Aims: This study examined the accuracy of self-reported HIV serostatus for injection drug users and the predictive validity of a measure of social desirability to detect inaccurate self-reports.
Design: Participants in the ALIVE study were provided testing for HIV antibodies and counseling every 6 months. In the separate SAFE study, the same volunteers were asked their HIV serostatus.
Participants: The 348 active injection drug users recruited from the ALIVE study, a longitudinal study of the natural history of HIV among injection drug users, volunteered for an HIV prevention study (SAFE study).
Measurements: The ALIVE baseline interview included a scale to measure "self-deception", a dimension of socially desirable responding.
Findings: Of one hundred and four HIV seropositive participants, 71 (68%) accurately reported their HIV status, 28 (27%) inaccurately reported their HIV serostatus, and five (5%) reported that they did not know or were not sure about their HIV serostatus. Of 242 HIV seronegative participants, 239 (98%) correctly reported their serostatus. Thus the sensitivity of self-report was 72%, specificity 99%, positive predictive value 97%, and the negative predictive value was 90%. For individuals who scored at or below the median on self-deception, the sensitivity of self-reported HIV serostatus was 81%.
Conclusions: The results suggest that high scores on some measures of social desirability may indicate questionable validity of self-reported HIV serostatus.