Cost-effectiveness of couples’ voluntary HIV counselling and testing in six African countries: a modelling study guided by an HIV prevention cascade framework

Introduction: Couples’ voluntary HIV counselling and testing (CVCT) is a high-impact HIV prevention intervention in Rwanda and Zambia. Our objective was to model the cost-per-HIV infection averted by CVCT in six African countries guided by an HIV prevention cascade framework. The HIV prevention cascade as yet to be applied to evaluating CVCT effectiveness or cost-effectiveness. Methods: We defined a priority population for CVCT in Africa as heterosexual adults in stable couples. Based on our previous experience nationalizing CVCT in Rwanda and scaling-up CVCT in 73 clinics in Zambia, we estimated HIV prevention cascade domains of motivation for use, access and effectiveness of CVCT as model parameters. Costs-per-couple tested were also estimated based on our previous studies. We used these parameters as well as country-specific inputs to model the impact of CVCT over a five-year time horizon in a previously developed and tested deterministic compartmental model. We consider six countries across Africa with varied HIV epidemics (South Africa, Zimbabwe, Kenya, Tanzania, Ivory Coast and Sierra Leone). Outcomes of interest were the proportion of HIV infections averted by CVCT, nationwide CVCT implementation costs and costs-per-HIV infection averted by CVCT. We applied 3%/year discounting to costs and outcomes. Univariate and Monte Carlo multivariate sensitivity analyses were conducted. Results: We estimated that CVCT could avert between 54% (Sierra Leone) and 62% (South Africa) of adult HIV infections. Average costs-per-HIV infection averted were lowest in Zimbabwe ($550) and highest in South Africa ($1272). Nationwide implementations would cost between 7% (Kenya) and 21% (Ivory Coast) of a country’s President’s Emergency Plan for AIDS Relief (PEPFAR) budget over five years. In sensitivity analyses, model outputs were most sensitive to estimates of cost-per-couple tested; the proportion of adults in heterosexual couples and HIV prevention cascade domains of CVCT motivation and access. Conclusions: Our model indicates that nationalized CVCT could prevent over half of adult HIV infections for 7% to 21% of the modelled countries’ five-year PEPFAR budgets. While other studies have indicated that CVCT motivation is high given locally relevant promotional and educational efforts, without required indicators, targets and dedicated budgets, access remains low