SETTING: Non-adherence to treatment is a frequent problem in the preventive chemoprophylaxis of tuberculosis. OBJECTIVE: To evaluate the usefulness of the Medication Event Monitoring System (MEMS) for following and improving patient adherence to 6-month treatment with isoniazid. DESIGN: Three methods of monitoring compliance, MEMS, pill count and a urine test for isoniazid, were compared prospectively in 30 patients. The efficacy of a combined intervention by the physician and the pharmacist was evaluated in non-compliant patients. RESULTS: According to the MEMS data, overall adherence to isoniazid therapy was 91.5%, and 86% of the patients were considered compliant throughout the period of observation. The pill count and the urine test tended to overestimate the overall compliance when compared to the MEMS. The combined intervention of the physician and pharmacist allowed drug adherence to be enhanced in non-compliant patients, but the effect was only transient if this was not repeated every month. CONCLUSION: Our results suggest that the MEMS system is a useful approach for monitoring and improving compliance with preventive chemotherapy for tuberculosis.