Modeling the epidemiology and economics of directly observed therapy in Baltimore.

SETTING: From 1958 to 1978, Baltimore maintained one of the highest pulmonary tuberculosis (TB) rates in the US. But, from 1978 to 1992 its TB rate declined by 64.3% and its ranking for TB fell from second highest among large US cites to twenty-eighth. This TB trend coincided with the implementation of an aggressive directly observed therapy (DOT) program by Baltimore's Health Department. OBJECTIVES: We used modeling to estimate the range of TB cases prevented in Baltimore under DOT. Case estimates equal the difference between the observed number of TB cases in Baltimore versus the expected number if Baltimore's TB trend was replaced by the TB trend for the US (low estimate) or the TB trend for all US cities with over 250,000 residents (high estimate). Economic savings are estimated. RESULTS: Without DOT we estimate there would have been between 1,577 (53.6%) and 2,233 (75.9%) more TB cases in Baltimore, costing $18.8 million to $27.1 million. Cases prevented and expenditures saved increased with increased DOT participation. CONCLUSION: Our model predicts that Baltimore's TB decline accompanying DOT resulted in health care savings equal to twice the city's total TB control budget for this period. These results are most plausibly due to DOT, since it was the only major change in Baltimore's TB control program, and rising TB risk factors-AIDS, injection drug use, poverty-in a city where TB had been epidemic should have triggered a TB increase as in comparable US cities, rather than the observed decline. As national TB rates continue to decline it will be important to identify ways to capture and reinvest these savings to support effective TB control programs.