Objective: To determine rates of compliance with doxycycline therapy for patients attending two inner city sexually transmitted diseases (STD) clinics using the MEMS (Medication Event Monitoring System) technology (Aprex Corporation, Fremont, CA). Design: An observational study. Setting: Two STD clinics in Brooklyn, New York and Birmingham, Alabama. Patients: Patients warranting doxycycline as antichlamydial therapy by usual clinical criteria (e.g., documented chlamydial infectious, gonococcal urethritis, mucopurulent cervicitis) were enrolled consecutively from both clinics into four separate categories according to gender and the presence or absence of symptoms: symptomatic men (77), asymptomatic men (30), symptomatic women (83), asymptomatic women (33). Intervention: In the clinic area, patients were given their doxycycline in standard 30-dram medication bottles fitted with the MEMS cap, which is capable of recording the date and time of each bottle opening and closing. This information was then retrieved using a software program developed by the manufacturer. Patients were instructed to return the bottle and cap at the completion of therapy. Efforts were made to contact those who did not return their bottles by both telephone and mail. Outcome Measures: Bottle openings as recorded by the MEMS were considered to represent use of medication. Patients were considered strictly compliant with prescription instructions if bottle openings and closings occurred at least twice dally for 6 consecutive days. Noncompliance was defined as initially opening the medication more than 48 hours after leaving the clinic or opening the bottle less than once daily for 5 consecutive days. Usage between these extremes was classified as intermediate. Results: Eighty percent of 223 patients enrolled completed the study by returning their bottles. The rate of strict compliance with prescription instruction was 25%. The rate of non- compliance was 24%. Fifty-one percent used some intermediate amount of medication. There was no statistical difference in compliance by gender, presence or absence of symptoms, or site of enrollment. Conclusions: Few patients administered doxycycline in an STD clinic can be expected to take medication precisely as prescribed. Although most probably take enough to eradicate uncomplicated chlamydial infections, a sizable portion can be expected to use an inadequate amount of medication. This may contribute to persistence of genital chlamydia infections in the community.