Aim: Failure to disclose cocaine use can have a negative impact on medical care and research validity. This study was performed to identify predictors of cocaine non-disclosure among individuals who self-reported heroin use during a medical care encounter.
Design: A prospective comparison of self-report of cocaine use among heroin users and hair analysis for cocaine.
Setting: Four health-care clinics at an academic, inner-city hospital.
Participants: Patients presenting for a health-care visit who were willing to self-report use of heroin and were not engaged in any form of drug treatment.
Measurements: (1) Self-report using standardized instruments: the Drug Addiction Severity Test (DAST), the Addiction Severity Index (ASI) and quantity/frequency questions for heroin and cocaine use. (2) Biochemical evidence: analysis of hair by radioimmunoassay (RIA) for cocaine and opiate levels.
Findings: Among 336 heroin users who tested positive for cocaine in hair, 34.2% did not report their recent cocaine use. The mean cocaine level for discordant individuals was significantly lower than for concordant individuals (109.6 ng/10 mg versus 470.57 ng/10 mg; P < 0.0001). Multivariate predictors of disclosure included opiate and cocaine levels in hair and the ASI drug severity subscore.
Conclusions: Although self-report has been validated for treatment system patients, almost a third of the out-of-treatment heroin users in this medical clinic study failed to disclose concomitant cocaine use. The likelihood of non-disclosure was greatest for heavy users of heroin and light users of cocaine. Confirmation of self-report with biochemical analysis in the medical setting may be necessary to improve both clinical care and research validity.