OBJECTIVES: Human deficiency virus 1 (HIV-1) is the principal causative agent of acquired immune deficiency syndrome (AIDS). With its high prevalence in South Africa, and its transmissible characteristics, HIV-1 presents unique modelling challenges for cost-effectiveness analysis. The study investigates the methods and results of previously published economic models evaluating treatments for people with HIV-1 in South Africa, to support the planning of future modeling studies.

METHODS: A targeted literature review of published health economic models of HIV-1 in South Africa was conducted through Medline via Ovid, and related modelling data from all available publications were extracted. Model structure, data sources, population characteristics, health states, results and other model characteristics were evaluated.

RESULTS: Fourteen studies met the requirements for inclusion; thirteen cost-effectiveness analyses and one comparative cost analysis. Within the cost-effectiveness analyses, microsimulations, markov models, and deterministic compartmental model types were used in nine studies. In four studies, decision trees or other models were utilized. Efficacy (reported in 11 studies) and utility (reported in five studies) data were obtained from publicly available sources and costs were sourced from local healthcare agencies. Ten studies reported analyses from the healthcare payer perspective, three included societal perspective and one did not report any perspective. Health states reported were limited to two studies and were based on CD4 count (CD4 >500, 351 – 500, 201 – 350, < 200) and death. All research employed a 3% discount rate, but time horizons varied.

CONCLUSIONS: There was a general lack of data source reporting in the identified analyses, and a wide variety of approaches were observed. However, Markov and micro-simulation were the most used model types, with data sources ranging from publicly available to specific local health care centers. Standardized methodology for developing cost-effectiveness models to assess HIV-1 in South Africa would be beneficial to ensure results are economically valid.