Directly observed therapy and treatment completion for tuberculosis in the United States: Is universal supervised therapy necessary?

Objectives. This study examined the relationship between directly observed therapy and treatment completion rates in the years before and after infusion of federal funding for tuberculosis (TB) control in 1993. Methods. An ecological study of estimated directly observed therapy rates and 12- month treatment completion rates from 1990 through 1994 was undertaken for TB control programs in all 25 cities and counties across the nation with 100 or more incident TB cases in any year from 1990 to 1993. Three cohorts were formed: high treatment completion, intermediate completion, and low completion. Results. In 1990, the median 12-month treatment completion rate was 80% for the entire study population, with a median estimated directly observed therapy rate of 16.8%. By 1994, those rates had increased to 87% and 49.4%, respectively, and increases were shown in all 3 cohorts. Conclusions. Directly observed therapy has had a marked impact on treatment completion rates in jurisdictions with historically low rates. But TB treatment completion rates of more than 90% can be attained with directly observed therapy rates far lower than those proposed by advocates of universal supervised therapy.