Data Mining Analysis with Orange in the Development of Tuberculosis Among Diabetes Mellitus Patients

Abstract

Abstract Prevention and treatment of diabetes will have a positive influence on tuberculosis (TB) since people may get TB because they have diabetes mellitus (DM). Recording and reporting through the TB Information System are not run optimally because of many factors. The information system must be strengthened to be used by private health facilities. This study used secondary data from the 2013 and 2018 Indonesian Basic Health Research (IBHR). The data was analyzed univariately and analyzed further using Orange Data Mining Tools to test the screening tool model used to predict TB in diabetic individuals. The total sample in this study from each data was 38,136 people. The 2013 IBHR stated that 749 people (2%) were diagnosed with pulmonary TB, while the 2018 IBHR stated that 97 people (0.3%) were diagnosed in the previous six months. The results of the Orange analysis showed that precision and recall calculations in this study were quite good, at 0.9. Therefore, the model would likely predict the occurrence of TB in diabetic individuals. According to Orange, the TB-DM electronic screening tool model tends to estimate the incidence of TB in diabetic individuals. Keywords: diabetes mellitus, Orange analysis, screening, screening tool, tuberculosis