The purpose of this report was to present findings from a pilot study conducted to explore the associations between sociodemographic, drug use, and health belief factors and perceived compliance with zidovudine (AZT) among African-American drug users. Data were collected in Washington, DC, USA from individuals who were African-American; were recent or current drug injectors or crack smokers; were HIV-seropositive, and were receiving treatment for HIV infection. Participants were recruited through local organizations that provide services to HIV-infected persons. Participants were interviewed using a questionnaire that solicited sociodemographic, lifetime and current drug use, current sexual behaviours, health status, HIV and drug treatment history, and health belief data. Analyses were limited to individuals currently using an illicit substance and who had received AZT during their medical treatment. Parametric (Pearson's r) and nonparametric (Spearman's rho) statistics were used to assess correlations between perceived compliance with AZT dosing and independent variables. As the study was intended to be both descriptive and exploratory, the level of statistical significance was set at 0.10, rather than the customary 0.05. Antiretroviral medications recognized and recalled by participants are presented. The most commonly recalled medication was AZT. Slightly less than one-third of participants reported being completely compliant with an AZT regimen. Perceived compliance was found to be negatively associated with 5 variables: age, homelessness, number of injections in the previous 30 days, trading sex for drugs, and the perception that AIDS is no longer a serious disease since the development of new antiretroviral medications. Intensity of feelings of joy, fear, and the belief that taking more anti-HIV medications would result in better health were found to be positively correlated. Bivariate associations between perceived compliance and sociodemographic, drug use, sexual behaviour, and health belief variables suggest further avenues of study and potential points for intervention to increase compliance with antiretroviral medications among racial/ethnic minority drug users receiving treatment for HIV infection.