Skip Navigation


Global Projects


Cost-Effectiveness of Antiretroviral Therapy Adherence Interventions

Cape Town, South Africa


As programs to provide highly active antiretroviral therapy (HAART) are scaled up in sub-Saharan Africa, achieving high rates of medication adherence will be necessary to maximize virologic and clinical outcomes. However, the cost-effectiveness of interventions aiming to increase adherence, particularly in the African setting, remains unknown. Development of a tool to estimate the cost-effectiveness of HAART adherence interventions could assist governments and donor agencies in deciding whether or not to implement specific programs aimed at increasing medication adherence. The aims of this project is: (1) To characterize the relationships between medication adherence, HIV disease progression, and health care costs in a cohort of patients participating in a South African private-sector HIV/AIDS management program; (2) To develop a decision tool for evaluating the cost-effectiveness of potential HAART adherence interventions in sub-Saharan Africa; (3) To apply this tool toward the evaluation of a specific intervention, namely a programmable prompting reminder device, as an adherence intervention. The study setting will be AfA, the largest private sector HIV/AIDS disease management program in Africa (approximately 50,000 registered patients from nine countries), based in Cape Town, South Africa. For specific aim #1, ART adherence will be assessed using monthly pharmacy claim data as reported to AfA; generalized estimating equations will be used to estimate the impact of adherence on HIV disease progression, as measured by CD4 count and viral load. Costs will be measured from the provider’s perspective on an individual patient basis and will include consultations, hospitalizations, investigations, and medications (both ART and non-ART). Total costs will be estimated for pre-defined strata of CD4 cell count, adherence, viral load, and duration of HAART and will be compared across strata using pairwise comparisons (Student’s t test), trend tests, and global tests (Wilcoxon rank sum test and ANOVA). For specific aim #2, a Markov-based decision analysis will be constructed, using the relationships defined in the first aim. The primary outcome of this analytic model will be the incremental cost-effectiveness ratio of interventions that aim to increase HAART adherence in the AfA patient population, expressed in terms of lifetime costs, life years and quality adjusted life years, as a function of the efficacy of adherence intervention. Cost-effectiveness will be reported as a function of program cost and expected efficacy (i.e., percentage increase in adherence), and results will be compared to those from widely-used models built on data from the developed world. This model will be made available for use by investigators and policymakers studying adherence interventions in other developing-world patient populations. For specific aim #3, published reports and local data will be used to estimate the expected program costs and increase in adherence for a programmable prompting reminder device. The model developed in aim #2 will be used to estimate the incremental cost-effectiveness ratio for this intervention, and first and second-order Monte Carlo simulation of model parameters will be used to characterize both the uncertainty of this estimate and its dependence on various model assumptions.


  • Start Date: 
  • End Date: