The economics of antiretroviral therapy in South Africa: The role of budget impact modelling in changing policy

Abstract

South Africa is home to the largest number of people living with HIV in the world, as well as the world’s largest public-sector antiretroviral treatment (ART) programme. Despite the absolute majority of it being domestically funded, planning and budgeting for this programme has in the past been based on assumptions regarding target population and unit cost and on politically expedient coverage targets. The aim of this thesis was to improve on this situation by developing a budget impact model that could project the number of adults and children in treatment based on sound epidemiological methods, and calculate the cost of treating them based on the results of detailed bottom-up cost analyses at relevant clinics and hospitals in South Africa. The thesis describes the methods used in generating the inputs for the model, including the outpatient and inpatient cost of ART provision to adults and children of different ages, and the rates of CD4 cell count development, mortality, loss to follow-up, treatment failure, and regimen switches that were used in the model.The model was used to illustrate the budget impact of a number of guideline changes under discussion by the South African government in 2009/10, including 1.) expanding eligibility to all adults with CD4 cell counts <350 cells/microl, as well as to all TB co-infected and pregnant patients and all children under the age of 12 months regardless of immunological status, and 2.) replacing stavudine in first-line regimens with tenofovir for adults and with abacavir for children, with concomitant changes to second-line regimens. Both 1.) and 2.) had been suggested by the 2009 World Health Organization (WHO) guidelines (“Full WHO guidelines”). A second scenario was considered that expanded eligibility at 350 CD4 cells/microl only to those adults who were pregnant or had active TB at initiation while also replacing the current drug regimens as under 2.) (“New guidelines”). Additional factors with an impact on cost that were considered in the model were a) the introduction of a task-shifting policy that allowed antiretroviral drugs to be prescribed by nurses instead of doctors, and dispensed by pharmacy assistants instead of pharmacists, and b) replacing the existing system of antiretroviral drug procurement via government tenders that favour domestic production with drugs sourced globally at ceiling prices based on the cheapest internationally available price for each drug, including fixed-dose combinations (FDCs) wherever possible. Combining all the inputs, the model showed that while the Full WHO guidelines scenario would increase total cost over the next two mid-term expenditure framework periods (2010/11 to 2016/17) by 35% to USD 19.1 billion, and the New Guidelines scenario by 19%, this increase could be more than offset by introducing the two additional policies. In this case, the total cost of the ART programme under the New Guidelines would be 32% less than under the Old Guidelines without FDCs and task-shifting (taken as government’s revealed willingness-to-pay), while reaching 14% more patients, and implementing the Full WHO Guidelines would still be 23% less costly than continuing the Old Guidelines, while reaching 23% more patients. Based in part on this analysis, the South African government increased treatment eligibility in two steps in April 2010 and in August 2011, introduced the improved drug regimens, established task shifting, and, using the proposed reference price list, negotiated significant drug price reductions for both the December 2010 and the December 2012 ARV drug tender. The budget impact model, named the National ART Cost Model, has been used in budget planning processes for the last seven financial years and, based in part on it, the government’s Conditional Grant for HIV/AIDS, the main vehicle for ART funding, was more than doubled in real terms over this time period. The thesis ends by presenting the results of a cost-benefit analysis of an alternative funding mechanism to public-sector funding, the provision of ART at the workplace, which was found from the company perspective to be cost-saving over no provision of ART, reducing the total cost due to HIV by 5%, and the cost per HIV-infected employee by 14%, over 20 years.