The most efficient use of resources to identify those in need of antiretroviral treatment in Africa: empirical data from Côte d'Ivoire's Drug Access Initiative

Abstract

Objective: To describe the cost and outcome associated with the use of CD4 cell count and viral load tests as part of screening strategies to identify persons eligible for subsidized antiretroviral therapy (ART) in Côte d'Ivoire.

Methods: Empirical data from the Drug Access Initiative in Côte d'Ivoire (DAI-CI) were used to describe the laboratory cost of patient screening using sequential clinical staging, CD4 cell count, and viral load and the proportion of screened patients identified as eligible for ART. We also estimated costs modelling a parallel screening algorithm, across a range of laboratory costs and with current international recommendations to assess treatment eligibility. Benefit was defined as being found eligible for ART.

Results: Of the 2138 HIV-positive, ART-naive, adults who presented to the DAI-CI between July 1998 and July 2000, median CD4 cell count was 172 x 10(6) cells/microl. DAI-CI criteria identified 2057 (96%) of these persons eligible for antiretroviral treatment. In a serial screening algorithm, 75% were eligible by CDC clinical stage B or C; 18% by CD4 cell count less than 500 x 10(6) cells/microl; and an estimated 3.9% by a viral load greater than 10 000 copies/ml. Use of the current US recommendations and a serial algorithm would have resulted in 1977 (92%) persons eligible for ART: 75% by CDC clinical stage B or C; 15% by CD4 cell count less than 350 x 10(6) cells/microl (including 8% < 200 x 10(6) cells/microl); and an estimated 3.6% due to viral load greater than 55 000 copies/ml. Using DAI-CI criteria and heavily subsidized laboratory test costs, the addition of CD4 cell count to clinical criteria cost US dollar 50 (serial algorithm) and US dollar 203 (parallel algorithm) to identify each additional eligible person. Modelling current recommendations with a serial algorithm, CD4 cell count cost an average US dollar 62/eligible person (US recommendations) and US dollar 109 (WHO recommendations). The addition of viral load cost between US dollar 108 (serial algorithm DAI) to US dollar 1700 (parallel algorithm DAI) to identify each additional eligible person.

Conclusion: In the African context of scarce resources and the huge unmet demands for voluntary HIV testing and for ART, simple screening strategies are needed to identify those most in need of ART. Health personnel should be trained to identify and refer clinically symptomatic persons. Viral load testing is of high cost and dubious benefit and should not be part of screening algorithms for initiating ART.