The influence of assay sensitivity and specificity on error rates in three anti-HIV testing strategies

Abstract

Tests for anti-HIV can be used (1) to establish for clinical or other reasons whether an individual has developed antibody to the virus, (2) to screen donated blood for evidence of infection and (3) to provide estimates of the prevalence of infection in a population. It is suggested that different testing strategies, using one or more distinct anti-HIV assays, are appropriate in these three situations. This paper examines the expected accuracy of the results that would arise if these strategies were implemented to test populations with various anti-HIV prevalences using assays with selected sensitivities and specificities. The expected outcomes are shown in tabular form for chosen values of the variables.