Eleven years of community-based directly observed therapy for tuberculosis.

OBJECTIVE

To evaluate community-based directly observed therapy (DOT) for tuberculosis (TB) control.

DESIGN

Ecological study.

METHODS

Three comparisons were made in this descriptive study. (1) An 11-year retrospective comparison of TB case rates, sputum conversion rates (SCRs), rates of therapy completion, and confounding factors (acquired immunodeficiency syndrome [AIDS], immigration, unemployment, and poverty) in Baltimore, Md, with those of the five major US cities having the highest TB incidence in 1981 but which did not have comprehensive DOT programs. (2) An 11-year trend of TB in Baltimore and the 19 major US cities with the highest TB incidence in 1981. (3) A 7-year trend in TB in both city groups between 1985 and 1992.

SETTING

Twenty US metropolitan cities with more than 250,000 residents.

RESULTS

Since 1981, Baltimore experienced the greatest decline in TB incidence (35.6 cases per 100,000 population, 1981; 17.2 cases per 100,000 population, 1992 [-51.7%]), and city rank for TB (sixth in 1981, 28th in 1992). Conversely, the average incidence of TB increased 2.1% in the five-city cohort and increased 1.8% in the 19-city cohort. Since 1985, TB incidence increased 35.3% in the five-city cohort and 28.5% in the 19-city cohort, but declined 29.5% in Baltimore. From 1986 through 1992, Baltimore's DOT-managed cases had the highest annual SCRs at 3 months (mean, 90.7%), and the highest completion rates for standard anti-TB therapy (mean, 90.1%) when compared with the five cities. These trends could not be attributed to differentials in AIDS, immigration, poverty, or unemployment. Increasingly, more Baltimore cases were treated under DOT (86.5%, 1993) over time. Disease relapse rates remained low, even among HIV-infected patients. Within Baltimore, the documented SCR was significantly higher among DOT-managed cases compared with non-DOT-managed cases (P < .05); multidrug resistance remains rare (0.57%). Within Maryland, Baltimore accounted for 44.4% of all TB cases in 1981, compared with 28.7% in 1992 (P < .001).

CONCLUSIONS

In contrast to the national TB upswing during the 1980s, Baltimore experienced a substantial decline in TB following implementation of community-based DOT, despite highly prevalent medicosocial risk factors. Directly observed therapy facilitated high treatment completion rates and bacteriologic evidence of cure. Directly observed therapy could help reduce TB incidence in the United States, particularly in cities with high case rates.