Thirty-month quasi-experimental evaluation follow-up of a national primary school HIV intervention in Kenya

Abstract

School-based programmes to combat the spread of HIV have been demonstrated to be effective over the short-term (i.e. 6–24 months), but few studies have addressed the sustainability of such interventions. Primary School Action for Better Health (PSABH) is an intervention delivered in upper primary-school grades in Kenya using regular classroom subjects and co-curricular activities. This paper reports results of an evaluation to assess the sustainability of PSABH and its impacts three years after teachers were trained in its delivery. Twenty primary schools participated in a pre-post evaluation of PSABH in Kenya over three consecutive school years. Surveys were completed by teachers (n = 40), pupils (n = 3403), and Quality Assurance Officers (n = 15); focus groups were held with pupils, and in-depth interviews with teachers. Teachers continued to deliver programme components three years after they were trained. Gains demonstrated in pupil knowledge, attitudes and risk-reducing sexual behaviours after one-and-a-half years of programme implementation were replicated in the third year of the intervention with additional gains in attitudes related to condoms and girls' reported use of condoms. These results support the conclusion that PSABH and its impact on student knowledge, attitudes and behaviour is sustainable across at least three school years and three student cohorts.

Acknowledgements

The research and intervention development were funded by the DFID UK. CfBT Education Trust, Kenya was responsible for oversight of the entire project. Janet Wildish and Mary Gichuru developed and delivered the intervention reported on in this paper while they were employed by CfBT Education Trust, Kenya. Eleanor Maticka-Tyndale designed and conducted the evaluation of the intervention reported on here. This was independent of the intervention development and delivery performed by her co-authors. Neither CfBT Education Trust nor DFID are responsible for any of the analyses or conclusions presented in this paper.

Notes

^1. See Maticka-Tyndale, Wildish, and Gichuru (2007) for a full description of the original sampling and evaluation methodology.