Partial Least Square Segmentation Modeling Using PATHMOX on Tuberculosis Patient Data

One multivariate statistical analysis technique that can explain intricate relationship patterns involving numerous variables is structural equation modeling, or SEM. The Variance-Based SEM (VB-SEM), also known as Partial Least Squares (PLS), is a more adaptable method within SEM that can handle several constraints, including variable indicator scales, limited sample numbers, and assumptions about data distribution. PATHMOX, a method for segmenting structural equation model findings to address observed heterogeneity, is a particular strategy in PLS-SEM. A person's behavior, diagnostic status, medical and treatment history, and socioeconomic variables can all be used to pinpoint specific issues in tuberculosis (TB) research. This work is to segment the structural model results of social determinants of TB using the PLS-PATHMOX approach and to create a structural equation model (PLS-SEM) that represents the clinical state of TB patients. The study uses secondary data from 1019 TB patients who were treated at the Paser District Health Office in 2022. The findings indicate that the incidence of tuberculosis is significantly influenced by social variables, behavior, medical and treatment history, and diagnostic status. The model has good explanatory power, as evidenced by the TB incidence coefficient of determination of 0.85. The PATHMOX-PLS segmentation results demonstrate that TB patients can be categorized according to segmentation criteria including gender and age category. All three split nodes were significant, with pvalues of 0.002 for the first, 0.0474 for the second, and 0.0033 for the third.