| In this dissertation, I undertake three empirical analyses using data from the Africa Centre for Health and Population Studies, which is located in the Hlabisa subdistrict of northern KwaZulu-Natal, South Africa. In the first analysis I assess if antiretroviral therapy (ART) usage in the household is associated with a reduction in individual HIV acquisition risk. To my knowledge, this analysis is the first attempt to quantify the preventive impact of a public sector treatment program based in a rural community with poor knowledge and disclosure of HIV status, frequent migration, late marriage, and multiple partnerships. I argue in the second analysis that efforts to optimize the preventive efficacy of ART in South Africa and elsewhere will be critically dependent on the ability of the public health sector to initiate and then keep HIV-infected patients on treatment. Here, I examine the socio-demographic and structural factors that are associated with poor or imperfect adherence to antiretroviral medications, which can be obtained for free at multiple health-care centers within the study area. The third analysis continues this work by examining the diagnostic performance and cost-effectiveness of two monitoring strategies---CD4 and HIV-1 viral load count testing---to detect poor patient response to ART. My approach is based on the idea that the cost-effectiveness of a treatment monitoring strategy is a function of its predictive performance. |
