Reliability, sensitivity and specificity of self-report of HIV test results

The objective of the study was to assess the reliability and validity of self-report of HIV testing questions on the Risk Behavior Assessment and Risk Behavior Follow-up Assessment. Study 1 had 219 injection drug users and crack cocaine smokers. Study 2 had 259 injection drug users and crack cocaine smokers. Study 3 had data from 17,408 injection drug users and crack cocaine smokers. When the question 'Have you ever been told that you were infected with the AIDS virus' was compared to ELISA result, the specificity was over 99% for both baseline and follow-up. The sensitivity was 32% on the Risk Behavior Assessment, but 61% on the Risk Behavior Follow-up Assessment. Those who were HIV positive at Risk Behavior Assessment (baseline) were less likely to have received their previous HIV test result. Two-thirds of the HIV test-related questions on the Risk Behavior Assessment and Risk Behavior Follow-up Assessment had acceptable reliability. The low sensitivity at Risk Behavior Assessment was probably due to the participants' failure to receive their previous test results.