Table 3 Details of studies included in review

Agha 2002, Agha and Van Rossem 2004 [9,30]

Secondary (boarding) schools in Lusaka (urban), Zambia; N = 913; N₁ = 759; mean age = 17.9; social class – 'somewhat wealthier backgrounds'; N₂ = 416.

Objective – to promote safer sex practices; theoretical basis – not specified; time setting – school hours; duration – 1 hour 45 minutes; main media – factual information and discussion, drama skits and leaflets; contents – transmission of HIV, prevention of HIV, promoting abstinence and condom use; facilitator – peers; control group-peer led water purification programme

Study type – Quasi-experimental longitudinal panel; (random) allocation – yes, 3 intervention & 2 control schools; follow-up assessment – 1 week and 6 months; analysis – multivariate logistic regression.

1) High attrition rate due to the Grade 12 at baseline, having passed out of school by second follow-up.

2) Sexual behaviours not expected to be different at 1 week intervention as report at this time a reflection of pre-intervention practice.

Fawole et al. 1999 [23]

Mixed-sex senior secondary school in Ibadan, (urban) Nigeria. N = 450; N1 = 433; social class-'poor'.

Objective – to evaluate knowledge attitude & sexual risk behaviours of students after a comprehensive health education intervention; theoretical basis-not specified; time setting – not specified; duration – 6 weeks (2 to 6-hour sessions weekly); main media – lectures, films, role plays, stories, songs, debates, essays; content – methods of prevention of HIV, demonstration of proper use of condoms; facilitator – 1 community physician and 2 trained teachers; control group – no programme.

Study type – (not stated) controlled longitudinal trial; (random) allocation – none, 2 intervention schools, unspecified number of control schools; follow-up assessment – 6 months; analysis – frequency distribution of variables, Chi² and ANOVA.

1) School principal approved demonstration of condom use.

2) Positive outcomes, though not statistically significant in condom use and recent history of STD.

3) Participants opinions regarding programme was assessed.

Fitzgerald et al. 1999 [25]

Grade 9 or 11 secondary school students in Omusati and Caprivi regions of Namibia; N = 515 of which mean age = 17 (range 15–18), median grade = 11; social class – 'class standing' = 46%; N₂ = 452.

Objective-to evaluate an intervention-'My Future Is My Choice' (based on one – 'Focus on Kids' that had been successful in reducing adolescent HIV risk in the U.S.) among adolescents residing in Namibia; theoretical basis – protective motivation theory; time setting – after school hours; duration – 7 weeks made of 14 (2-hour) sessions, main media – narratives, games, facts and exercises and questions and discussions; contents-basic facts about reproductive biology, HIV/AIDS, alcohol, substance abuse and violence within a relationship, communication skills across genders and age differences and a frame work for decision making; facilitator-volunteer teacher and out of school youth; control group – no programme.

Study type-randomised trial; (random) allocation – yes, of students from 10 schools; follow-up assessment – 6 months; analysis-Chi² and ANOVA.

1) Attrition rate higher (statistically significant) among control youth than intervention youth.

2) Other risky behaviour like alcohol use and weapon carrying assessed.

Harvey et al. 2000 [4]

Standard 8 secondary school students in 5 districts (4 rural, 1 urban) of KwaZulu Natal, South Africa; N = 1080 of which mean age = 17.6 (range 13–29); social class – 'average annual income per household ≤ $150.'

Objective – to evaluate effectiveness of a high school drama-in education programme 'DramAide', to increase AIDS awareness; theoretical basis-not specified; during school hours; duration – unspecified; main media – 1^st phase of play presentation by DramAide team, 2^nd by running drama workshops with students and teachers, and 3^rd culminating in an open day of drama, song, dance, poetry and posters all prepared and presented by students; contents – 'issues surrounding HIV/AIDS'; facilitator – DramAide team of nurses, teachers and actors; control group – "booklet intervention schools" students given a 10-page booklet in Zulu about prevention and transmission of HIV/AIDS.

Study type – RCT; (random) allocation – yes, 7 schools to each of the groups; follow up assessment – 6 months; analysis – linear regression modelling.

1) Areas of questionable validity and relevance noted by the medical research workers at the Public Health Laboratory Service, U.K relative to questionnaire design.

2) The booklets were written in Zulu whereas questionnaires were administered in English.

3) Difference in intensity, supervision and monitoring of intervention between the two study groups.

James et al. 2005 [31]

Grade 11 secondary school students in the Midlands (a rural and urban, largely Zulu speaking) district of the Province of KwaZulu-Natal, South Africa. N = 1168; N₁ = 867 N₂ = 722; social class – not specified.

Objective – to test the effects of a systematically developed photo-novella ('Laduma') on knowledge, attitudes, communication and behavioural intentions with respect to sexually transmitted infections(STIs); theoretical basis – health promotion and social learning; time setting – during school hours; duration – approximately 1 hour; main medium – print media, a comic called Laduma; contents – factual information about STIs, protection against STIs including illustration of correct condom use; facilitator – printed material; control group-received normal(not AIDS related school lessons) at time of intervention, then Laduma at end of study.

Study type-RCT; random allocation – yes, 10 schools to control group and 9 schools to intervention group; follow-up assessment – 3 weeks from baseline (time of intervention delivery) and 6 weeks post intervention; analysis – Chi² and logistic regression analysis.

1) Larger percentage of participants spoke Zulu, but intervention was delivered in English.

2) High attrition rate (38.2%), attributed to impending school examination.

3) The time to evaluation of 6 weeks too short to validate measured behaviour change.

4) Bias in reporting of sexual behaviour with only changes in favour of intervention regarding condom use being reported. Change regarding abstinence (seemed opposite to desired outcome) was not reported.

Klepp et al. 1994, Klepp et al. 1997 [21, 29]

Grade six and grade seven primary school pupils in Arusha and Kilimanjaro regions (mixed urban and rural) Tanzania. N = 2026; N₁ = 1785; N₂ = 814 social class – not specified.

Objective – to test effects of an HIV/AIDS education programme (NGAO); theoretical basis – theory of reasoned action and social learning; time setting-during school hours; duration – 20 hours over 2–3 months; main media – instruction by teachers, crafts(like posters), plays, role plays, poetry, group discussions and theme T-shirt; contents – communication about AIDS, transmission and prevention of HIV, abstinence and reduced intention to be sexually active and caring for PWA; facilitator – 2 teachers and 1 local health worker per school; control-group – no programme during study but to receive 'NGAO' at end of study..

Study type – RCT; random allocation – yes, 12 control schools and 6 intervention schools; follow-up assessment-6 months and 12 months; analysis – ANOVA.

1) Attrition higher among control schools (though explained to be incidental),

2) Attrition greater among semi urban than rural schools; catholic than protestants; persons with lower baseline level of AIDS information and persons with subjective norms in favour of becoming sexually active.

3) Positive attitude changes regarding sexual intercourse improved in the intervention group (though not reaching statistical significance).

4) Difference between N₁ and N₂ due to initial 7th grade students having passed out.

Kuhn et al. 1994 [27]

A high (secondary) school in Cape Town (urban), South Africa. N = 567 of which, mean age = 18 years (range 12–30); N₁ = 482; social class – 'economically disadvantaged'

Objective – to raise awareness about AIDS; theoretical basis – not specified; time setting-during and after school hours; duration – 2 weeks; main media – structured classroom information, role plays, games, structured group work, videos and leaflets; contents – 'key AIDS messages', condoms made available; facilitator – teachers, assisted by nurses; control group – no programme delivered.

Study type – pilot study; random (allocation) – unclear, 1 intervention and 1 control school, follow-up assessment – immediate; analysis – Chi² tests.

1) Programme sometimes led to disruption in normal school activities.

2) Full parental involvement despite taboos.

3) HIV positive individual gave talk to school staff.

4) Condoms were actually made available.

5) Intervention integrated into an already existing 'language curriculum'.

6) Questionnaires were translated into Xhosa (the mother tongue).

7) Some material described as racist.

Munodawafa et al. 1995 [24]

5 rural secondary forms 2 and 3 (grade 9 and 10 U.S A. equivalent) in Masvingo Province (rural) Zimbabwe. N = 315; N₁ = 285; social class – not specified.

Objective-to assess the impact of health instruction on knowledge targeting prevention of STD, AIDS and drug abuse and to assess utility of student nurses as health instructors; theoretical basis – not specified; time setting-during school hours; duration – 7 weeks (2 sessions per week, 40 minutes per session); main media – health lessons; contents – transmission and prevention and psychosocial issues of STD and AIDS, responsible sexual behaviour and problem-solving and decision-making strategies and others relating to drug abuse; facilitator – student nurses; control group – no programme delivered.

Study type – quasi-experimental; (random) allocation – none, 2 schools to control group and 3 to intervention group; follow-up assessment – immediate; medium of assessment – 'inventories'; analysis – ANCOVA and Chi².

1) Study not very focused relative to sexual health intervention as emphasis more towards promoting use of student nurses as facilitator of intervention. Greater part of study evaluation is the perceived performance of the student nurses in intervention delivery.

2) No attempt seems to have been made to validate the questionnaires. Item on 'possibility of getting AIDS by giving blood' actually yielded the reverse of expected outcome.

3) Alcohol, marijuana and tobacco use, part of the content of programme.

Rusakaniko et al. 1997 [18]

All types of secondary schools (boarding and day schools, mixed- & single-sex schools) in rural & urban areas Zimbabwe. N = 1673; N₁ = 1568 and N₂ = 1589; school classes – forms 1 to 6; social class – unspecified.

Objective- to determine the impact of an intervention package on knowledge of various health issues; theoretical basis- not specified; time setting-not specified; duration- not specified; main media-lectures, videos, and leaflets and pamphlets; contents-reproductive biology, STD and AIDS, responsible sexual behaviour and unwanted pregnancy and contraception; facilitator-teachers and printed material; control group – No programme delivered.

Study type – (wrongly described as an RCT); (random) allocation-none, purposive sampling of schools and random sampling of students; follow-up assessment – 5 months and 7 months; analysis-linear trend and Chi² test.

1) Stated to be RCT, had random sampling rather than randomisation.

2) Religious bias noted.

3) Not quite focused in terms of study population. Though schools were stratified for representatives, results are analysed as if study sample is homogenous.

4) Some documented 'decrease in knowledge level' following intervention in relation to HIV/AIDS transmission and pregnancy

7) Aims and objectives not clear throughout the study.

Stanton et al. 1998 [26]

10 secondary schools in Caprivi and Omusati districts of Namibia. N = 515; median age = 17 and median grade= 11; N₁ = 452; N₂ = 379; N₃ = 359; social class – not specified.

Objective- to evaluate HIV risk-reduction intervention 'My Future is My Choice', based on 'Focus on Kids' curriculum developed for African-American youth in US; theoretical basis – social cognitive theory; time setting – after school hours; duration – 7 weeks (14 × 2-hourly sessions); main media – narratives, games, facts & exercises, questions & discussion; contents – basic facts about reproductive biology, HIV/AIDS and other risk behaviours includ. alcohol, intra-relationship violence, communication skills and framework for decision-making; facilitator – volunteer teacher & out-of-school youth; control group – delayed-control condition.

Study type – RCT; (random) allocation – yes of individuals; follow-up assessment – immediate, 6 months and 12 months; analysis – Chi ² analysis.

1) Attrition rate significantly higher in control group, older respondents, and in Caprivi region.