TABLE 2.

Variables used in cost-effectiveness and sensitivity analyses^a

Variable	Base value	Lower bound	Upper bound
Sensitivity (%) of MTD in smear-positive specimensb	99.6	91.8 (17)	100 (2, 6, 8, 11, 16, 20)
Specificity (%) of MTD in smear-positive specimensb	99.7	95	100 (2, 11, 19)
Proportion (%) of smear-positive specimens containing detectable MTD inhibitorsb	0, 2.3c	0 (2, 11, 19)	7.2d (17)
Proportion (%) of first smear-positive respiratory specimens that are culture-positive for M. tuberculosis	31.4e	25	98.4 (7)
Annual no. of MTD-candidate patients submitting smear-positive respiratory samples	14e	1	500
Hourly cost of labor ($)	25.47e	15	35
Marginal cost of reagents per test ($)	24e	0	100f
Marginal daily cost of respiratory isolationg ($)	27.77e
Daily cost of four-drug presumptive TB therapyg ($)	5.66h
Median length of time from MTD result to M. tuberculosis speciation resultg (days)	6e

^aReferences from the literature are listed in parentheses. Lower and upper bounds refer to the range over which each given variable was tested in a one-way sensitivity analysis.

^bSee Table 1 for a description of how the variable was estimated.

^c0% in individuals without TB; 2.3% in individuals with TB.

^dSensitivity of the MTD without a test for inhibitors. (A test for inhibitors was assumed in the present analysis.)

^eEstimate obtained directly from our hospital.

^fThe marginal reagent cost per test of a program in which reagents are purchased at list price, and only 15 of 50 tests are performed before the reagent expiration date.

^gVariable not used in the decision analysis to determine the cost of MTD testing per early TB exclusion. Rather, the variable factors into the value that individual hospitals will assign to an early TB exclusion. See Table 3 and Fig. 1.

^hAverage wholesale price in 2001.