The purpose of this trial was to test a simple method of measuring and identifying non-compliance with antiretroviral (ARV) therapy and cotrimoxazole (CTX) prophylaxis in a resource-limited setting. A cross-sectional study was undertaken among HIV patients attending two outpatient clinics in Bangui, Central African Republic. Compliance with ARV and CTX treatment was assessed based on 5 measurement modalities, i.e., skipping medication during the 4 days prior to attendance, attendance assiduity, number of remaining tablets, patient visual analogue scale (VAS), and physician VAS. These measures were combined to obtain an overall medication compliance score. A total of 141 patients were interviewed including 89 using ARV and 52 using CTX. Compliance scores varied according to measurement modality from 66.3% to 96.6% for ARV and from 67.3% to 90.4% for CTX. The only significant difference between patients using ARV and CTX involved physician VAS that was significantly lower for CTX than ARV (p=0.04). Overall medication compliance scores classified 117 patients (83%) as compliant and 21 patients (17%) as non-compliant. According to this study the level of medication compliance was relatively good (83%). Findings also indicated that measurement of compliance was feasible using a combination of remaining tablet count and assessment of medication skipping in the last 4 or 7 days. Routine clinical use of these modalities allows assessment and improvement of medication compliance.