PRIORITIZING SUPPORT FOR INTERVENTION STUDIES THAT AIM TO IMPROVE RETENTION IN ANTIRETROVIRAL THERAPY PROGRAMS IN EAST AFRICA

Purpose: HIV outcomes worldwide are substantially impaired because many people are not retained in care, and the best interventions to promote retention in care are unclear. Our objective was to evaluate whether investing in randomized control trials for retention interventions is worthwhile.

Method: We calculated the expected value of perfect information (EVPI) and expected value of partial perfect information (EVPPI) for a range of willingness to pay (WTP) thresholds using data generated by a dynamic compartmental model for HIV transmission.¹ The EVPI analysis compared each intervention to standard care, and for the EVPPI analysis we compared parameters reflecting two components of a hypothetical outreach intervention (tracing patients with missed appointments and relinking patients back to care). Prior distributions for intervention effect sizes were established based on  a systematic literature review and meta-analysis. Costs were estimated based on the literature. We assumed an effective population of 70,000 with an annual 6% discount rate. Preliminary results are presented.

Result: EVPI for the risk reduction intervention peaked at $31,125,800 when WTP=$5,500 and peaked at $318,480 when WTP=$2,600 for the outreach intervention. For the risk reduction intervention, when applying a WTP of $3,700 (approximately three times the per-capital GDP of Kenya), EVPI=$0 indicating that additional investments may not be worthwhile. For the outreach intervention, as EVPI peaked below the $3,700 threshold, investments in additional information may be worthwhile. For the outreach intervention, EVPPI for tracing peaked at $314,120 when WTP=$2,600 and for relinking peaked at $22,560 when WTP=$2,550, indicating the importance of resolving uncertainty around the tracing parameter relative to relinking.

Conclusion: EVPI quantifies the maximum dollar amount the health care system should invest to generate additional information for a given WTP threshold, and should the cost of conducting additional intervention research be less than the calculated EVPI it is potentially cost effective to conduct further research.² Our analyses indicate that no more than $318,480 should be spent on funding research on outreach interventions when WTP=$3,700, and that generating additional information for tracing be prioritized over relinking. Further, no additional funds should be invested in risk reduction intervention research unless WTP can be raised to at least $5,500. VOI can help bridge the gap from implementation back to research by identifying areas where investments in future research can potentially net the most health.