HIV PARTNER SERVICES: WHEN IS IT WORTH THE ADDITIONAL COST?

Purpose: A critical function of HIV partner services (PS) is the notification of sex and drug-injection partners of a person newly diagnosed with HIV (index case) and offer of HIV testing.  We evaluate whether a healthcare provider should add provider-based PS to an HIV screening program using cost per new HIV diagnosis as measure of cost-effectiveness.

Methods: Based on published studies, we estimated the cost of implementing PS per index case (Cps) and the yield of PS, or average number of newly diagnosed partners per index case (Y). We used published data on the cost of screening HIV positive and negative individuals (Cpoz and Cneg) in an STD clinic.  In our base case, we assumed Cps=$254, Y=0.09, Cpoz=$76 and Cneg=$18. We also consider varying levels of prevalence (P) of undiagnosed HIV infection in the screened population.  We compared the cost per new HIV diagnosis when PS is added to HIV screening to that under screening alone to determine whether PS was cost-effective for a healthcare provider. We identified threshold levels of Y, Cps and P where the addition of PS would result in a greater number of new HIV diagnoses.

Results: We demonstrate that PS is cost-effective relative to screening alone when the yield of PS is greater than the ratio of the cost of PS to the cost of HIV screening only.  That is, iff Y > Cps*P/(Cpoz*P+Cneg*(1-P)). Under our assumptions for costs and yield, PS generates more new HIV diagnoses than screening alone when the prevalence of undiagnosed HIV infection is less than 0.68%. At P=0.68%, the cost per new diagnosis without PS equals the cost per new diagnosis with PS, at $2,711. At a higher P, a greater yield and/or a lower PS cost is required for PS to be a cost-effective addition to HIV screening.

Conclusions: Diagnosing HIV is associated with reduced transmission and benefits for the infected person. Given a fixed budget, healthcare providers should consider whether adding PS to HIV screening strategies will yield more new diagnoses than screening alone. We provide a metric for making that decision. Although not included in this analysis, the potential for PS to identify individuals earlier in the course of their disease could increase its cost-effectiveness.