) | NR, focus on values, knowledge, gender, and empowerment | NA | NA |
| |||||||||
Ross et al. [ |
Tanzania | 1998–2002 | Repeat C/S, randomized communities | Ross: 9219 (5103/4116) |
Participatory, teacher-led, peer-assisted, in-school program, youth-friendly health services, condom promotion and distribution, and youth health days and video shows in community (3 years) | School + Community (rural) | Social Learning Theory | Development of intervention | No |
| |||||||||
Maticka-Tyndale et al. [ |
Kenya | 2002-2003 | Repeat C/S, randomized schools | 7392 (3636/3764) | Peer education on level of teachers and students, question boxes, school health clubs, information corners and assemblies, drama, music and literary performances (18 months) | Primary school (urban + rural) | Social Learning Theory and Scripting Theory | Development of intervention | Yes |
| |||||||||
Rijsdijk et al. [ |
Uganda | 2008 | Cohort, randomized schools | 1986 (1096/889) | low-tech, computer-based, interactive comprehensive sex education programme, teacher-led (6 months) | Secondary school (urban + rural) | Theory of Planned Behavior and Health Belief Model | Development of intervention | No |
| |||||||||
Southern Africa | |||||||||
| |||||||||
Kuhn et al. [ |
South Africa | 1990 | Repeat C/S, quasiexperimental | 567 (not reported) | Intense, high-profile focus on AIDS in the school by teachers (2 weeks) | Secondary school (urban) | NR, focus on knowledge and attitudes | NA | NA |
| |||||||||
Harvey et al. [ |
South Africa | 1993-1994 | Cohort, randomized schools | 1080 (447/633) | “School open day” with drama, song, dance, poetry, and posters prepared and presented by students (3 days) | Secondary school (urban + rural) | Applied behaviour change framework | Development of questionnaire/evaluation | No |
| |||||||||
Meekers [ |
South Africa | 1994–1997 | Repeat C/S, quasiexperimental | 226 (0/226) | Mass media campaign, peer education and condom promotion and distribution (35 months) | Community (urban) | Health Belief Model | Development of questionnaire/evaluation | No |
| |||||||||
Fitzgerald et al.; Stanton et al. [ |
Namibia | 1996 | Cohort, randomized participants | 515 (236/279) | Curriculum taught by a teacher and out-of-school youth (7 weeks) | Secondary school (urban + rural) | Social Cognitive Theory/Protective Motivational Theory | Development of intervention and questionnaire/evaluation | No |
| |||||||||
Kim et al. [ |
Zimbabwe | 1997-1998 | Repeat C/S, quasiexperimental | 1426 (713/713) | Mass media campaign, community drama groups, peer educators, youth-friendly health services (6 months) | School + Community (urban) | Steps to Behaviour Change Framework | Development of intervention | No |
| |||||||||
James et al. [ |
South Africa | 1998 | Cohort, randomized schools | 1168 (542/616) | Reading of a comic book (1 hour) | Secondary school (urban + rural) | Theory of Health Promotion and Social Learning | Development of intervention | No |
| |||||||||
Visser [ |
South Africa | 1998–2000 | Repeat C/S, pre post-controlling for exposure | 873 (410/463) | Trained teachers and professionals provide life skills and HIV/AIDS education. Parents included in action committee (1 year) | Secondary school (urban) | Health Belief Model | Development of intervention | No |
| |||||||||
Underwood et al. [ |
Zambia | 1999-2000 | Repeat C/S, quasiexperimental | 921 (378/543) | Participatory developed mass media campaign (7 months) | Community (urban + rural) | Stage Theory of Behaviour Change | Development of intervention | No |
| |||||||||
Magnani et al. [ |
South Africa | 1999–2001 | Cohort, pre post-controlling for exposure | 3052 (1375/1677) | Life skills curriculum taught by teachers (2 years) | Secondary school (urban) | Social Learning Theory | Development of intervention | No |
| |||||||||
Agha [ |
Zambia | 2000 | Cohort, randomized schools | 481 (268/213) | Peer educators using discussion and drama skits (1 hour 45 min) | Secondary school (urban) | NR, focus on knowledge, normative beliefs, and risk perception | NA | NA |
| |||||||||
James et al. [ |
South Africa | 2001 | Cohort, randomized schools | 936 (456/466) | Life skills intervention taught by trained teachers (20 weeks) | Secondary school (urban + rural) | Social Cognitive Theory and Theory of Planned Behaviour | Development of questionnaire/evaluation | No |
| |||||||||
Plautz et al. [ |
Madagascar | 2001-2002 | Cohort, pre post-controlling for exposure | 1785 (1000/785) | Youth-friendly services, mass media, and interpersonal communication by peer educators (23 months) | Community (urban + rural) | Social Learning Theory, Health Belief Model, and Theory of Reasoned Action | Development of intervention and questionnaire/evaluation | No |
| |||||||||
Karnell et al. [ |
South Africa | 2002 | Cohort, randomized schools | 661 (324/337) | Peer educators using recorded monologues of fictional characters, teacher support (8 weeks) | Secondary school (urban) | Social Learning Theory, Social Inoculation, Cognitive Behaviour Theory | Development of intervention and questionnaire/evaluation | Yes |
| |||||||||
Visser [ |
South Africa | 2002-2003 | Repeat C/S, quasiexperimental | 1918 (858/1060) | Peer education (18 months) | Secondary schools (urban) | Systems Theory | Development of intervention | No |
| |||||||||
Jewkes et al. [ |
South Africa | 2003-2004 | Cohort, randomized communities | 2776 (1360/1416) | Participatory learning approaches taught by facilitators, peer group meeting, community meeting (6–8 weeks) | Community (rural) | Participatory Learning Approach and Adult Education Theory | Development of intervention | No |
| |||||||||
Tibbits et al. [ |
South Africa | 2004-2005 | Cohort, randomized schools | 4040 (2020/2020) | Comprehensive, risk-reduction life skills curriculum for adolescents, teacher-led (24 months) | Secondary school, urban | Selective optimization with compensation, Self-Determination Theory, and Social Cognitive Theory | Development of intervention | Yes |
| |||||||||
Mason-Jones et al. [ |
South Africa | 2007-2008 | Cohort, quasiexperimental | 3934 (1661/2211) | Government-led peer education project, in class standard curriculum, conversations outside class, referral (18 months) | Secondary school (urban + rural) | NR, knowledge and psychosocial characteristics | NA | NA |
| |||||||||
Baird et al. [ |
Malawi | 2008-2009 | Cohort, randomized schools | 3796 (0/3796) | Monthly cash transfer programme to reduce the risk of STI infection (24 months) | School + community (urban + rural) | NR, focus on structural factor (poverty and education) and knowledge | NA | NA |
| |||||||||
Burnett et al. [ |
Swaziland | NR | Cohort, randomized youth | 204 (101/103) | Teacher-led life-skills HIV prevention education program, curriculum, interactive techniques, role playing, and group discussions (13 weeks) | Secondary school (urban) | Self-efficacy theory and Protection Motivation Theory | Development of intervention | No |
| |||||||||
Western Africa | |||||||||
| |||||||||
Brieger et al. [ |
Nigeria and Ghana | 1994–1997 | Repeat C/S, quasiexperimental | 1784 (not reported) | Peer educators, promotion of community-level networks, referral to services (30 months) | School + Community (urban) | NR, focus on knowledge and attitudes | NA | NA |
| |||||||||
Fawole et al. [ |
Nigeria | 1996 | Cohort, pre post-controlling for exposure | 450 (204/246) | Education sessions by community physicians with help of teachers (1 month) | Secondary school (urban) | NR, focus on knowledge and attitudes | NA | NA |
| |||||||||
Okonofua et al. [ |
Nigeria | 1997-1998 | Repeat C/S, randomized schools | 1896 (877/1008) | Establishment of reproductive health club in school, health awareness campaigns by professionals, distribution of print material, peer education, youth-friendly services (11 months) | Secondary school (urban) | NR, focus on knowledge and barriers | NA | NA |
| |||||||||
Van Rossem and Meekers [ |
Guinea | 1997-1998 | Cohort, quasiexperimental | 2016 (925/1091) | Peer educators (discussion and theatre), condom promotion, billboards, youth-friendly services and contraception distribution (8 months) | Community (urban) | Health Belief Model | Development of questionnaire/evaluation | No |
| |||||||||
Atwood et al. [ |
Liberia | 2007-2008 | Cohort, randomized schools | 812 (455/357) | Curriculum-based program by health educators (8 weeks) | Primary school (urban) | Social Cognitive Theory and Theory of Reasoned Action | Development of intervention | No |
C/S: Repeated cross-sectional design.
NR: No theory is explicitly reported, dominant constructs used in the intervention.
NA: Not applicable.
About three quarters of the studies—25 of 34—mentioned having used at least one theory. In total, 19 different theories were mentioned 42 times. Several stated that they had applied two or more theories, with three papers reporting that the intervention design drew on three theories.
Of all the theories mentioned, the SCT was most prominent (
Description of the main activities indicates that most interventions use one or a combination of participatory learning techniques, such as drama plays, poetry, songs, club formation, peer education (role modelling), and discussions and debates. This suggests that the learning strategies of most interventions were based on participatory learning approaches.
A small, but considerable proportion of interventions [
There is no clear evolution detectable over time in the frequency of use of different theories; of the 20 studies which began in the decade 1990–1999, 14 reported theory use, while 11 of the 14 beginning after 2000 used theory.
Of 25 interventions that mentioned a theory, 7 said that the theory was used to both inform the content of the intervention (e.g., the curriculum) and to inform the evaluation or questionnaire design. In 13 studies, theory was reportedly used only to inform the intervention content, and in 4 only for designing the evaluation or questionnaire. One study mentioning theory use did not specify how this was applied [
Only three articles provided information on why a particular theory was selected [
Overall, the behavioural outcomes of the 34 studies were markedly heterogeneous, with little reduction in heterogeneity after stratifying by theory use (Table
Description of study outcomes stratified by role of theory use in each study.
Condom use at last sex | Ever/consistently used condom | Sexual debut, proportion of sexually active youth | Sexual intercourse in past months | Number of sexual partners | HIV incidence | HSV-2 | Other STIs | Overview | |
---|---|---|---|---|---|---|---|---|---|
Theory used for development of intervention | Visser, 2005° | Ross, 2007++ | Visser, 2005−− | Maticka-Tyndale, 2007° | Visser, 2005° | Ross, 2007° | Ross, 2007° | Ross, 2007° | 44 outcomes |
Magnani, 2005++ | Underwood, 2006++ | Magnani, 2005−− | Visser, 2007++ | Magnani, 2005++ | Jewkes, 2008° | Jewkes, 2008++ | Jewkes, 2008° | ++7 | |
Maticka-Tyndale, 2007+ | Kim, 2001° | Maticka-Tyndale, 2007+ | Kim, 2001+ | Ross, 2007+ | Doyle, 2010° | Doyle, 2010° | Doyle, 2010° | +10 | |
Ross, 2007+ | Magnani, 2005++ | Ross, 2007° | Visser, 2007−− | °22 | |||||
Jewkes, 2008° | Atwood 2012° | Klepp, 1997+ | Kim, 2001+ | −1 | |||||
Underwood, 2006° | Kim, 2001+ | Atwood, 2012−− | −−4 | ||||||
Visser, 2007° | Underwood, 2006− | Doyle, 2010+ | |||||||
Doyle, 2010° | Atwood, 2012+ | ||||||||
Tibbits, 2011° | Burnet, 2011° | ||||||||
Doyle, 2010° | |||||||||
Tibbits, 2011° | |||||||||
| |||||||||
James, 2006+ | Harvey, 2000++ | Harvey, 2000° | James, 2006+ | Harvey, 2000° | 12 outcomes | ||||
Meekers, 1998° | Meekers, 1998° | Van Rossem, 1999+ | Van Rossem, 1999° | ++1 | |||||
Theory used for development of evaluation or questionnaire | Van Rossem, 1999+ | Van Rossem, 1999+ | Meekers, 1998° | +5 | |||||
°6 | |||||||||
−0 | |||||||||
−−0 | |||||||||
| |||||||||
Van Rossem, 2000° | Van Rossem, 2000++ | Fitzgerald, 1999° | Fitzgerald, 1999° | Fitzgerald, 1999° | 15 outcomes | ||||
Fitzgerald, 1999° | Meekers, 2005+ | Van Rossem, 2000° | Van Rossem, 2000+ | ++1 | |||||
Theory used for development of intervention and evaluation or questionnaire | Meekers, 2005+ | Fitzgerald, 1999° | Plautz, 2003− | +3 | |||||
Plautz, 2003° | Plautz, 2003° | °10 | |||||||
Karnell, 2005° | −1 | ||||||||
−−0 | |||||||||
| |||||||||
Theory used, but uncertain in which phase of study | Shuey++ | 1 outcome | |||||||
++1 | |||||||||
| |||||||||
Speizer+ | Agha, 2002° | Speizer++ | Speizer−− | Agha, 2002++ | Baird, 2012++ | Baird, 2012++ | Baird, 2012° | 18 outcomes | |
Agha, 2002° | Okonofua++ | Brieger, 2001−− | Agha, 2002++ | Fawole, 1999° | ++6 | ||||
Interventions not explicitly based on theory | Erulkar, 2004+ | Kuhn° | Fawole, 1999° | Erulkar, 2004° | Erulkar, 2004+ | +4 | |||
Fawole, 1999° | Fawole, 1999° | Erulkar, 2004+ | °12 | ||||||
Mason-Jones, 2011° | Baird, 2012° | Baird, 2012° | −0 | ||||||
Mason-Jones, 2011−− | −−3 |
++significant positive intervention effect on outcome variable for the whole study population.
+significant positive intervention impact on outcome variable for a subgroup of the target population, and no significant impact on the whole study population or whole population impact not reported.
°no significant intervention impact on the outcome variable.
−significant negative intervention effect on outcome variable in a sub-group of the target population, and no significant impact on the whole study population or whole population impact not reported.
−−significant negative intervention effect on the outcome variable for the whole study population.
Four studies reported biological measures of intervention effectiveness [
Four studies refer to their theoretical basis in their conclusions, criticizing the theory, specifically “the theoretical approaches underlying the program have built in shortcomings which could result in the program not having significant impact on the students' behavioural intentions” [
The review found that the majority of HIV prevention interventions targeted at youth in sub-Saharan Africa use theory-based approaches. A wide range of theories have been employed, but three behavioural theories predominate: SCT, HBM, and TRA/TPB. No one theory emerged dominant, as reporting on the choice, use, and specific evaluation of theory was low.
Broadly, the results are consistent with reviews of HIV risk-reduction interventions elsewhere, though some variation in use of theory can be noted across these reviews. Pedlow and Carey [
By focussing on cognitive constructs of behaviour, the interventions explicitly or implicitly start from the assumption that cognitions influence the person's thinking and decision making, and thus drive sexual behaviour [
Firstly, cognitive behavioural models aim to explain a particular behaviour. The theoretical constructs that influence behavioural decisions may vary, depending on the behaviour in question. This poses marked challenges for HIV prevention interventions, since they generally attempt to influence a wide range of behaviours—for example, increasing condoms use, reducing the number of sexual partners, minimising sexual activity, delaying the onset of sexual debut—which are influenced by different factors. To further complicate matters, sexual decisions may vary depending on the reasons for sexual intercourse (ranging between, e.g., intimacy or desire, external factors, and affect management). These, in turn, are further influenced by gender and psychological characteristics (e.g., depression, self-esteem, and impulsiveness) [
Second, the applicability of cognitive behavioural models to youth sexual behaviours may vary between development stages. For instance, applying cognitive theory to young people with no or limited sexual experience may be difficult. This group may not yet have well-anchored ideas, and consequently their attitudes, norms, and beliefs about safe sexual behaviour may be less clear and stable than for their adult counterparts [
Third, these theories seem to ignore the fact that sexual intercourse takes place between two persons, within a relationship. Sexual decisions do not depend on the individual, but also on the sexual partner and the type of relationship. Young people might have specific types of partners that may influence sexual decision making. For example, relationships with someone who is much older are risky because it exposes the younger person (mostly girls) to a partner who is more likely to be sexually experienced and hence more likely to be HIV-positive [
Fourth, while some interventions recognize the importance of involving the community, only the SCT explicitly stresses the influence of contextual and structural factors on an individual’s behaviour. Even though the theory states that the social environment is an important determinant of the behaviour, many interventions based on SCT did not attempt to include or influence environmental factors. Most interventions are limited to providing information and teaching skills. TRA/TPB implicitly includes this level by stating that personal attitudes, and norms are influenced by behavioural and normative beliefs in the society, which is useful for tracking varying modes of sexual socialisation. However, this is an indirect effect of the environment on individual behaviour, while still ignoring the broader structural factors that shape sexual behaviour. Many recent studies have demonstrated the contribution of structural factors to young people’s vulnerability for HIV [
According to Gielen and Sleet [
Finally, while cognitive behavioural theories of change might be successful in altering cognitions and behavioural intentions, they provide insufficient directives on translating this into actual behaviour change. Thus interventions could be regarded as successful in having altered motivations and intentions, even though behavioural change may not result. Similarly, interventions based on the HBM, might increase the perceived severity and susceptibility of a person, and relieve barriers to behaviour change, but in itself, might be insufficient to alter the sexual behaviour. Clearly, motivations or beliefs about behaviour change on a cognitive or rational level need to be accompanied by a clear strategy for introducing a new behaviour [
In the end, it boils down to two key questions: what determines sexual behaviour of young people? And what frameworks are most useful for making sense of and impacting positively on determinants of youth sexual behaviour? Recognizing the complexity and heterogeneity of this particular behaviour, theory can provide help in generalizing key determinants and making them operational. Theories aim to describe determinants and processes that account for or guide behaviour (change) through the rationalization of individual decisions. This aids in understanding human behaviour, and when used appropriately, can provide a solid grounding for program development and evaluation. The strength of theory is to generalize and simplify complex situations. However, in the case of HIV prevention interventions for young people, the dominant theories might oversimplify sexual behaviour. While such cognitive behavioural models can explain the links between intention and behaviour, particularly at an intrapersonal level, they are less able to account for interpersonal and contextual factors related to the complexity of sex, the experience of youth and disparities in social, cultural, and economic realities of youth in sub-Saharan Africa.
K. Michielsen acknowledges the Research Foundation Flanders (FWO) for financial support. The authors acknowledge Stanley Luchters and Petra De Koker for data extraction.