TUAD0201
Share
 
Title
Modelling combination interventions including increased school attendance to prevent HIV among girls of school age in South Africa (HPTN 068)
Presenter
Marie Stoner
Authors
M. Stoner1, J. Ahern2, K. Kahn3,4, R. Twine3, F.X. Gomez-Olive3,4, S. Lippman3,5, A. Pettifor1,3,6
Institutions
1University of North Carolina at Chapel Hill, Carolina Population Center, Chapel Hill, United States, 2University of California, School of Public Health, Division of Epidemiology, Berkeley, United States, 3University of the Witwatersrand, Johannesburg, South Africa, MRC/Wits Rural Public Health and Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, Johannesburg, South Africa, 4INDEPTH Network, Accra, Ghana, 5University of California, Department of Medicine, San Francisco, United States, 6University of North Carolina at Chapel Hill, Department of Epidemiology, Chapel Hill, United States

Background: Combination prevention interventions may be an effective way to prevent HIV in adolescent girls and young women. However, combinations require enormous resources to implement and evaluate. Modelling strategies are a cost-effective way to determine what might work before implementation. We modelled HIV prevention interventions including 1) staying in school or attending 80% or more school days; 2) intervention 1 plus eliminating depression; 3) intervention 1 plus reducing depression by 50%; 4) intervention 2 plus eliminating physical intimate partner violence (IPV) in the last 12 months; and 5) intervention 3 plus reducing IPV by 50%.
Methods: We used data from the main trial period of the HIV Prevention Trials Network (HPTN) 068 study from 2011 to 2015 when girls were of school age. Our study includes 2,328 HIV negative adolescent girls and young women aged 13-20 years in rural South Africa. We used the parametric G-formula where we simulated a population of 10,000 girls from observed data in which confounding from observed covariates was removed. Confounders were age, intervention assignment, socioeconomic status, alcohol use, anxiety, and orphanhood. We then manipulated variables in this population and determined how these changes corresponded with changes in the cumulative incidence of HIV over the study period. Interaction terms were included between all exposures. Confidence intervals were calculated using the standard deviation of results from 200 bootstrap samples with replacement.
Results: The observed risk of HIV over 3.5 years was 4.5%. School attendance, depression and IPV were each independently associated with incident HIV infection. Compared to the observed risk of HIV (4.5%), HIV risk was 4% for intervention 1 (Risk Ratio (RR) 0.90; 95% Confidence interval (CI): 0.84, 0.96; 3.5% for intervention 2 (RR 0.77; 95 % CI: 0.70, 0.84); 3.9% for intervention 3 (RR 0.87; 95% CI: 0.80, 0.94); 3.4% for intervention 4 (RR 0.75; 95% CI: 0.69, 0.82); and 3.6% for intervention 5 (RR 0.80; 95% CI: 0.75, 0.86).
Conclusions: Combination prevention interventions that include increasing school attendance, reducing depression and reducing IPV may be an effective way to reduce risk of HIV in adolescent girls and young women in South Africa.