Results indicated a modest correlation between the self-reported frequency of heroin use among the inmates and the morphine concentrations in their hair samples. Hair analysis based on immunoassay was determined to be inadequate as a primary measure of treatment outcome because it underestimated heroin use in the treatment population. Hair analysis was useful, however, at detecting the under-reporting of heroin use during treatment. As such, hair analysis may serve as valuable supplementary evidence alongside of self-reported data on heroin use. The use of sectional hair analysis to determine the duration of drug use, on the other hand, was found to be both uninformative and largely impractical due to the length of hair sections needed to complete the analysis. Participants were 382 male prisoners who were randomly assigned to either a methadone treatment program or a 4-month wait list control group. Participants were interviewed at in-take and at a 4-month follow-up about their drug use, injecting behavior, and syringe sharing. Participants also provided hair samples for heroin testing. Over 50 hairs were taken from each participant and were analyzed using enzyme-linked immunosorbent assay with a morphine specific antibody. Each hair section was matched to the self-reported heroin use for the corresponding month using kappa, McNemar’s test of symmetry, and Pearson’s correlation coefficient r. Tables, figure, references