Background: In Vancouver, the availability of prescription refill data for all HIV-infected individuals plus a prospective cohort of injection drug users (IDUs) permitted an examination of the validity of self-reported highly active antiretroviral therapy (HAART) adherence among IDUs. Methods: Self-reported HAART adherence among Vancouver Injection Drug Users Study participants was compared with pharmacy refill rates from the British Columbia Drug Treatment Program database. Pearson's correlation coefficient and Pearson's χ2 test were used to assess associations between adherence as measured by self-report and pharmacy refill data. Results: Among 88 HIV-infected IDUs, 48 (55%) had an adherence rate of ≥75% as measured by pharmacy refill adherence, whereas 81 (92%) had an adherence rate of ≥75% as measured by self-report. Self-reported adherence was not statistically associated with pharmacy refill adherence (P > .1). Conclusion: These findings suggest that the validity of self-report measures may be limited when applied to community-recruited IDUs.