Importance of confirmatory test characteristics in optimizing community-based screening for tuberculosis: An epidemiological modeling analysis.

Background: Current active case-finding (ACF) efforts for tuberculosis (TB) are limited by the costs, operational barriers, and sensitivity of available tools to confirm a TB diagnosis. However, it is not well understood which of these limitations has the greatest epidemiological relevance and might therefore warrant prioritization in test development.

Methods: We developed a state-transition model of a one-time, community-based ACF intervention, with a fixed budget of one million United States dollars for screening and confirmatory testing. Assuming an adult population with four time the national prevalence of Uganda, we compared the impact of this intervention on TB diagnoses, mortality, and transmission when using a currently available confirmatory test (mirroring sputum-based Xpert Ultra) versus an improved confirmatory test. We considered the following test improvements: (1) increased sensitivity (from 69% to 80%), (2) non-sputum specimen type (increasing specimen availability from 93% to 100%), (3) immediate turn-around of test results (increasing delivery of positive results from 91% to 100%), (4) reduced costs (from $20 to $9 per confirmatory test). For those individuals not included in ACF efforts, TB outcomes under routine care were informed by recent natural history models.

Results: In a simulated target population of 400,000 adults, 6,421 (1.6%; 95% uncertainty range [UR] 5,316-7,531) had TB disease, and 873 (612-1,182) were projected to die of TB in the absence of ACF. Assuming current tests, ACF efforts could reach 83,808 (59,388-118,601; 21% of the target population) people under the allotted budget, connecting 651 (429-983) individuals with TB to treatment and averting 76 (39-132) deaths. Of all hypothetical confirmatory test improvements modeled, higher diagnostic sensitivity most increased the number of people with TB who received treatment as a result of ACF (by 14% [4-26%]). However, considering mortality or transmission as a metric, the largest reductions resulted from tests that provided immediate turn-around of results (by 11% [5-18%]).

Conclusion: Making confirmatory tests for community-based TB screening more accessible and rapid may lead to greater population health benefits than further increasing sensitivity. Nonetheless, achieving large (>20%) increases in the health impact of ACF will require improvements to components of ACF other than the confirmatory diagnostic test.