This study examined under naturalistic assessment conditions the validity of self-reported opiate and cocaine use among 175 veterans enrolled in methadone treatment, and factors related to self-report validity, such as stage in treatment and drug of abuse. Veterans were interviewed by clinical staff about past 30-day drug use with the addiction severity index (ASI), and urinalysis results were obtained for the same 30-day interval assessed with the ASI. Analysis revealed that urinalysis generally produced higher rates of substance use than patient self-report, and with the exception of reported opiate use among new patients presenting for treatment, validity of patient self-reported drug use generally was poor with patients under-reporting both opiate and cocaine use. The findings are in marked contrast to those obtained in other studies in which participants are ensured confidentiality regarding their self-reports. Further, the results raise questions about the utility of self-report measures of substance use to assess patient progress or methadone program performance. Copyright (C) 2000 Elsevier Science Ireland Ltd.