Increasing smear positive tuberculosis detection using a clinical score - A stepped wedge multicenter trial from Africa.

Abstract Background The Bandim TBscore is a clinical score that predicts treatment outcome in Tuberculosis (TB) patients and proved useful as an indicator of which healthcare-seeking adults to refer for sputum smear microcopy. We aimed to test in a randomized trial if the TBscore could be used to enhance the detection of smear positive (SP) TB. Methods We carried out a stepped wedge cluster-randomized trial at six health centers in Bissau, Guinea-Bissau, and Gondar, Ethiopia. The primary outcome was diagnostic yield for SP TB. Secondary outcomes were successful treatment and effect on overall 12 months mortality. The study was registered at the Pan African Clinical Trials Registry (PACTR201611001838365). Results We included 3571 adults. Overall, there was no effect of the intervention on SP PTB detected (OR 1.39 (95%CI 0.75 – 2.56). Analysis stratified by country, showed that the TBscore increased case detection in Gondar (OR 4.05 (95%CI 1.67 – 9.85)) but no effect was found in Bissau (OR 0.47 (95%CI 0.22 – 1.05)) where take-up was much lower. Overall mortality decreased during the intervention (HR 0.31 (95%CI 0.13-0.72)). Conclusion Using the TBscore for triage before smear microscopy may improve case detection and decrease mortality if there is sufficient laboratory capacity to increase sputum smears.