Socio-economic factors predicting drop-out of paediatric HIV prevention programming among adolescent mothers in South Africa

Sub-Saharan Africa has the highest rates of both HIV infection and unintended adolescent pregnancy worldwide; two challenges in which adolescent girls and young women bear the greatest socio-economic and health consequences. In South Africa, which is home to the largest population of people living with HIV globally, adolescent women still account for nearly a third of all new infections, due to several social, relational, and biological risk factors such as age-disparate relationships, transactional sex, and intimate partner violence. Critical to the control of the HIV epidemic is the prevention of mother-to-child transmission (PMTCT), as this causes several adverse health and socio-economic outcomes for mothers, their children, and society more broadly. Adolescent mothers in South Africa have up to three times higher rates of mother-to-child transmission compared to adults, making them a critical, yet often overlooked, priority population for prevention programming intervention. Thus, there exists an urgent need to understand where, when, and why South African adolescents and their babies drop out of PMTCT services, and where these services can be improved to ultimately eliminate paediatric HIV infection. The proposed DPhil research aims to answer these questions by quantifying retention in PMTCT programming among adolescent mothers pre- and post-COVID-19-onset, and determining which socio-economic factors predict poor paediatric HIV prevention outcomes. To this end, the research will draw upon data from a large, multi-wave cohort of adolescent mothers and their children living in the Eastern Cape Province, South Africa, enrolled in the HEY BABY longitudinal study (a research collaboration between the Universities of Cape Town and Oxford). This data is unique in three ways: i) the inclusion of over 1000 mother-child pairs, with linked data for both, ii) its longitudinal nature, spanning three waves pre- and post-onset of the COVID-19 pandemic, and iii) multiple data sources, including self-report interviews, patient files, and medical laboratory records for each mother-child pair. This rich data allows for the use of more advanced statistical methodology, including marginal effects modelling, classification and regression tree analysis, population attributable fraction analysis, and path analysis structural equation modelling, all of which will be employed to determine which stages of PMTCT programming should be improved, and which modifiable socio-economic factors have the most significant impact on paediatric HIV outcomes. This will provide evidence-based policy and programming guidance, identifying key areas where government and funders can effectively prioritise their investments and intervention efforts.