Abstract

Tuberculosis (TB) is a killer disease, and its root can be traced to Mycobacterium tuberculosis. As the world population increases, the burden of tuberculosis is growing along. Low-and-middle-income nations are not exempted from the tuberculosis crisis. Due to a shortage of medical supplies, tuberculosis bacteria have become a huge public health concern. This study reviewed recent literature from 2015 to 2020 to critically examine what earlier researchers have done about TB burden and treatment. The data used were based on the hospital’s medical department’s record and used a machine-learning algorithm to predict and determine the risk factors associated with the disease. Furthermore, it developed five predictive models to offer the medical managers a valid alternative to the manual estimation of TB patients’ status as cured or not cured. The overall classification showed that all the classification methods performed well for classifying the TB treatment outcome (ranging between 67.5% and 73.4%). Our findings showed that MLP (testing) is the best model to predict TB patients’ treatment outcomes. Age and length of stay were identified as significant risk factors for TB patients in this study. This study explains the study’s limitation, contributions, managerial implications, and suggest future work.