Several prophylactic medications for opportunistic or recurrent infections are used in human immunodeficiency virus-infected individuals. Essential to the efficacy evaluation of these agents is the accurate reporting of medication compliance. We hypothesized that poor patient compliance with thrice-weekly fluconazole prophylaxis would correlate with the occurrence of clinical events. Fluconazole compliance was monitored electronically by using the Medication Event Monitoring Systems with 19 women receiving fluconazole at 50 mg thrice weekly for prophylaxis of recurrent mucocutaneous candidiasis. During 202 patient-months of follow-up, eight breakthrough episodes of mucocutaneous candidiasis developed in four women; compliance data were available for seven of these episodes. At 6 months of therapy, more women with greater than or equal to 80% compliance were disease free compared with women with less than 80% compliance (P < 0.05; the Fisher exact test). These data suggest that documentation of medication compliance is essential in studies of chronic prophylaxis in human immunodeficiency virus-infected patients to properly evaluate drug efficacy and to avoid erroneous conclusions concerning drug failure.