PURPOSE: CDC guidelines recommend pre-exposure prophylaxis (PrEP) for all persons at high risk of HIV infection. Previous evaluations support this, suggesting that PrEP might be cost-effective when prescribed to HIV-negative men who have sex with men (MSM) with HIV-positive main partners (serodiscordant partnerships). Our objective was to assess the cost-effectiveness of targeting PrEP more effectively based on the antiretroviral treatment (ART) status of the HIV-positive partner.
METHOD: We focused on two strategies: 1) providing PrEP to HIV-negative MSM in serodiscordant main partnerships versus 2) only to HIV-negative MSM with an infected partner not on ART. We linked a set of discrete time-inhomogeneous Markov Chain (DTMC) models to an established microsimulation model of HIV disease (CEPAC). The first DTMC models HIV-negative MSM at risk for HIV infection; the second DTMC tracks MSM if and when they get HIV infection. States in both DTMCs are stratified by partnership status and HIV status of a main partner. The proportion of unprotected sexual acts (12% - 52%, depending on the partner’s HIV status and main/casual partnership type), and annual number of sex acts (80 for main/4 for casual) are state-dependent. HIV-related costs, survival, and utilities are derived from the microsimulation output. We assume PrEP is limited to those who accept PrEP and take it. Baseline PrEP efficacy and annual cost (varied in sensitivity analyses) are 95% (i.e. full adherence) and $13,000. We assume 94% of HIV-positive individuals on ART achieve HIV viral suppression. PrEP costs are captured for 5 years, and transmission, total QALYs and costs are projected until death. We used a cost-effectiveness threshold of $100,000/QALY.
RESULT: The incremental cost-effectiveness ratio (ICER) of targeting PrEP to HIV-negative MSM in a serodiscordant main partnership where the infected partner is not on ART was $900 compared to no PrEP. Expanding PrEP to include all HIV-negative MSM in serodiscordant main partnerships, regardless of ART status, yielded a base-case ICER of $286,000. This ICER remained greater than $100,000/QALY as long as the proportion of unprotected sex acts while on ART was less than 67% or the annual cost of PrEP was above $5500. Conclusion: PrEP is cost-effective if targeted to MSM in serodiscordant partnerships where the HIV-positive partner is not on ART; however, PrEP is not likely cost-effective if targeted to all MSM in serodiscordant partnerships.
CONCLUSION: PrEP is cost-effective if targeted to MSM in serodiscordant partnerships where the HIV-positive partner is not on ART; however, PrEP is not likely cost-effective if targeted to all MSM in serodiscordant partnerships.
