Setting: Hlabisa health sub-district, KwaZulu-Natal, South Africa.
Objective: To describe the establishment of a community-based multidrug-resistant tuberculosis (MDR-TB) treatment programme embedded in the district TB control programme and to evaluate whether early outcomes are comparable to those in the traditional hospital-based model of care.
Design: Cases who initiated community-based MDR-TB treatment (CM) between March and December 2008 were compared with patients who initiated MDR-TB treatment under the traditional hospital-based model of care (TM) between January 2001 and February 2008. Time to initiation of treatment and time to sputum smear and culture conversion were compared for the two groups in Kaplan-Meier survival curves using the Mantel-Cox log-rank test.
Results: Overall, 50 CM cases and 57 TM cases were included; 39 of the 50 CM cases (78.0%) were human immunodeficiency virus positive. The median time to initiation of treatment was 84 days for CM and 106.5 days for TM (P = 0.002). Median time to sputum smear conversion was shorter for CM than TM (59 vs. 92 days, P = 0.055), as was time to sputum culture conversion (85 vs. 119 days, P = 0.002).
Conclusion: Community-based treatment for MDR-TB can be implemented within the existing TB control programme in rural South Africa and should be scaled up where resources allow.