Abstract
BACKGROUND: Community-based active case finding (ACF) may reduce tuberculosis incidence and mortality but is not widely implemented due to resource constraints. How novel screening test characteristics would affect the impact and costs of ACF strategies is not clear. METHODS: We developed mathematical models of tuberculosis natural history and transmission, calibrated to data from 3 high tuberculosis burden countries (India, Vietnam, and the Philippines) and simulated the impact of annual community-based ACF programs. We compared annual screening, with or without an initial symptom screen, for 10 years with 2 hypothetical screening assays based on target product profile minimum targets: a low complexity (reaches 60% of the population) test with high sensitivity (90%) and moderate specificity (80%) costing $3; and a near point-of-care (reaches 80% of the population) test with moderate sensitivity (55%) and high specificity (98%) costing $2. We projected reductions in tuberculosis incidence, mortality, and costs under each scenario. RESULTS: Annual screening with a low-complexity, high-sensitivity screening assay was projected to reduce tuberculosis mortality by 24%-38%, at an annual cost of $5.04-$5.83 per capita. Annual screening with a near-point-of-care, moderate-sensitivity assay was projected to achieve comparable reductions in mortality (24%-42%) at $3.49-$3.63 per capita. Symptom-based screening would yield more modest reductions in tuberculosis mortality (17%-24%) at an annual cost of around $1 per capita. CONCLUSIONS: Near point-of-care, moderate-sensitivity screening assays for tuberculosis could achieve substantial impact on tuberculosis mortality-equivalent to that of high-sensitivity assays and at lower cost-if they can be performed at scale in community settings.