A decision analysis was conducted to evaluate the cost-effectiveness of programs in which the Amplified Mycobacterium Tuberculosis Direct test (MTD) (Gen-Probe) is used to rapidly exclude Mycobacterium tuberculosis complex as a cause of disease in smear-positive respiratory specimens. MTD sensitivity, specificity, and probability of inhibition for smear-positive specimens were estimated from literature reports. Costs and laboratory performance characteristics were determined from review of records and practices at an urban hospital in the mid-Atlantic United States. In the base case, 31.4% of smear-positive specimens were assumed to be culture positive for M. tuberculosis. Under these conditions, the marginal cost of the MTD testing program was estimated as $338 per smear-positive patient, or $494 per early exclusion of tuberculosis based on negative MTD results. By comparison, the cost of respiratory isolation ($27.77/day) and drugs ($5.66/day) averted by MTD testing was estimated at $201 per early tuberculosis exclusion. MTD testing was therefore not cost-effective in this scenario. Sensitivity analysis revealed that cost-effectiveness estimates are sensitive to the number of smear-positive specimens processed annually, the relative prevalence of M. tuberculosis in smear-positive specimens, and the marginal daily cost of respiratory isolation. A decision tool is therefore presented for assessing the cost-effectiveness of MTD under various combinations of those three variables. While routine MTD testing of smear-positive specimens is not expected to be cost-saving for most individual hospitals, centralized reference laboratories may be able to implement MTD in a cost-effective manner across a wide range of situations.