This study compares drug patterns and prevalence of risk behaviors in a randomized trial using two methods of administration, Audio Computer-Assisted Self-Interview (ACASI) and Interviewer-Administered Questionnaire (IAQ), among drug users seeking treatment in a drug treatment center. We randomized 735 participants: 367 to ACASI and 368 to IAQ. No significant difference in sociodemographic variables were found between subjects in the two arms of the study. Those interviewed by ACASI were more likely to report use on 7 of 10 substances assessed. Rates of reporting of sexual risk behaviors (male-to-male and commercial sex) were higher among participants in the ACASI arm. ACASI seems to be a key resource in improving the reporting of sensitive data in Brazil, as it has been in prior international studies.