Setting: The potential cost-effectiveness of improved diagnostic tests for tuberculosis (TB) in resource-limited settings is unknown.
Objective: To estimate the incremental cost-effectiveness of a hypothetical new point-of-care TB diagnostic test in South Africa, Brazil and Kenya.
Design: Decision-analysis model, adding four diagnostic interventions (sputum smear microscopy, new test, smear plus new test and smear plus TB culture) to a baseline of existing infrastructure without smear.
Results: Adding sputum smear was estimated to be more cost-effective (incremental cost per disability-adjusted life year [DALY] of $86 [South Africa], $131 [Brazil], $38 (Kenya]) than a new TB diagnostic with 70% sensitivity, 95% specificity and price of $20 per test ($198 [South Africa], $275 [Brazil], $84 [Kenya]). However, compared to sputum smear, smear plus new test averted 46-49% more DALYs per 1000 TB suspects (321 vs. 215 [South Africa], 243 vs. 166 [Brazil], 790 vs. 531 [Kenya]), at an incremental cost of $170 (Kenya) to $625 (Brazil) per DALY averted. Cost-effectiveness was most sensitive to the specificity and price of the new test, the baseline TB case detection rate and the discount rate.
Conclusion: Novel diagnostic tests for TB are potentially highly cost-effective. Cost-effectiveness is maximized by high-specificity, low-cost tests deployed to regions with poor infrastructure.