A community-based TB treatment programme of fully supervised, intermittent (twice weekly) ambulatory (SIAT) treatment, in Zululand, S. Africa, is described. The area served was about 3,000 sq. kms. and 200,000 people who lived in scattered kraals. SIAT points were designated, starting with clinics and community health workers, and involving stores, tea rooms, schools and other non-health care sites as need arose. All patients, including children, were offered SIAT and the only indication for hospital admission was severe illness. Each patient was allocated a supervisor of his/ her choice and the emphasis was on the convenience of the patient, not the health service. All patients were transported to their supervisor who was given a 6-month supply of treatment for the patient. Verbal and written instructions were given to all supervisors, who were asked to watch the patient take the medication and then sign the TB card which they retained. The TB health worker visited each supervisor monthly, checked compliance, only visited patients if there was a problem with compliance, and attempted to trace defaulters. Most of the patients who absconded and were not traced had left the area in search of work. Over the study period, only one store refused to supervise a patient, and over 60 different stores were used. Non-health worker supervisors were unpaid.

The findings showed that 89% of surviving patients completed treatment under programme conditions. It was concluded that high completion of treatment rates were possible if services were well-structured, use an intermittent regime, utilise all possible community resources to ensure full supervision of treatment, and are regularly audited. Above all, the service must actively involve and be fully acceptable to the patient.