Clinical decision analysis of HIV screening.

Abstract

BACKGROUND: Various health organizations, and government agencies have developed differing recommendations to address the issue of who should be tested for HIV (human immunodeficiency virus). This study used decision analysis to evaluate approaches to testing for HIV infection in a variety of clinical situations. METHODS: Four decision tree branches were evaluated to determine the relative importance of HIV seroprevalence, test sensitivity, test specificity, and other factors in four testing scenarios: 1) no HIV testing of anyone, 2) mandatory HIV testing for all health care workers and hospitalized patients younger than age 65, 3) testing of persons determined to be at high risk for HIV infection based on risk assessment, and 4) voluntary HIV testing and reporting. RESULTS: No HIV testing results in a situation in which HIV status is unknown for all patients and health providers; issues of test sensitivity and specificity are not pertinent. Mandatory testing of health care workers and hospitalized patients, and testing only of high-risk persons, is heavily dependent on assumptions regarding HIV disease prevalence and on test sensitivity and specificity. Voluntary testing is more dependent on political decisions regarding the consequence of positive tests than it is on HIV prevalence or test characteristics. CONCLUSIONS: Policies regarding HIV testing are heavily dependent on the prevalence of HIV seropositivity, the sensitivity and specificity of HIV blood tests, and public policy.