The Regai Dzive Shiri Project: a cluster randomised controlled trial to determine the effectiveness of a multi-component community-based HIV prevention intervention for rural youth in Zimbabwe – study design and baseline results
Frances M. Cowan,
Centre for Sexual Health and HIV Research, University College London, London, UK
Department of Community Medicine, University Of Zimbabwe, Harare, Zimbabwe
Corresponding Author Frances M. Cowan, Centre for Sexual Health and HIV Research, Mortimer Market Centre Off Capper Street, London WC1E 6AU, UK. Tel.: +44 207 530 3639; Fax: +44 207 530 3646; E-mail: email@example.com
Objective To assess the effectiveness of a community-based HIV prevention intervention for adolescents in terms of its impact on (1) HIV and Herpes simplex virus type 2 (HSV-2) incidence and on rates of unintended pregnancy and (2) reported sexual behaviour, knowledge and attitudes.
Methods Cluster randomised trial of a multi-component HIV prevention intervention for adolescents based in rural Zimbabwe. Thirty communities were selected and randomised in 2003 to early or deferred intervention implementation. A baseline bio-behavioural survey was conducted among 6791 secondary school pupils (86% of eligibles) prior to intervention implementation.
Results Baseline prevalences were 0.8% (95% CI: 0.6–1.0) for HIV and 0.2% (95% CI: 0.1–0.3%) for HSV-2. Four girls (0.12%) were pregnant. There was excellent balance between study arms. Orphans who made up 35% of the cohort were at increased risk of HIV [age–sex adjusted odds ratio 3.4 (95% CI: 1.7–6.5)]. 11.9% of young men and 2.9% of young women reported that they were sexually active (P < 0.001); however, there were inconsistencies in the sexual behaviour data. Girls were less likely to know about reproductive health issues than boys (P < 0.001) and were less likely to have used and to be able to access condoms (P < 0.001).
Conclusion This is one of the first rigorous evaluations of a community-based HIV prevention intervention for young people in southern Africa. The low rates of HIV suggest that the intervention was started before this population became sexually active. Inconsistency and under-reporting of sexual behaviour re-emphasise the importance of using externally validated measures of sexual risk reduction in behavioural intervention studies.
Objectif: Evaluer l’efficacité d’une intervention de prévention contre le VIH basée sur la communauté chez les adolescents, en fonction de l’impact sur (1) l’incidence du VIH et le virus Herpes simplex de type 2 (HSV-2) et sur les taux de grossesses non désirées, (2) le comportement sexuel, les connaissances et les attitudes rapportés.
Méthode: Essai randomisé contrôlé en grappe d’une intervention à multiples composantes pour la prévention du VIH chez les adolescents, basé en zone rurale du Zimbabwe. 30 communautés ont été sélectionnées et randomisées en 2003 pour l’application d’une intervention imminente ou différée. Une enquête sur le comportement de base a été menée auprès de 6791 élèves d’école secondaire (86% des éligibles) avant l’application de l’intervention.
Résultats: Les prévalences de base étaient de 0,8% (IC95%: 0,6-1,0) pour le VIH et 0,2% (IC95%: 0,1-0,3%) pour le HSV-2. Quatre filles (0,12%) étaient enceintes. Il y avait un excellent équilibre entre les groupes étudiés. Les orphelins qui constituaient 35% de la cohorte étaient exposés à un risque plus accru pour le VIH (odds ratio ajusté pour l’âge et le sexe, OR = 3,4; IC95%: 1,7-6,5). 11,9% des jeunes hommes et 2,9% des jeunes femmes ont déclaréêtre sexuellement actifs (p < 0,001). Toutefois, il y avait des incohérences dans les données relatives au comportement sexuel. Les filles étaient moins susceptibles d’avoir des connaissances sur la santé de reproduction que les garçons (p < 0,001) et étaient moins susceptibles d’avoir utilisé et d’être en mesure d’accéder à des préservatifs (p < 0,001).
Conclusion: Cette étude est une des premières évaluations rigoureuses d’une intervention de prévention contre le VIH basée sur la communauté pour les jeunes en Afrique australe. Le faible taux de prévalence du VIH donne à penser que l’intervention a débuté avant que cette population devienne sexuellement active. L’incohérence et la sous-estimation du comportement sexuel soulignent de nouveau l’importance d’utiliser dans les études d’intervention comportementale, des moyens validés ailleurs sur la réduction des risques sexuels.
Objetivo: Evaluar la efectividad de una intervención basada en la comunidad para la prevención del VIH entre adolescentes en términos de su impacto sobre (1) VIH y virus de Herpes Simplex virus tipo 2 (VHS-2) incidencia y tasa actual de embarazos no deseados; (2) comportamiento sexual reportado, conocimientos y actitudes.
Métodos: Ensayo aleatorizado en grupos de una intervención multicomponente para la prevención del VIH en adolescentes de un área rural de Zimbabwe. Se seleccionaron 30 comunidades en el 2003 y se aleatorizaron para la implementación temprana o diferida de la intervención. Se realizó un estudio de biocomporatamiento de base entre 6791 estudiantes de secundaria (86% elegibles), antes de la implementación de la intervención.
Resultados: La prevalencia de base era del 0.8% (95% IC: 0.6-1.0) para VIH y 0.2% (95% IC: 0.1-0.3%) para VHS-2. Cuatro niñas (0.12%) estaban embarazadas. Existía un balance excelente entre los dos brazos del estudio. Los huérfanos, que componían un 35% de la cohorte, tenían un mayor riesgo de VIH (odds ratio ajustado por sexo y edad 3.4 (95% IC: 1.7-6.5). Un 11.9% de los chicos y un 2.9% de las chicas reportaron ser sexualmente activos (p < 0.001). Sin embargo existían inconsistencias en los datos sobre su conducta sexual. Las niñas eran más propensas que los chicos a no tener conocimiento sobre cuestiones sanitarias reproductivas (p < 0.001) al igual que a haber utilizado o tener acceso a preservativos (p < 0.001).
Conclusiones: Esta es una de las primeras evaluaciones rigurosas sobre una intervención basada en la comunidad para la prevención del VIH entre jóvenes de sur de África. Las bajas tasas de VIH sugieren que la intervención comenzó antes de que esta población fuese sexualmente activa. Las inconsistencias y el bajo reporte de comportamiento sexual reafirman la importancia de utilizar mediciones externas validadas de reducción de riesgo sexual en estudios de intervención sobre comportamiento.
In 2004, the Joint United Nations Programme on HIV/AIDS (UNAIDS) commissioned a series of systematic reviews to provide policy makers with evidence on the effectiveness of HIV prevention for young people (UNAIDS Interagency Task Team on Young People 2006). Evidence was assembled for interventions in schools (Kirby et al. 2006), in communities (Maticka-Tyndale & Brouillard-Coyle 2006) and to improve young people’s access to health services (Dick et al. 2006). While there was good evidence that school-based interventions can reduce reported sexual risk taking and that providing training to make health clinics more ‘youth friendly’ increases clinic usage by young people, data to support or argue against the implementation of broader, more community-based approaches, which aim to change societal norms in order to support individual behaviour change, were relatively sparse.
In this paper, we report the design and baseline results of a community randomised trial of a multi-component community-based HIV prevention intervention for young people underway in rural Zimbabwe. The main aim of the trial is to measure the effectiveness of the intervention delivered to secondary school students, out-of-school youth and the wider community in reducing the rates of HIV-1, Herpes simplex virus type 2 (HSV-2) and of unintended pregnancy among young Zimbabweans after 4 years of intervention delivery. The project is called the Regai Dzive Shiri Project (taken from the Shona proverb ‘Regai dzive shiri mazai haana muto’– which literally translates as ‘Give the eggs a chance to hatch, because birds are better to eat than eggs which have no sauce’. The proverb is used to support the notion of giving young people time to grow up. In rural Zimbabwe, in 2003, a majority (>90%) of young people attended secondary school, transitioning from primary to secondary school at around 11–13 years, and secondary school students were therefore representative of Zimbabwean youth more generally.
Study area and population
The study is being conducted in seven districts in three provinces in south-eastern Zimbabwe (Figure 1). Thirty rural communities were selected for inclusion. A community is defined as the rural health clinic, its catchment population and the secondary schools; in nine communities, there are two clinics rather than one. Communities were selected if there were at least 250 Form 2 (ninth year of schooling) students attending local secondary schools and there was an absence of HIV prevention specifically targeting young people.
The intervention is being delivered to in- and out-of-school youth, parents and rural clinic health staff. The impact is being measured in a cohort of 6791 young people who were attending Form 2 in study secondary schools between March 2003 and June 2003, when school-going youth were broadly representatives of young people living in rural Zimbabwe. School attendance has fallen since then as a consequence of the deteriorating economic situation.
The overall design of this trial is illustrated in Figure 2. In brief, a cohort of young people enrolled in Form 2 at study schools, was recruited in 2003. The baseline survey was conducted in this cohort March–June 2003. An interim survey was conducted among cohort members in 12 of 30 randomly selected communities in 2006, in order to re-ascertain the likely power of the trial by determining HIV prevalence and rates of loss to follow-up at year 3. The final evaluation survey was run in 2007 (year 4). The results of the interim and final survey will be published elsewhere.
Community mobilisation, access and consent processes
Community sensitisation took place at national, provincial and district levels before entering study communities. The study was approved by the Zimbabwe Ministry for Health and Child Welfare and the National AIDS Council. Ethical approval was obtained from the Medical Research Council of Zimbabwe and the ethics committees of University College London and the London School of Hygiene and Tropical Medicine. Strong links were established with other relevant ministries and community stakeholders.
Five people from each community were trained as community advisory board (CAB) members (150 in all) and were crucial to successful community sensitisation and cohort recruitment. A list of potential study participants (all Form 2 students attending study schools) was drawn up from school registers. Community meetings were held to inform the community about the study and to obtain written consent from parents for their child’s participation in the research (consent was not sought for participation in the intervention itself). If parents failed to attend these meetings, home visits were undertaken to obtain consent.
The randomisation process
In February 2003, a meeting was held at which randomisation of the 30 communities to early or deferred intervention implementation took place. More than 100 provincial, district and community representatives attended the meeting. Restricted randomisation was used to ensure balance between arms of the study (see below) (Hayes et al. 2000).
The process of randomisation was as follows. The communities were placed into three strata based on the distance of the clinic from a tarred road. There were 16 299 360 ways that the 30 communities could be allocated to the two arms ensuring balance across these strata. Randomisation was then restricted to ensure an equal number of schools in each arm; balance across districts; and an average sample size per community between 255 and 261 in each arm. The 8575 allocations satisfying these conditions were listed, and one was selected randomly at the randomisation meeting.
The baseline survey of the study cohort
The baseline survey took place between March 2003 and June 2003. Young people whose parents had consented were asked to sign to confirm their assent or refusal to take part.
Participants in the baseline survey were asked to self-complete a questionnaire as it was read out loud in Shona (the indigenous language) by a trained surveyor. Participants provided a finger-prick blood sample for HIV & HSV-2 antibody testing and girls provided a urine sample for pregnancy testing. The questionnaire collected information on socio-demographic characteristics, socio-economic status (SES) of their household (using questions from the Zimbabwe census), self-esteem, self-efficacy, sexual knowledge and attitudes, sexual behaviour including sexual intercourse, other risky behaviours and their hopes and aspirations for the future.
Design of main intervention
The intervention is theoretically based in social learning theory and the stages of change model (Prochaska & DiClemente 1992; Bandura et al. 1997). It aims to achieve change in societal norms within communities. The intervention has three components described in detail in Table 1:
Table 1. The Regai Dzive Shiri intervention
In-school programme (First 3 years of the intervention; delivered to all students in years 1–4)
The 3 year MEMA kwa Vijana curriculum (Hayes et al. 2005; Obasi et al. 2006; Plummer et al. 2007) was adapted for use in Zimbabwe. Additional introductory sessions on self-awareness and communication were added in year 1 and on self belief and gender issues in year 2.
Supplemented with materials developed locally by the Zimbabwe CDC AIDS Project (Talktime and Mopani Junction), Training and Research Support Centre (Auntie Stella –http://www.auntiestella.co.zw) and JSI UK (Young People We Care – an innovative programme which integrates HIV prevention for young people into a broader package of providing treatment and care for those affected/infected by HIV).
In year 4, the programme became entirely community based.
In years 1–3, community groups were established, which undertook HIV prevention activities using the supplementary materials outlined above.
In year 4, a 24-session out-of-school youth programme based on the ‘Stages of Change’ theoretical model was developed and implemented de novo. This is a highly participatory course which uses techniques developed for rural development, including risk and body mapping, drama, story telling and role play. The programme is run with groups of 20–30 participants over 4 weeks and groups are then encouraged to continue with community outreach to families and young people affected by HIV. Participants who attend at least 75% of sessions received an attendance certificate and a project T-shirt.
Five day residential youth-friendly clinic staff training for at least one nurse in each clinic with refresher training after 2 years. On-site training was provided for remaining clinic staff.
Monthly support visits by project staff when key features of clinic accessibility were assessed and on-site training tailored to any deficiencies noted.
Standards for youth-friendly services provision were developed by clinic nurses, who are independently assessed against these standards at 6 monthly intervals. Prizes are awarded to clinics performing well. Each clinic receives detailed independent feedback on their performance at these standard assessments. The district nursing leadership is an integral part of this process.
The community component aims to (i) raise awareness of issues relating to adolescent sexuality among parents and other adults in the community; (ii) improve communication between parents and their children; (iii) make the community a safer place for young people to be and (iv) enable adults to support young people to reduce their reproductive health risk.
The community intervention comprises two modules (eleven 3-h sessions for each module) delivered by carefully selected, trained and supported community facilitators. The sessions are designed to be as participatory as possible in order to maximise ownership of the learning points, encourage development of life skills, encourage internalisation of attitude change and be more enjoyable for participants. It is hypothesised that people will be more likely to accept new ideas if they come from within the community.
In addition to participatory techniques, the sessions use proverbs and bible verses to support new ideas that community members may be wary of or have difficulty understanding. Traditional games are also used to help support and clarify learning points. Participants can only progress to Module 2 if they have completed at least 75% sessions in Module 1.
• The youth programme for in- and out-of-school youth is delivered by carefully selected and trained Zimbabwean school leavers in the year between leaving school and starting university. These school leavers work as volunteers and go to live and work in the rural communities for 8–10 months of the year. They act both as role models for young people and as a bridge between adults and youth within communities. These professional peer educators (PPEs) use well-structured, theoretically-based materials which they deliver in a highly participatory way. The programme is delivered to all students and out-of-school youth who want to take part and not just those enrolled in the trial cohort.
• The programme for parents and community stakeholders is a 22-session community-based programme, which aims to improve knowledge about reproductive health, communication between parents and their children and community support for adolescent reproductive health. The community component arose from focus group discussions held with parents during the feasibility study (Cowan et al. 2002; Power et al. 2004). Parents lamented the collapse of traditional communication structures and acknowledged their lack of communication skills. Interestingly young people also said that they struggled to communicate with their parents especially about reproductive health issues and that they saw this as an important barrier to staying safe.
• The programme for nurses and other staff working in rural health clinics aims to improve accessibility of clinics for out-of-school youth.
These three components are highly integrated. For example, nurses trained to run the clinic intervention also run sessions within the youth and parents programmes and in so doing publicise the accessibility of the clinic. PPEs help run the ‘youth corners’ at the clinics and help facilitate sessions in the parents programme. Integrating the three components in this way makes them mutually supportive and reinforcing. By living and working so closely with the community, it is hypothesised that the PPEs are able to change the norms of that community through challenging the norms that may be detrimental to adolescent reproductive health and reinforcing those that are beneficial.
Intervention in the deferred intervention arm
No specific intervention was introduced into the deferred intervention arm, but standard HIV prevention activities were implemented through the District AIDS Action Committees by local and international governmental and non-governmental organisations across both early and deferred intervention communities. Details of these activities are being recorded. The intervention was delivered in these communities in 2007 after final survey completion.
The project provided voluntary HIV counselling and testing through rural health clinics in all 30 communities on 1 day a month for the duration of the study. Uptake and acceptability was recorded and will be reported elsewhere.
Process evaluation of the intervention
Monitoring and evaluation data were collected throughout the study to document the study and non-study activities underway in communities, the number of people attending these activities and the frequency of their attendance. Clinic attendance figures were collected from study clinics. A team of social scientists undertook regular assessment of the intervention to assess whether each component was being delivered as intended. The results of these evaluations were fed back to the intervention teams who modified the intervention and or its delivery as necessary.
Blood samples were collected onto filter paper [US National Committee for Clinical Laboratory Standards (NCCLS), unpublished observation]. The samples were tested for HIV-1 antibody at the National Microbiology Reference Laboratory in Harare using a validated testing algorithm (US Department of Health and Human Services et al., unpublished observation). All specimens were tested using two ELISA tests (Vironostika® HIV Microelisa System BioMerieux, Inc., Durham, NC) and AniLabsytems EIA kit (AniLabsystems Ltd, Vantaa, Finland), with western blot used in the case of discrepant results. Dried blood spot (DBS) samples were tested for antibodies to HSV-2 using a type-specific HSV-2 assay (Focus HerpeSelect EIA, Focus Technologies, Cypress CA) with the index for diagnosing positive samples raised to >3.4 to minimise the number of false positives (Morrow et al. 2006). Urine samples were tested on-site for pregnancy using Cortez OneStep hCG Rapidip InstaTest®.
Data handling and analysis
All questionnaire data from the baseline survey were double-entered into an Access database and range and consistency checks were performed. An overall SES scale was created from household asset data. Each item was coded on a scale of 0 to 1 where 0 was considered to be the poorest SES category and 1 the richest. The scores were then added and the sum of each scale used to calculate a mean score for each participant. SES categories were then created by dividing this mean score distribution into quintiles. All statistical analyses were performed using stata 9.2 (College Station, TX).
On completion of the trial, the primary analysis will be based on the cumulative incidences of HIV, HSV-2 and unintended pregnancies at the final follow-up survey, at median age 19 years. Statistical methods that allow for clustering of events within community will be used for the analysis (Donner & Klar 2000; Hayes et al. 2000). Further analyses will be carried out to adjust for covariates showing baseline imbalance.
Secondary analysis to compare HIV incidence by different levels of intervention intensity will be undertaken, adjusting for potential confounding. Incidence rates will be compared both within the intervention communities (e.g. comparing those with limited exposure to those who had reasonable exposure to the intervention) and to control communities where intervention was not implemented. Rates of losses to follow-up, and the characteristics of those lost, will be examined to investigate bias.
Similar methods will be used to analyse the cumulative incidence of HSV-2 and unintended pregnancies. We will also examine the factors associated with the acquisition of HIV, using logistic regression on individual-level data. Terms will be included in the model to adjust for communities.
Sample size justification
The sample size calculations made the following assumptions: 15 communities per arm from which an average of 220 Form 2 students would be enrolled, 4% of students would be excluded from analysis because they are HIV-infected at baseline and 25% would be lost to follow-up by the end of the trial.
Assuming that the cumulative incidence of HIV in the ‘deferred intervention’ arm would be 4% over the 4 years of the study (National AIDS Council, Ministry of Health & Child Welfare & Zimbabwe National Family Planning Council 2002) and a coefficient of variation of cumulative incidence between communities within strata of k = 0.2, we will have 80% power to detect as statistically significant (at the 5% two-sided significance level) a cumulative incidence of 2.4% or lower in the ‘early intervention’ arm (reduction of 40%) (Hayes et al. 2000). As cumulative incidence of HSV-2 and unintended pregnancy will likely be higher than 4%, we will have greater power to detect a smaller difference in these factors.
There were 7885 Form 2 students eligible for inclusion in the cohort. Parents/guardians of 91.5% (n = 7215) of these agreed to their child’s participation, 8.1% declined (mostly because of religious beliefs around blood draw) and 0.2% of parents were not contactable. 94.1% of young people whose parents had given consent agreed to participate (n = 6791), giving an overall consent rate of 86.3%.
Baseline characteristics of participants
The baseline characteristics of study participants are shown in Table 2. Boys were 6 months older than girls (P < 0.001). While both girls and boys reported relatively low levels of knowledge about sexual matters, girls reported poorer knowledge than boys (P < 0.05). For example, 78% girls compared with 71% of boys (P < 0.001) said they knew little or nothing about preventing pregnancy (including two girls who were pregnant); 70% of girls compared with 62% of boys (P < 0.001) said they knew little or nothing about HIV [including 32 (63%) of those infected]; and 81% of girls compared with 68% of boys (P < 0.001) said they knew little or nothing about other sexually transmitted infections. Girls were also significantly less likely than boys (P < 0.001) to report that they would find it easy to access condoms (7% compared with 24%), or to use a condom in the future (17% compared with 46%), and to tell their partner he must use condoms (24% compared with 49%).
Table 2. Baseline characteristics of respondents by intervention arm
Early (n = 1777)
Deferred (n = 1744)
Early (n = 1604)
Deferred (n = 1666)
19 years and over
Sex of head of household
Both parents alive
Both parents dead
Overall SES score (quintiles)
SES1 – Lowest
SES5 – Highest
Ever had sex
Number of lifetime partners
Only 1 partner
3 or more partners
Age at 1st sex
Under 15 years
15 or older
Ever used a condom
Used a condom at last sex
Knowledge & Attitudes
Condoms can prevent HIV
No or don’t know
Condoms can prevent pregnancy
No or don’t know
Self Esteem & Mattering
I am a failure
Neither agree or disagree
I do lots of things well
Neither agree or disagree
People do not care what happens to me
Neither agree or disagree
Twenty (0.6%) boys and 31 (1.0%) girls had HIV-1 antibody detected (Table 3). Only 12 had HSV-2 antibody (0.2%). Four (0.12%) girls were pregnant. There was excellent balance between early and deferred intervention arms in terms of rates of HIV-1 infection and other behavioural and socio-demographic variables.
Table 3. Baseline survey results: prevalence of HIV and HSV-2 infection by gender and pregnancy in girls only
HIV-positive n (%) [95% CI]
10 (0.6) [0.3–1.0]
15 (0.9) [0.5–1.6]
25 (0.7) [0.5–1.1]
10 (0.6) [0.3–1.1]
16 (1.0) [0.6–1.6]
26 (0.8) [0.5–1.1]
20 (0.6) [0.4–0.9]
31 (1.0) [0.7–1.4]
51 (0.8) [0.6–1.0]
HSV-2-positive n (%) [95% CI]
2 (0.1) [0.0–0.3]
2 (0.1) [0.0–0.3]
4 (0.1) [0.0–0.2]
2 (0.1) [0.0–0.3]
6 (0.4) [0.1–0.6]
8 (0.2) [0.1–0.4]
4 (0.1) [0.0–0.2]
8 (0.2) [0.1–0.4]
12 (0.2) [0.1–0.3]
Pregnant (%) [95% CI]
3 (0.18) [0.00–0.39]
1 (0.06) [0.00–0.19]
4 (0.12) [0.03–0.31]
Overall, 2.9% of the girls reported being sexually active compared with 11.9% of the boys (P < 0.001) [median age at first intercourse 12 years for both males (range 3–21 years) and females (range 3–16 years)]. On the whole, 1% of female respondents consistently reported having had sex and 86% consistently reported not having sex. Among male respondents, 5% consistently reported having sex and 67% consistently reported not having sex. Girls appeared more likely to under-report their sexual activity. Overall they had higher rates of HIV infection but reported less sexual activity. Additionally, none of the four pregnant girls reported that they had sex in the questionnaire.
Thirty-five per cent (95% CI: 33.9–36.2%) of the cohort reported having lost one or both parents, with 9.5% (95% CI: 8.8–10.2%) indicating that they had lost both, 19.9% (95% CI: 19.0–20.9%) reporting having lost their father only and 5.6% (95% CI: 5.0–6.1%) having lost their mother only. Orphaned participants were at increased risk of HIV [age–sex and clustering adjusted odds ratio 3.4 (95% CI: 1.7–6.5; P < 0.001)]. Majority of the HIV-positive participants reported having lost one or both parents (65%; 95%CI: 51–78%).
Based on the individual question items relating to self-esteem, it appears that young women generally have higher self-esteem than men (P < 0.05 for all items – data not shown). Likewise, females were more likely to perceive that they mattered to people (P < 0.05 for all items).
In this paper, we outline the design of a community randomised trial of a multi-component community-based adolescent HIV prevention intervention. The study is being conducted in rural Zimbabwe. We intend to evaluate the effectiveness of the intervention by measuring its impact on HIV incidence, HSV-2 incidence and prevalence and incidence of unintended pregnancy among a cohort of young people who were enrolled in Form 2 of secondary school in 2003. In addition we will measure impact on reported behaviour, self-efficacy and self-esteem and on attitudes to gender issues. Over the first 3 years of the study, the intervention has evolved from a primarily school-based intervention supported by an out-of-school youth programme and a programme for parents to using a broader community-based approach. The intervention has evolved in this way in response to the economic climate in Zimbabwe which has impacted severely on rural secondary school attendance. Importantly, while the site of intervention delivery has changed, the target population (in- and out-of-school youth), the delivery method (carefully selected, recruited and trained PPEs) and the theoretical basis for the intervention have remained constant, thus maintaining the integrity of the intervention (Hawe et al. 2004) and thereby its replicability and generalisability.
It is encouraging to note that our intervention fulfils many of the criteria outlined for potentially effective community-based interventions in the systematic reviews of HIV prevention among young people in developing countries (United Nations General Assembly Special Session on HIV/AIDS 2001). A cornerstone of the intervention is that its facilitators are very carefully selected, trained and supported. The facilitators who implement the intervention with parents go through a similarly rigorous selection process and undergo 2 weeks of residential training on the course materials and facilitation skills. Both groups are supported by the project intervention team who provide ongoing mentoring. Although the initial recruitment and training of facilitators is time consuming (and accounts for the bulk of intervention delivery costs), it is a manageable and easily replicable activity that then forms the backbone of the intervention.
Baseline data suggest that there is good balance between trial arms in terms of both biological and behavioural markers at the start of the study. The levels of HIV, HSV-2, pregnancy and reported sexual behaviour in both arms were remarkably low before the intervention was implemented, suggesting that the intervention started before the majority of young people in our target population became sexually active. HSV-2 antibody was detected using DBS samples. In our pre-study evaluation of the assay, we found that sensitivity of DBS specimens was 91% compared with sera suggesting that HSV-2 sero-prevalence may be slightly underestimated in our baseline survey. Specificity was high compared with sera at 99.3% (data not shown). While communities are well separated from each other, the high rates of mobility in Zimbabwe make it impossible to prevent the possibility of contamination between communities. While other HIV prevention programmes may be implemented within study communities, all prevention activities are coordinated by the District AIDS Action Committee (DAAC). Project staff work closely with the DAAC and in this way we hope to minimise the risk of differential implementation of other programmes by study arm. All non-study HIV prevention activities that take place in study communities are recorded.
As others have found when using quantitative survey methods to measure sexual behaviour in young people, there was considerable inconsistency in reporting (Plummer et al. 2006). While young women were more consistent in their reporting, they were also more likely to under-report the extent of their sexual experience; for example, none of the four women who were pregnant reported having had sexual intercourse. This finding underscores the importance of using externally valid outcomes measures when assessing the effectiveness of sexual behavioural interventions especially as reporting bias may be differential in the context of these interventions. Nonetheless, understanding behaviours and their determinants is important for designing and implementing appropriate and effective reproductive health interventions and finding methods to improve the validity of these data is critical.
After many years of very high prevalence of HIV, data suggest that the tide has turned and that rates of infection in Zimbabwe are now falling (UNAIDS 2005). Evidence from one study in eastern Zimbabwe indicates that this is likely the result of behaviour change (specifically reduction in casual partners and delay in sexual debut) (Gregson et al. 2006). Rates of HIV among our study participants were lower than we were anticipating based on our pilot study data (Cowan et al. 2002) and data from a large national representative survey of young people conducted in 2001 (National AIDS Council, Ministry of Health & Child Welfare & Zimbabwe National Family Planning Council 2002) This encouraging trend in HIV incidence may have implications for the power of this study to detect a difference in cumulative HIV incidence between arms, although it is likely that the study will still be adequately powered to detect a difference in cumulative incidence of HIV when combined with incidence of HSV-2 and unintended pregnancy.
Data from this study suggest that HIV is not distributed uniformly across the population. Orphans had higher rates of HIV than non-orphans. Household structure also seemed to be important. A nested qualitative study has been conducted to explore this vulnerability in more depth and has been linked to expanded data collection as part of the interim survey data so that these causes of vulnerability can be quantified. These data will be important for more effectively targeting HIV prevention in this rapidly changing environment.
Few rigorous evaluations of HIV prevention in young people have been conducted in Africa. The MEMA kwa Vijana trial is evaluating the impact of a package of interventions on HIV, other STIs, pregnancy, reported sexual behaviours, reported attitudes towards sexual risk and knowledge of sexual and reproductive health among young people (Hayes et al. 2005; Ross et al. 2007). In South Africa, Jewkes et al. (2006) have evaluated the effect of the Stepping Stones intervention on HIV and HSV-2 incidence.
In summary, this is one of the few rigorous evaluations of a community-based HIV prevention intervention for young people to take place in southern Africa. The intervention being evaluated targets both young people and their parents and adults in the community, and aims to change behaviour at an individual level in addition to changing societal and cultural norms surrounding adolescent sexuality more broadly in order to reduce their environment of risk. The UNAIDS Interagency Taskforce on Youth has advised against further expansion of community-based HIV prevention in youth without more rigorous evaluation of the processes and outcomes of such programmes. We believe that this study will make an important contribution to the evidence base required by policy makers on the likely effectiveness and costs of such programmes.
The study is funded through the National Institute of Mental Health.