Directly observed therapy for tuberculosis in rural South Africa, 1991 through 1994

Objectives. This paper describes an audit of a community-based tuberculosis treatment program involving directly observed therapy in South Africa. Methods. A program audit of 2473 consecutive tuberculosis patients in Hlabisa Health District, KwaZulu/Natal, South Africa, was conducted between 1991 and 1994. Results. Monthly admissions increased from 34 per month in 1991 to 66 in 1994. Of 2186 patients managed in Hlabisa, 1903 (87%) received directly observed therapy. Of those receiving directly observed therapy, 1034 (55%) were supervised by volunteers; 743 (72%) of these were supervised by storekeepers. Among those patients managed locally, 1679 (85%) of 1967 surviving patients completed treatment. Completion rates for patients supervised by health workers and non-health workers were the same. Completion fell from a high of 90% in 1992 to 78% in 1994. Mortality increased from 5% in 1991 to 10% in 1994. Conclusions. Community-based directly observed therapy that uses an intermittent drug regime and volunteers as supervisors can achieve high treatment completion rates for tuberculosis, even in resource-poor settings.