TB: Development of Prevention and Control Pulmonary Tuberculosis Process in Prisons

Abstract

A retrospective analytic study without concurrent controls aimed to develop a screening and diagnosis process for tuberculosis (TB) patients in prisons, to determine the appropriate time interval for TB screening of “long-term” inmates, and to study the effectiveness of an AI-based system in screening for TB. The sample group consisted of 22,000 long-term captives from 14 prisons in the Region 4 Health Provider. Collection of data carried out from November 1, 2021 through March 31, 2022. The screening process conducted finding cases in the time every 3, 6, and 12 months using digital X-ray systems, and for those with abnormal lung tissues to be tested for TB using the GeneXpert MTB/RIF or RT-PCR method. Analyses of the data by using the Mann-Whitney U test needed to compare the image scores of inmates with TB to those without TB. The study found that all the long-term inmates were screened for TB, and the prevalence of TB in the 3 and 6-month screenings was 0.62% and 0.63% respectively, which was higher than the 12-month screening. Prisons with screening systems did not find any long-term inmates had TB, and the AI-based system was able to significantly distinguish TB patients from non-TB patients. Furthermore, 56.1% of TB patients experienced delayed treatment. The AI-based system can screen for TB in prisons faster than former methods, and it is recommended that prisons installing an AI-based X-ray system to screen long-term prisoners every 3 or 6 months is able to reduce the spread of TB and to prepare the lab for an increase in the amount of sputum tests.