Objective: To determine the prevalence of recent cocaine use and the reliability of patient self-reported cocaine use.
Design: A survey with blinded comparison to a criterion standard.
Setting: Walk-in clinic of a large public hospital in metropolitan Atlanta, Ga.
Participants: Male patients, aged 18 to 39 years, presenting to the triage desk for immediate care during weekdays. Of the 415 eligible men who agreed to participate (acceptance rate, 82%), the average age was 29.5 years, 91.6% were black, and 89% were uninsured.
Intervention: None.
Main outcome measures: Comparison of self-reported illicit drug use with results from urinary immunoassays for benzoylecgonine, a major cocaine metabolite. Determination of which drug history questions produce the most accurate responses using anonymous urine testing as the criterion standard.
Results: Thirty-nine percent of patients tested positive for the presence of benzoylecgonine and were statistically more likely to be older, black, and have a prior history of sexually transmitted disease (P less than .01). Seventy-two percent of men with positive urinary assays denied illicit drug use in the 3 days prior to sampling. When queried with several formats, subjects with positive urine assays were more likely to admit to "any illegal drug" use (87.5%) than to the more specific "any form of cocaine" use (60.6%) within the prior year (P less than .0001).
Conclusions: These results underscore the magnitude of cocaine abuse among black, inner-city men. Patient self-report of illicit drug use is highly inaccurate. Accuracy of self-report may be increased by asking less specific questions.