Optimising screening for recurrent tuberculosis: a microsimulation-based study.

Abstract

Background: People recently treated for tuberculosis (TB) (TB survivors) face substantial risk of recurrence. The optimal implementation and expected impact of active case-finding are uncertain.

Methods: We developed an individual-level stochastic simulation model of TB recurrence among recently treated people in India, informed by empirical data on the cumulative incidence and timing of TB recurrence, the extent and durations of symptoms and bacteriological positivity and screening-related costs. We modelled post-treatment symptom screening at quarterly or semiannual intervals, with or without universal near-point-of-care bacteriological testing at the first screening encounter. We estimated months of symptomatic TB and of potentially infectious (bacteriologically positive, regardless of symptom status) TB averted over a 2-year period after treatment, along with corresponding screening programme costs, estimated as medians and 95% uncertainty ranges across simulations.

Results: Symptomatic TB recurred within 2 years post-treatment in 10% (8%-12%) of recent TB survivors and undiagnosed prevalence peaked in month 2 (1-3). Symptom screening 6, 12 and 18 months post-treatment (as per Indian guidelines) averted 16% (8%-32%) of symptomatic and 8% (5%-17%) of infectious time. An 'early' screening schedule (3, 6 and 9 months post-treatment) increased case detection and health impact (26% (15%-43%) of symptomatic and 16% (9%-27%) of infectious time averted). Adding 'universal bacteriologic testing' at the 3-month encounter doubled the impact on transmission potential (34% (21%-55%) of symptomatic and 37% (22%-62%) of infectious time averted). In this context, 'early' symptom screening cost $190 ($100-$350) per case detected, and adding universal bacteriological testing cost an additional $520 ($240-$1070) per case incrementally detected by the screening programme.

Conclusions: Screening for recurrent TB in high-risk settings such as India could reduce morbidity and time spent with potentially infectious TB. Concentrating screening within the first year has greatest yield. Symptom-based screening is most affordable, but testing for asymptomatic recurrence could further reduce transmission potential.