Trade-offs in Diagnostic Algorithms for Community TB Screening

Diagnostic strategy trade-offs in community TB screening: study design and core findings:

• The study uses a deterministic compartmental TB model to explore how different elements of community screening—diagnostic algorithms, population coverage, and duration—shape reductions in symptomatic TB incidence and TB mortality over a 10-year horizon.
• The model distinguishes symptomatic infectious TB, asymptomatic infectious TB (defined by absence of symptoms at screening), and noninfectious TB.

• Diagnostic algorithms examined:
• Targeting symptomatic infectious TB: prolonged cough with confirmatory Xpert Ultra.
• Targeting infectious TB: Xpert Ultra.
• Targeting all TB: chest X-ray.
• Population coverage and duration were varied, including a scenario with maximum coverage (100%) and five annual screening rounds.
• Primary outcomes were changes in symptomatic TB incidence and TB mortality over 10 years, with attention to potential rebound after screening ceases.
• The model assumes homogeneous population structure and relies on parameter assumptions where data are sparse, including performance of algorithms, treatment acceptance, and completion across disease states.

• Maximum coverage and five annual rounds yield:
• Symptomatic TB incidence reductions of about 26.9% (uncertainty interval 22.8–31.5%) when screening targets symptomatic TB.
• Reductions of about 74.0% (UI 68.5–79.1%) when screening targets infectious TB.
• Post-screening dynamics:
• Incidence rebound erases about 9.8% and 15.9% of the prior reductions within five years for symptomatic-targeted and infectious-targeted approaches, respectively.
• All-TB screening (via chest X-ray) demonstrates greater potential for rapid incidence reductions (exceeding 98% in the model) with minimal rebound, but substantial overdiagnosis emerges due to limited diagnostic accuracy, estimated at 7.2 false positives per true positive in a single round.
• Broader definitions (infectious or all TB) generally achieve larger epidemiologic impact with lower coverage and/or shorter duration, highlighting efficiency in reaching broader case definitions.
• Mortality outcomes tracked showed broadly similar patterns to incidence across the explored scenarios.

• The study is inherently constrained by simplifications (homogeneous population) and by necessary assumptions about diagnostic performance, treatment uptake, and completion across disease states.
• Limitations include potential overdiagnosis with chest X-ray–based approaches and mismatches between model parameters and real-world program implementation.
• The authors emphasize the need for accurate diagnostic tools and appropriate treatment regimens for noninfectious TB to sustain impact.

• The results suggest substantial potential benefits from community screening but underscore trade-offs between symptom-focused versus broader, symptom-agnostic algorithms.