Share this article on:

Biological and behavioural impact of an adolescent sexual health intervention in Tanzania: a community-randomized trial

Ross, David Aa,b,c; Changalucha, Johnb; Obasi, Angela INa,c; Todd, Jima,b,c; Plummer, Mary La,b,c; Cleophas-Mazige, Bernadettec; Anemona, Alessandraa,b,c; Everett, Deana,b,c; Weiss, Helen Aa; Mabey, David Ca; Grosskurth, Heinera; Hayes, Richard Ja,*

doi: 10.1097/QAD.0b013e3282ed3cf5
Epidemiology and Social

Objective: The impact of a multicomponent intervention programme on the sexual health of adolescents was assessed in rural Tanzania.

Design: A community-randomized trial.

Methods: Twenty communities were randomly allocated to receive either a specially designed programme of interventions (intervention group) or standard activities (comparison group). The intervention had four components: community activities; teacher-led, peer-assisted sexual health education in years 5–7 of primary school; training and supervision of health workers to provide ‘youth-friendly’ sexual health services; and peer condom social marketing. Impacts on HIV incidence, herpes simplex virus 2 (HSV2) and other sexual health outcomes were evaluated over approximately 3 years in 9645 adolescents recruited in late 1998 before entering years 5, 6 or 7 of primary school.

Results: The intervention had a significant impact on knowledge and reported attitudes, reported sexually transmitted infection symptoms, and several behavioural outcomes. Only five HIV seroconversions occurred in boys, whereas in girls the adjusted rate ratio (intervention versus comparison) was 0.75 [95% confidence interval (CI) 0.34, 1.66]. Overall HSV2 prevalences at follow-up were 11.9% in male and 21.1% in female participants, with adjusted prevalence ratios of 0.92 (CI 0.69, 1.22) and 1.05 (CI 0.83, 1.32), respectively. There was no consistent beneficial or adverse impact on other biological outcomes. The beneficial impact on knowledge and reported attitudes was confirmed by results of a school examination in a separate group of students in mid-2002.

Conclusion: The intervention substantially improved knowledge, reported attitudes and some reported sexual behaviours, especially in boys, but had no consistent impact on biological outcomes within the 3-year trial period.

From the aLondon School of Hygiene and Tropical Medicine, London, UK

bNational Institute for Medical Research, Mwanza Centre, Mwanza, Tanzania

cAfrican Medical and Research Foundation (AMREF), Mwanza, Tanzania.

*See Appendix for additional co-authors.

Received 10 December, 2005

Revised 25 December, 2006

Accepted 29 February, 2007

Correspondence to David A. Ross, Infectious Disease Epidemiology Unit, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK. E-mail:

Back to Top | Article Outline


Youth (15–24 years) in sub-Saharan Africa account for almost one quarter of those living with HIV worldwide [1–3], and several studies have demonstrated high rates of sexually transmitted infections (STI), pregnancy, and their complications [4,5]. In the absence of a vaccine or cure, preventive behavioural interventions have been advocated as the most effective HIV control strategy, especially among youth [1,6]. Even in the worst affected countries, prevalences of HIV and STI are very low in 15 year olds, but rise steeply thereafter [3,7,8]. Interventions focusing on adolescents might therefore have a substantial impact on the HIV epidemic, but evidence on the efficacy of behavioural interventions is contradictory [9–17]. Whereas most programme evaluations in developing countries have shown an improvement in knowledge, reported communication about sexual matters and reported attitudes, most showed no changes in reported sexual behaviours or changes that were demonstrated only in subgroups [16].

Few previous trials have measured biomedical endpoints, and this is the first to measure the impact on HIV as well as other STI and pregnancy. The inclusion of such outcomes is critically important because of known limitations in the validity of reported sexual behaviour, particularly in young people [18–23]; the potential for interventions to increase bias in reported behaviour towards more socially desirable behaviours; and because reductions in HIV, STI and pregnancy are usually the ultimate objectives for these interventions.

We conducted a community-randomized trial to evaluate the impact of the MEMA kwa Vijana (‘Good things for young people’) intervention on HIV incidence, the prevalence or incidence of other STI and pregnancy, and sexual health knowledge, attitudes and reported sexual behaviour.

Back to Top | Article Outline


The design of the trial [24] and of the intervention [25] are described in detail elsewhere.

Back to Top | Article Outline

Study population

The trial was conducted in 20 well-separated rural communities in Mwanza Region, Tanzania, (Fig. 1). The study communities were grouped into three risk strata using data from a previous population-based survey [5]. Restricted randomization was used to balance HIV and chlamydia prevalence between the two trial arms [24]. Ten communities (58 primary schools, 18 health facilities) received the intervention, the other 10 (63 primary schools, 21 health facilities) acting as comparison communities.

Back to Top | Article Outline

Intervention design

To ensure sustainability and replicability, the intervention was delivered by government workers through existing structures and supervision systems, who were trained and supported by eight staff from the African Medical and Research Foundation. The main aims of the intervention were to provide young people with the knowledge and skills to enable them to delay sexual debut, reduce sexual risk-taking by sexually active youth (including reducing numbers of sexual partners and promoting condom use) and increase their appropriate use of sexual health services (e.g. STI treatment, family planning). It had four major components [25].

The most intensive component was a participatory, teacher-led, peer-assisted, in-school programme, comprising an average of 12 40-min sessions per year, held in normal school hours in years 5–7 of primary school. The programme built on the experience of adolescent reproductive health projects in Tanzania [26–28] and elsewhere [29–31], and aimed to include all 10 characteristics previously identified as associated with effective programmes [11].

Second, two to four health workers per government facility were trained for one week in the provision of youth-friendly sexual and reproductive health services, and were supervised quarterly. This was in addition to the provision of family planning services and improved case management of STI, which were available in all facilities in both intervention and comparison communities throughout the trial, with drugs and other supplies ensured.

The third component, community-based condom promotion and distribution by youth, was introduced early in the second year of the trial in response to the results of process evaluation. Four to five youth per village were elected by their peers and trained in the social marketing of condoms.

Finally, community-wide activities included initial community mobilization followed by annual youth health weeks focused around interschool competitions and performances by local youth groups, twice-yearly youth health days at health facilities, and quarterly video shows linked to discussions that were open to all community members.

Surveys in the trial communities showed that sexual health activities in the comparison communities were very limited, and similar to non-MEMA kwa Vijana activities in the intervention communities.

Back to Top | Article Outline

Process evaluation

Extensive process evaluation of the intervention was conducted and fed back to the intervention team throughout the trial. This included questionnaires to trainees before and after all training courses; quarterly supervision visits to every intervention school and clinic; observation of in-class sessions, clubs and clinic sessions; checks on exercise books to see which sessions had been taught; annual feedback workshops with teachers; two externally conducted evaluation surveys with interviews and data collection from district to community levels [32]; regular feedback from the social science team who were studying the social and sexual norms of local young people (J. Wamoyi, D. Wight, M. Plummer, G. Mshana, D. Ross, Exchanging sex for gifts or money among young people in rural northern Tanzania, in preparation) [33]; and evaluations of programme components by international and national experts (D. Kirby, The MEMA kwa Vijana curriculum: a review, 2001, unpublished report; W. Lugoe. Evaluation of the teachers' training sessions for the MEMA kwa Vijana teacher-led component, 2001, unpublished report) [34].

Back to Top | Article Outline

Impact evaluation

The impact of the intervention was evaluated in a cohort of 9645 adolescents. All those aged 14 years or more (mean 15.7 years) in late 1998, who were in years 4–6 of all 121 government primary schools within the 20 trial communities (and about to enter years 5–7) were eligible for enrolment. An interim follow-up survey was conducted in 2000 and final follow-up between October 2001 and April 2002, approximately 3 years after recruitment. Strenuous attempts were made to locate cohort members. These included up to six household visits, and attempts to trace out-migrants using address information supplied by household members. Cohort members who were in year 7 when the intervention commenced (January 1999) could only receive one year of the in-school programme, those in year 6 2 years, and those in year 5 the full 3 years.

The predefined primary trial outcomes were HIV seroincidence during follow-up and HSV2 seroprevalence at final survey. Secondary outcomes were six further biological, five behavioural, one attitudinal and three knowledge outcomes (see Table 3). Each of the attitudinal and knowledge outcomes was based on the answers to three questions (Table 1). After registration, detailed identity checks and informed consent, and data on knowledge, attitudes and sexual behaviour were collected through a 15–25 min interview administered by a same-sex, 20–24-year-old research assistant. Laboratory specimens were then collected by trained technicians, and a clinician checked for clinical symptoms (male and female participants) and signs (male participants only) of sexually transmitted diseases and offered HIV counselling and testing.

In addition to the cohort evaluation, a formal test of knowledge and attitudes was carried out in July 2002 among all 4707 pupils who were in year 7 in the intervention and comparison schools. These were pupils in the year below the youngest group in the study cohort, and so were not cohort members. All pupils taking the test in the intervention schools had received all 3 years of the in-school intervention, which continued to be implemented after the trial period in the 10 intervention communities only. Tests were administered under examination conditions, supervised by a teacher from a different school.

Back to Top | Article Outline

Laboratory methods

Only urine specimens were collected at the baseline survey, and tested for HIV-1 antibodies using a semiquantitative particle agglutination test (GACPAT; Central Public Health Laboratory, Colindale, London, UK) [35]. Reactive specimens were tested with Wellcozyme HIV1+2 GAC enzyme-linked immunosorbent assay (ELISA; Murex Biotech Ltd., Dartford, UK) [36], which provided the definitive result. Urine was also tested for Chlamydia trachomatis and Neisseria gonorrhoeae by polymerase chain reaction (PCR; Amplicor, Roche Diagnostics, Branchburg, New Jersey, USA). Samples were pooled for PCR, with a pool size of five, with samples from reactive pools then tested individually [37]. Urine specimens from female participants were tested for pregnancy using an IPAS dipstick (Quickstick One Step HCG Pregnancy Test; IPAS/Pharmatech Inc., Denver, Colorado, USA).

At the final survey, serum and urine were collected from all participants and two self-administered vaginal swabs were collected by female subjects. Sera were tested for HIV-1 and HIV-2 using the Murex HIV Ag/Ab Combination ELISA (Murex Biotech). No specimen was reactive for HIV-2. Specimens reactive in the Murex HIV Ag/Ab ELISA were sent to the UK Central Public Health Laboratory for confirmation using a battery of antibody and antigen assays, including PCR when necessary. Full details of the HIV testing algorithm are given elsewhere [38]. Fifteen positive baseline HIV results (based on urine testing) were reclassified as negative when serum testing at the final survey showed these participants to be HIV negative.

Sera were tested for antibodies to HSV2 using a monoclonal enzyme immunoassay (HSV2 IgG; Kalon Biological Ltd., Ash Vale, Surrey, UK) [39]. Lifetime exposure to syphilis was examined using a Treponema pallidum particle agglutination (TPPA) test (Serodia TPPA; Fujirebio Inc., Tokyo, Japan). Urine from male subjects and vaginal swabs from female subjects were tested for C. trachomatis and N. gonorrhoeae using the same PCR methods as at baseline. Each vaginal swab was also tested for Trichomonas vaginalis using two independent sets of PCR primers [40–42]. A specimen was only defined as T. vaginalis positive if it was positive by both PCR assays. Urine specimens from female participants were tested for pregnancy using the same dipstick method as at baseline.

Back to Top | Article Outline

Statistical considerations

Based on previous survey data from the study communities [5], cumulative HIV sero-incidence over the 3 years of the trial was expected to be 1.6%. Assuming a between-community coefficient of variation for HIV incidence of 0.2 and 30% loss to follow-up, the sample size of 9645 would have given 75% power to detect a 50% reduction in HIV incidence in intervention communities if this had been the incidence rate within the trial.

Impact measures were based on ratios of prevalences, risks or rates in the intervention and comparison arms. For risks, the geometric mean risk for the 10 communities in each arm was computed, and the unadjusted risk ratio (RR) was calculated as the ratio of these geometric means. An approximate variance for the log(geometric mean) in each arm was obtained based on the residual mean square from a two-way analysis of variance (ANOVA) of community log-risk on stratum and study arm. A 95% confidence interval (CI) for the RR was calculated from this variance using a t-statistic with 14 degrees of freedom [43]. Analogous methods were used for prevalences and rates.

Logistic or Poisson regression were used to adjust for individual-level covariates [43]. The regression model included terms for risk stratum and predefined adjustment factors, but not study arm. For each community, the fitted model was used to compute the ratio of observed to expected events (O/E). The adjusted RR was obtained as the ratio of the geometric mean of these O/E estimates for the two study arms, and variances and CI were obtained from an ANOVA of log(O/E) on stratum and study arm.

For outcomes with zero cases in some communities, unadjusted and adjusted RR were obtained as the ratios of arithmetic mean risks and O/E, and approximate variances and CI were obtained from ANOVA of untransformed community risk or O/E on stratum and study arm.

Intervention impact was also analysed separately among those who were in years 4, 5 and 6 at baseline. The trend in adjusted RR with school year was assessed by regression of O/E on study arm and school year as a linear variable. A random effects model was used to allow for clustering by community. The statistical significance of the trend was obtained from the interaction between study arm and school year. A similar method was used to assess the interaction between study arm and marital status.

The analytical plan was approved by the trial's data and safety monitoring committee before the impact analysis commenced.

Back to Top | Article Outline

Ethical considerations

The trial protocol received ethical and research clearance from the Tanzanian National Medical Research Coordinating Committee and the Ethics Committee of the London School of Hygiene and Tropical Medicine. Primary school committees (which include parent representatives) gave written approval for schools to participate. Before each survey round, the study was explained to teachers and students, who were given a leaflet to take home to their parents that included notice of a meeting for parents and the right to inform the school if they did not want their child to take part. Signed informed consent was obtained from each student on the day of the survey round. At each round, cohort members could opt to receive their HIV test result after pretest and posttest counselling.

Back to Top | Article Outline


Process evaluation

The results of the process evaluation are described elsewhere [25]. Briefly, these showed that the intervention had been delivered to a high standard and with high coverage. For example, supervision visits to the schools showed that most teachers taught the sessions well, most class peer-educators ably performed brief dramas that were used as discussion starters, and a qualitative study confirmed that the teaching was well received by most pupils and communities [44]. Over 80% of sessions had already been taught 2–3 months before the end of each school year during the trial [45]. A simulated patient study found that health workers in intervention facilities were more respectful and empathic to youth than in comparison facilities [46]. Over 57 500 condoms were sold by the youth condom promoters/distributors in the 2 years of this intervention component.

Back to Top | Article Outline

Baseline characteristics

The baseline characteristics of the intervention and comparison groups were generally similar (Table 2). Slight baseline imbalances in ethnic group and lifetime number of partners were adjusted for in all analyses of trial outcomes. There were substantial differences between male and female participants, so outcomes were analysed separately by sex.

Back to Top | Article Outline

Completeness of follow-up

This is shown in Fig. 2. In summary, 7040 (73%) of the 9645 eligible cohort members were seen at the final survey. Follow-up rates were similar in intervention (72%) and comparison (74%) communities, higher among male (77%) than female (69%) participants (P < 0.001), and higher in younger participants (age at recruitment: 14 years, 74%; 15 years, 73%; 16+ years, 71%; P value for trend < 0.001).

Back to Top | Article Outline

Impact on knowledge and reported attitudes

At the final survey, there were substantial and statistically significant differences in the proportions of both male and female participants who answered all three questions ‘correctly’ for each of the three knowledge outcomes, and for the reported attitude outcome (Table 3, Fig. 3). The adjusted RR for these four outcomes ranged from 1.28 to 1.77 for male and from 1.41 to 1.58 in female participants. These results were independently confirmed by the results of a school examination administered to year 7 students in July 2002 (Table 4).

Back to Top | Article Outline

Impact on reported sexual behaviours

The proportion of young men reporting sexual debut during follow-up was 60% in the intervention and 72% in the comparison communities (adjusted RR 0.84, CI 0.71, 1.01). There was little difference among young women (adjusted RR 1.03, CI 0.91, 1.16; Table 3, Fig. 3). Similarly, the proportion of male subjects reporting more than one sexual partner in the past year was significantly lower in the intervention (19%) than in the comparison communities (28%; adjusted RR 0.69, CI 0.49, 0.95), but no significant difference was seen in female participants (adjusted RR 1.04, CI 0.58, 1.89).

The proportions who reported initiating condom use during follow-up were substantially and significantly higher in intervention communities among both male and female participants. The proportions reporting condom use at last sex were higher in intervention communities in both sexes, but this was only significant in young men, and absolute levels of condom use at last sex remained relatively low (< 30%; Table 3).

Back to Top | Article Outline

Impact on clinical symptoms and signs

At the final survey, the proportion of participants reporting genital pus or abnormal genital discharge during the past year was substantially lower in intervention communities, both among male (adjusted RR 0.58, CI 0.41, 0.83) and female (adjusted RR 0.59, CI 0.43, 0.80) participants. Among those reporting STI symptoms, however, there was no significant difference in the proportion who reported having sought care at a local health facility for their most recent STI episode during the past year, in either sex (Table 3).

Back to Top | Article Outline

Impact on HIV, sexually transmitted infections and pregnancy

The two primary outcomes of the trial, HIV incidence and HSV2 prevalence, were based on biological tests. Only 45 participants (five boys and 40 girls) seroconverted to HIV during 23 730 person-years of follow-up. After adjustment, HIV incidence in female subjects was 25% lower in the intervention communities, but this difference was not statistically significant (adjusted RR 0.75, CI 0.34, 1.66; Table 3). Overall, 12% of male and 21% of female participants were HSV2 seropositive at the final survey, but there was no difference by trial arm for either male (adjusted RR 0.92, CI 0.69, 1.22) or female (adjusted RR 1.05, CI 0.83, 1.32) participants (Table 3, Fig. 3).

Secondary biological outcomes included TPPA seroprevalence (syphilis), the prevalence of C. trachomatis and N. gonorrhoeae, and, in young women only, the prevalence of T. vaginalis and pregnancy, and the reported incidence of pregnancy during follow-up. There was no evidence of a protective effect of the intervention on any of these outcomes. In female participants, the prevalence of N. gonorrhoeae was higher in the intervention arm, and this difference was of borderline significance (Table 3). This difference was, however, only seen in those who were in year 6 at enrolment, who only had the potential to receive one year of in-school intervention (Table 5). A substantial proportion of young women were pregnant at the final survey (19%), and 46% reported having been pregnant for the first time during the 3-year follow-up period.

Back to Top | Article Outline

Trends by year of enrolment

In general, intervention/comparison differences in knowledge, reported attitudes and reported risk behaviours were greater among participants with more years of potential exposure to the in-school intervention, especially in young men (Table 5).

Back to Top | Article Outline

Impact by marital status

The impact of the intervention on pregnancy prevention knowledge was greater for those never married, among both female and male participants (P = 0.06 and P = 0.01, respectively), and there was some evidence of a greater impact on STI knowledge among never-married women (P = 0.07). The effect on HIV incidence was substantially greater in never married women (adjusted RR 0.40, CI 0.10–1.59) than ever-married women (adjusted RR 0.98, CI 0.47–2.03), although this was not statistically significant. There was no consistent difference by marital status for any of the behavioural or other biological outcomes (results not shown).

Back to Top | Article Outline


The study has confirmed that young people in this rural African population were at high risk of adolescent pregnancy and STI, with a low incidence of condom use and a high proportion of the cohort reporting sexual debut during 3 years of follow-up. There is, therefore, an urgent need for effective and affordable preventive interventions.

This trial has demonstrated the feasibility of large-scale implementation of an adolescent sexual health intervention using existing government staff and structures in sub-Saharan Africa. External evaluations of this multicomponent package of interventions showed that they were of high quality, well implemented and achieved high coverage. The average annual cost of the intervention was almost US$30 000 per trial community (approximately equivalent to an administrative ward, mean total population approximately 15 000) during the trial phase, including all start-up and capital costs, approximately equivalent to US$10 per adolescent per year within the primary target age range (12–19 years). Within a district-wide programme, first-year costs are projected to be US$22 000 per ward (US$7.30 per 12–19 year old), decreasing to US$3600 per ward (US$1.20 per 12–19 year old) in subsequent years [47].

Important limitations of the trial were that the interventions were deliberately constrained to be affordable and replicable on a large scale, and that the trial cohort included some young people who only received one or 2 years of the main, in-school component, rather than the full 3 years. The trial design also meant that mass media and other national or region-wide approaches could not be included. HIV incidence was much lower than predicted based on a previous survey of 15–19 year olds in the same communities [5]. This may have been a result of the closed nature of the cohort, which excluded in-migrants, whereas those who were lost to follow-up may have been at higher risk than those followed up. Despite considerable efforts to trace cohort members who were absent during the follow-up rounds, 27% were lost to follow-up (Fig. 2). As follow-up rates were very similar in the intervention and comparison arms, it is unlikely that this would have biased the results.

The intervention led to substantial and statistically significant improvements in knowledge and reported sexual attitudes in both sexes. There was no evidence that the intervention increased sexual activity. On the contrary, young men reported delayed sexual debut and a reduction in the reported number of partners in the past year. Reported behavioural effects were stronger in male than female participants, possibly because young women were exposed to older male partners who had not benefited from the programme, or because young men understood the intervention messages better or were better able to act on them. For some outcomes, especially among male participants, the data suggested a dose–response effect with greater impact among those receiving 2 or 3 years of the in-school programme (Table 5). Previous studies have shown that youth interventions can improve knowledge and attitudes in the short-term (< 6 months) [16]; this study has shown a strong and sustained impact on knowledge and reported attitudes, especially in young men. Young people in the intervention communities were also less likely to report STI symptoms within the past year.

Despite these differences in knowledge, and reported attitudes and behaviours, there was no consistent impact on biological outcomes, including HIV incidence, the prevalence of other STI or pregnancy rates. There are several potential explanations for this discrepancy.

First, such interventions may only change knowledge and reported attitudes, but not actual risk taking, at least in the short term. Reported behaviour can be unreliable in young people [23], and may be subject to differential reporting bias in intervention and comparison arms. Therefore, observed differences in reported behaviour may have reflected a better knowledge of the promoted behaviours rather than changes in actual behaviour.

Second, it may be that the differences in reported sexual behaviour were valid, but the behavioural changes were not large enough to impact significantly on biological outcomes, at least within the follow-up period.

Third, young people may need longer exposure to such interventions. Overall, 42% of the cohort only had the potential to receive one year of the in-school intervention, and a further 32% could only have received 2 years. The tendency for greater beneficial impact on knowledge and reported attitudes among those who had the potential to receive 3 years of in-school intervention lends some support to this hypothesis.

Fourth, young men have considerably more decision-making power within sexual partnerships than young women [33]. Our study cohort were on average several years older than the young women [48], and would not have benefited from the school-based intervention.

Fifth, those youth who leave school early or attend irregularly may be at disproportionately high risk, but would miss the in-school component of the intervention.

Finally, it may be that additional interventions are needed to make an impact within the short term. The interventions that were tested within this trial were all directly targeted to adolescents themselves. Cultural norms, however, such as gendered and age-related power relationships and marriage and fertility norms within the wider community, compromise the ability of adolescents to change their sexual behaviour [33]. Community-wide interventions aimed at changing societal norms may be particularly important [49].

Accurate knowledge and skills are a prerequisite for young people who want to change their behaviour to reduce their risk of HIV and other sexual health problems. This trial has shown that it is possible for a large-scale intervention, implemented through local government workers and supervision structures, substantially to improve knowledge, reported attitudes, and reported behaviours. The effects on these outcomes were at least as great as those recorded in previous studies in Africa [15–17], and there was no evidence of increased risk-taking behaviour. Furthermore, these effects were present after a relatively long follow-up period (approximately 3 years) after the interventions were initiated.

It has often been assumed when interpreting the results of previous studies that improvements in self-reported sexual behaviours can be used as a proxy for reductions in true behaviour and hence reductions in HIV, other STI or pregnancy [9–12,15–17]. The lack of any consistent effect on the biological outcomes measured in our trial, despite substantial impacts on knowledge, reported attitudes and self-reported sexual behaviours, raises serious questions about the interpretation of previous studies, and argues strongly for the inclusion of biological outcomes in future programme evaluations.

Back to Top | Article Outline


This study was a collaboration between the Tanzanian National Institute for Medical Research (NIMR), the African Medical and Research Foundation (AMREF), the London School of Hygiene and Tropical Medicine (LSHTM), and the Ministries of Health and of Education and Culture of the Government of Tanzania.

The authors would like to thank the entire intervention, survey, data, laboratory, social science and administrative staff in Mwanza, and in particular Joseph Chilongani, Kenneth Chima, Robert Lugundu, Julius Mngara, Godwin Mmassy, Obed Mrita, Gerry Mshana, Zachayo Salamba, and Joyce Wamoyi. They also thank Aura Beltran, France Bernier, Tania Crucitti, Sandra Molina, Angels Natividad Sancho, Iain Tatt, and Rosanna Peeling, who either conducted or supervised additional laboratory analyses, Lilani Kumaranayake who supervised the costing study, and Sarah Henson, Eloise Turner and Onno Dekker, who provided administrative support to the study in Europe. The main authors apologize to the co-authors of this paper whose names are only listed in an appendix. This was due to the limits imposed by the journal editors.

The authors are grateful to the members of the study's Steering and Scientific Advisory Committees, which were co-chaired by the Director of Preventive Services in the Tanzanian Ministry of Health, and the Director of Basic Education in the Ministry of Education and Culture, and to the members of the Data and Safety Monitoring Committee: Peter Smith, Andrew Kitua, and Helen Weiss (before she became the trial statistician). They thank the Tanzanian Ministry of Health and Ministry of Education and Culture, National AIDS Control Programme, and Medical Research Co-ordinating Committee, and the Ethics Committee of the LSHTM, for permission to carry out and publish the results of this study. They are also grateful to the Regional Medical and Education Officers of Mwanza and the National Directors of NIMR and AMREF for their support.

This report is dedicated to the young people of Mwanza Region who participated in the study, and the teachers, health workers, and community members who implemented the interventions.

Sponsorship: The study was supported by grants from the European Commission, Development Cooperation Ireland (now Irish Aid), UK Medical Research Council, UNAIDS, and UK Department for International Development. The sponsors of the study had no role in the study design, data collection, analysis, interpretation, or writing of the report.

Back to Top | Article Outline


1. UNAIDS. Young people and HIV/AIDS: opportunity in crisis. Geneva, Switzerland: UNAIDS, WHO, UNICEF; 2002.
2. UNAIDS. AIDS epidemic update: December 2002. Geneva, Switzerland: UNAIDS; 2003.
3. UNAIDS and WHO. AIDS epidemic update: December 2003 (03.39E). Geneva, Switzerland: UNAIDS; 2003.
4. Temmerman M. Sexually transmitted diseases and reproductive health. Sex Transm Dis 1994; 21(Suppl 2):S55–S58.
5. Obasi AI, Balira R, Todd J, Ross DA, Changalucha J, Mosha F, et al. Prevalence of HIV and Chlamydia trachomatis infection in 15 to 19 year olds in rural Tanzania. Trop Med Int Health 2001; 6:517–525.
6. Pequegnat W, Stover E. Behavioral prevention is today's AIDS vaccine. AIDS 2000; 14(Suppl 2):S1–S7.
7. MacPhail C, Williams BG, Campbell C. Relative risk of HIV infection among young men and women in a South African township. Int J STD AIDS 2002; 13:331–342.
8. Buvé A, Carael M, Hayes RJ, Auvert B, Ferry B, Robinson NJ, et al. Multicentre study on factors determining differences in rate of spread of HIV in sub-Saharan Africa: methods and prevalence of HIV infection. AIDS 2001; 15(Suppl 4):S5–S14.
9. Robin L, Dittus P, Whitaker D, Crosby R, Ethier K, Mezoff J, et al. Behavioral interventions to reduce incidence of HIV, STD, and pregnancy among adolescents: a decade in review. J Adolesc Health 2004; 34:3–26.
10. Johnson BT, Carey MP, Marsh KL, Levin KD, Scott-Sheldon LAJ. Interventions to reduce sexual risk for the human immunodeficiency virus in adolescents, 1985–2000: a research synthesis. Arch Pediatr Adolesc Med 2003; 157:381–388.
11. Kirby D. Sex and HIV/AIDS education in schools. BMJ 1995; 311:403.
12. Jemmott JB, Jemmott LS. HIV risk reduction behavioral interventions with heterosexual adolescents. AIDS 2000; 14(Suppl 2):S40–S52.
13. DiCenso A, Guyatt G, Willan A, Griffith L. Interventions to reduce unintended pregnancies among adolescents: systematic review of randomised controlled trials. BMJ 2002; 324:1426–1430.
14. Wight D, Raab GM, Henderson M, Abraham C, Buston K, Hart G, Scott S. The limits of teacher delivered sex education: interim behavioural outcomes from a randomised trial. BMJ 2002; 324:1430–1433.
15. Grunseit A. Impact of HIV and sexual health education on the sexual behaviour of young people: a review update. Geneva, Switzerland: UNAIDS; 1997.
16. Gallant M, Maticka-Tyndale E. School-based HIV prevention programmes for African youth. Soc Sci Med 2004; 58:1337–1351.
17. Shuey DA, Babishangire BB, Omiat S, Bangarukayo H. Increased sexual abstinence among in-school adolescents as a result of school health education in Soroti district, Uganda. Health Educ Res 1999; 14:411–419.
18. Catania JA, Gibson DA, Chitwood DD, Coates TJ. Methodological problems in AIDS behavioral research: influences on measurement error and participation bias in studies of sexual behavior. Psychol Bull 1990; 108:339–362.
19. Peterman TA, Lin LS, Newman DR, Kamb ML, Bolan G, Zenilman J, et al, The Project RESPECT Study Group. Does measured behavior reflect STD risk? An analysis of data from a randomized controlled behavioral intervention study. Sex Transm Dis 2000; 27:446–451.
20. Clark LR, Brasseux C, Richmond D, Getson P, D'Angelo LJ. Are adolescents accurate in self-report of frequencies of sexually transmitted diseases and pregnancies? J Adolesc Health 1997; 21:91–96.
21. McFarlane M, St Lawrence JS. Adolescents' recall of sexual behavior: consistency of self-report and effect of variations in recall duration. J Adolesc Health 1999; 25:199–206.
22. Wight D, West P. Poor recall, misunderstandings and embarrassment: interpreting discrepancies in young men's reported heterosexual behaviour. Culture Health Sexual 1999; 1:55–78.
23. Plummer ML, Ross DA, Wight D, Changalucha J, Mshana G, Wamoyi J, et al. “A bit more truthful”: the validity of adolescent sexual behaviour data collected in rural northern Tanzania using five methods. Sex Transm Infect 2004; 80(Suppl II):ii49–ii56.
24. Hayes R, Changalucha J, Ross DA, Gavyole A, Todd J, Obasi AIN, et al. The MEMA kwa Vijana Project: design of a community-randomised trial of an innovative adolescent sexual health intervention in rural Tanzania. Contemp Clin Trials 2005; 26:430–442.
25. Obasi AIN, Cleophas B, Ross DA, Chima KL, Gavyole A, Plummer ML, et al. Rationale and design of an adolescent reproductive health intervention in Mwanza Region, Tanzania. AIDS Care 2006; 18:311–322.
26. Matasha E, Ntembelea T, Mayaud P, Saidi W, Todd J, Mujaya B, Tendo-Wambura L. Sexual and reproductive health among primary and secondary school pupils in Mwanza, Tanzania: need for intervention. AIDS Care 1998; 10:571–582.
27. Klepp K, Ndeki SS, Leshabari MT, Hannan PJ, Lyimo BA. AIDS education in Tanzania: promoting risk reduction among primary school children. Am J Pub Health 1997; 87:1931–1936.
28. Schapink D, Hema J, Mujaya B. Youth and HIV/AIDS programmes. In: Ng'weshemi J, Boerma T, Bennett J, Schapink D, editors. HIV prevention and AIDS care in Africa: a district level approach. Amsterdam: Royal Tropical Institute; 1997. pp. 163–184.
29. Kirby D. Reducing adolescent pregnancy: approaches that work. Contemp Pediatr 1999; 16:83–94.
30. Wight D, Abraham C. From psycho-social theory to sustainable classroom practice: developing a research-based teacher-delivered sex education programme. Health Educ Res 2000; 15:25–38.
31. WHO, UNESCO. School health education to prevent AIDS and STD; a resource package for curriculum planners: students' activities. Geneva, Switzerland: WHO; 1994.
32. Mosha F, Masanja J, Ross D, Hayes RJ, Todd JE, Obasi A, et al. Monitoring attitudes towards adolescent sexual health among programme implementers in Tanzania. In: XIVth International AIDS Conference. Barcelona, Spain, 7–12 July 2002. Abstract no. MoPeF3950.
33. Wight D, Plummer M, Mshana G, Wamoyi J, Salamba Z, Ross D. Contradictory sexual norms and expectations for young people in rural northern Tanzania. Soc Sci Med 2006; 62:987–997.
34. Guyon A, Obasi A, Lugoe W, Wamoyi J, Mshana G, Plummer M, et al. Evaluation of HIV/AIDS peer education in primary schools of Mwanza, Tanzania. In: XIIIth International AIDS Conference. Durban, South Africa, 9–14 July 2000. Abstract LbPeD7125.
35. Connell JA, Parry JV, Mortimer PP, Duncan RJS, McLean KA, Johnson AM, et al. Preliminary report: accurate assays for anti-HIV in urine. Lancet 1990; 335:1366–1369.
36. Connell JA, Parry JV, Mortimer PP, Duncan J. Novel assay for the detection of immunoglobulin G antihuman immunodeficiency virus in untreated saliva and urine. J Med Virol 1993; 41:159–164.
37. Peeling RW, Toye B, Jessamine P, Gemmill I. Pooling of urine specimens for PCR testing: a cost saving strategy for Chlamydia trachomatis control programmes. Sex Transm Infect 1998; 74:66–70.
38. Everett DB, Weiss H, Changalucha J, Anemona A, Chirwa T, Ross DA, et al. Low specificity of the Murex fourth generation HIV-enzyme immunoassay in Tanzanian adolescents (in press).
39. Van Dyck E, Buvé A, Weiss HA, Glynn JR, Brown DWG, De Decken B, et al. Performance of commercially available enzyme immunoassays for detection of antibodies against herpes simplex virus type 2 in African populations. J Clin Microbiol 2004; 42:2961–2965.
40. Kengne P, Veas F, Vidal N, Rey JL, Cuny G. Trichomonas vaginalis: repeated DNA target for highly sensitive and specific polymerase chain reaction diagnosis. Cell Mol Biol 1994; 40:819–831.
41. Shaio MF, Lin PR, Liu JY. Colorimetric one-tube nested PCR for detection of Trichomonas vaginalis in vaginal discharge. J Clin Microbiol 1997; 35:132–138.
42. Crucitti T, van Dyck E, Tehe A, Abdellati S, Vuylsteke B, Buve A, Laga M. Comparison of culture and different PCR assays for detection of Trichomonas vaginalis in self collected vaginal swab specimens. Sex Transm Infect 2003; 79:393–398.
43. Bennett S, Parpia T, Hayes R, Cousens S. Methods for the analysis of incidence rates in cluster randomized trials. Int J Epidemiol 2002; 31:839–846.
44. Plummer ML, Wight D, Obasi AI, Wamoyi J, Mshana G, Todd J, et al. A process evaluation of a school-based adolescent sexual health intervention in rural Tanzania: the MEMA kwa Vijana programme. Health Educ Res: Theory Pract, 3 October 2006. E-pub ahead of print.
45. Chima K, Obasi A, Cleophas-Frisch B, Makokha M, Ross D, Gavyole A, et al. Participatory reproductive health education for primary school: lessons learned during the first 3 years of an intervention in Mwanza, Tanzania. In: XIVth International AIDS Conference. Barcelona, Spain, 7–12 July 2002. Abstract no. ThPeF8139.
46. Cleophas-Frisch B, Obasi A, Mshana G, Wamoyi J, Plummer M, Rwakatare M, et al. Evaluation of the youth-friendliness of reproductive health services in rural Tanzania using simulated patients. In: XIVth International AIDS Conference. Barcelona, Spain, 7–12 July 2002. Abstract no. WeOrD1277.
47. Terris-Prestholt F, Kumaranayake L, Obasi A, Cleophas-Mazige B, Makokha M, Todd J, et al. From trial intervention to scale-up: costs of an adolescent sexual health program in Mwanza, Tanzania. Sex Transm Dis 2006; 33(10 Suppl):S133–S139.
48. Boerma JT, Urassa M, Nnko S, Ng'weshemi J, Isingo R, Zaba B, Mwaluko G. Sociodemographic context of the AIDS epidemic in a rural area in Tanzania with a focus on people's mobility and marriage. Sex Transm Infect 2002; 78(Suppl 1):i97–i105.
49. Campbell C. ‘Letting them die’: Why HIV/AIDS prevention programmes fail. Oxford, UK: International African Institute; 2003.
Back to Top | Article Outline


Additional co-authors: Rebecca Baliraa, Daniel Wightb, Awene Gavyolec, Maende J. Makokhac, Frank Moshaa, Fern Terris-Prestholtd, John V. Parrye

aNational Institute for Medical Research, Mwanza Centre, Mwanza, Tanzania

bSocial and Public Health Sciences Unit, Medical Research Council, Glasgow, UK

cAfrican Medical and Research Foundation (AMREF), Mwanza, Tanzania

dLondon School of Hygiene and Tropical Medicine, London, UK

eSexually Transmitted and Blood Borne Virus Laboratory, Central Public Health Laboratory, Health Protection Agency, London, UK

Contributors: D.A. Ross was the project leader in Mwanza throughout, contributed to all aspects of the study, and took the lead in drafting the report. J. Changalucha co-supervised the laboratory testing, and contributed to the design and interpretation of the impact evaluation component of the study. A.I.N Obasi led the team that designed and implemented the interventions and contributed to the design of the impact evaluation. J. Todd was in charge of the data management and analysis from 1997 to 2000, and played a leading role in the design and implementation of the initial survey and baseline and interim follow-up survey rounds. M.L. Plummer led the team that conducted detailed qualitative social research that contributed to the design of the interventions and the interpretation of the findings, and played a leading role in the design and implementation of the final follow-up survey round. B. Cleophas-Mazige was directly responsible for the design and implementation of the health services component of the intervention, and led the intervention team during 2002. A. Anemona was in charge of the data management and analysis from 2000 to 2002. M. Makokha was actively involved in the implementation of the intervention from 2001 onwards, and played a major role in the design of the in-school component in 2001 and 2002. R. Balira played a major role in the data management for the trial. D. Everett co-supervised the laboratory testing of the interim and final follow-up survey rounds. H. Weiss took the lead in the analysis of the trial results. D. Wight contributed to the design of the intervention and interpretation of the trial results, and supervised the qualitative social research. F. Mosha was in charge of the field data collection teams throughout the trial. J. Parry was responsible for external quality control on the HIV testing, and for resolving the HIV status of problem sera from the final survey. F. Terris-Prestholt was responsible for the costing of the intervention. A. Gavyole was AMREF's Programme Coordinator for the Lake Zone of Tanzania throughout the trial, and advised on the design and implementation of the intervention and the interpretation and policy implications of the findings. D.C. Mabey contributed to the design of the trial and supervised the laboratory procedures. H. Grosskurth contributed to the design of the trial and the intervention, and advised on other aspects of the study. R.J. Hayes took the lead on the design of the trial and the interpretation of the results, and contributed to all other aspects of the study. All co-authors contributed to the writing of the report.


adolescents; HIV; pregnancy prevention; sexually transmitted infections; Tanzania

© 2007 Lippincott Williams & Wilkins, Inc.