Accelerating tuberculosis diagnosis in Mozambican prisons using digital chest X-rays with computer-aided detection: a longitudinal, comprehensive health intervention

Mozambique has a high burden of tuberculosis (TB) and in 2021, an estimated 18,000 persons with TB nationwide were not diagnosed. Estimates suggest that globally 47% of TB disease in prisons is undiagnosed. We implemented an integrated-care model health intervention to enhance the diagnosis and treatment of TB disease, provide TB preventive treatment (TPT), and identify and treat other undiagnosed health conditions in three prisons in Maputo, Mozambique. From July 11, 2023, through the second quarter of 2024, we systematically screened for TB using digital chest X-rays with computer-aided detection (DCXR-CAD). This was combined with symptom screening in a parallel algorithm. We conducted clinical assessments for other health conditions and delivered TPT. Graphs visually compare TB case notification trends with two control prisons, which continued the usual standard of care, symptom screening for TB. An interrupted time series (ITS) analysis was used to evaluate TB case notification counts before and after the intervention in the intervention prisons. Of 7912 individuals screened, 264 new diagnoses of TB disease were notified, a TB screening yield of 3.34% and a number needed to screen of 30, and 1346 persons were initiated on TPT. Other conditions diagnosed and treated include HIV (28), malnutrition (830), and skin conditions (462). TB case notifications were higher in the intervention prisons during the intervention than in the preceding period in the same prisons. Comparing case notifications during the intervention with an equivalent period prior to the intervention showed an increase of about 52% in intervention prisons but a decrease in control prisons by the same magnitude. The ITS analysis found the effect of the intervention on TB case notification counts statistically significant in two of the intervention prisons: (incidence rate ratio IRR 3.96, 95% confidence interval CI 1.51–10.40; p = 0.005) and (IRR 6.17, 95% CI 2.17–17.55; p < 0.001). Strengthening local infrastructure and implementing DCXR-CAD for screening likely contributed to a higher TB yield in the intervention prisons. Screening and diagnosing TB, delivering TB disease treatment and prevention, and addressing other health conditions simultaneously in this vulnerable population were feasible and important.