COST-EFFECTIVENESS ANALYSIS OF HIV PREVENTION FOR NEGLECTED, AT-RISK WOMEN IN VIETNAM: IMPLICATIONS FOR PRIORITIZING BEHAVIORAL INTERVENTION RESOURCES

Purpose:

   HIV/AIDS programs in Vietnam have not included female sexual partners (FSPs) of men who inject drugs (MWIDs). This study evaluated the cost-effectiveness of a rare behavior change intervention to reduce risky sexual behaviors and thereby prevent HIV among these FSPs in Hanoi, Vietnam.

Methods:

   Based on the sexual behaviors and HIV prevalence among MWIDs and their FSPs, the Bernoulli model was used to estimate a reduced annual probability of infection as a result of the behavior change intervention. We used a microcosting approach to estimate the intervention and treatment costs. Our cross-sectional surveys were part of the data sources for the health outcomes. We used a state-transition Markov model to simulate the natural history of HIV infection and subsequent disease progression in 1,500 FSPs from age 25 (intervention until age 50) throughout their lifetimes. From a healthcare payer perspective, our analysis used costs in 2013 US$ and effectiveness in quality-adjusted life years (QALYs). Incremental cost-effectiveness ratios (ICERs) were estimated to compare the intervention with no intervention. ICER robustness was examined in deterministic and probabilistic sensitivity analyses. We considered a willingness-to-pay threshold of three times the gross domestic product per capita per QALY (3xGDPc/QALY) as cost-effective and 1xGDPc ($1,911) per QALY as highly cost-effective.

Results:

   In the base-case analysis, for each FSP, the behavior change intervention extended HIV-free time by 1.39 life-years, increased quality-adjusted life expectancy by 0.307 QALY and yielded an ICER of $1,522/QALY, compared with no intervention. The intervention averted 14.7 HIV infections for the whole cohort. In the deterministic sensitivity analysis, the cost-effectiveness of the intervention was most sensitive to the intervention cost, the discount rate, and the annual probability of HIV infection. The cost-effectiveness acceptability curves showed that the intervention had higher chances of being cost-effective at the willingness-to-pay thresholds of 3xGDPc/QALY and 1xGDPc/QALY (Figure 1). 

Conclusion:

   The behavior change intervention to reduce risky sexual behaviors among FSPs of MWIDs in Hanoi, Vietnam, proved highly cost-effective over no intervention. This study provides evidence that this HIV program results in positive health outcomes and is worth the financial investment. Additionally, as FSPs play an instrumental role in disease transmission to the general population, health program organizers and policy makers should prioritize resources to implement this intervention to all affected individuals.