Modeling The Number Of Tuberculosis Cases In West Java Using The Negative Binomial Approach

Abstract

Objective: This study aims to model the number of Tuberculosis cases in West Java Province in 2021 using the Negative Binomial Regression approach. Methods: This study used quantitative analysis uses secondary data from the Central Bureau of Statistics website and the Health Office of West Java Province. 27 West Java districts/cities were studied. The number of tuberculosis cases was assumed to be affected by population density, poverty, sanitation, and health complaints in the past month. Negative Binomial Regression was used to analyse data. Results: The results showed that Poisson Regression caused overdispersion, which was solved using the Negative Binomial Regression approach. The Negative Binomial Regression model passed a detailed test. The partial test showed that only the variable percentage of low-income persons and the variable percentage of people with health concerns significantly affected the model with regression coefficients of 0.8755 and 1.0318, respectively. The final Negative Binomial Regression model with the lowest Akaike Information Criterion value of 491.9 is best for this investigation. Conclusion: The most suitable model for modelling the number of Tuberculosis cases in West Java Province in 2021 is the Negative Binomial Regression model with independent variables that significantly influence the model, namely the percentage of poor people and the percentage of people who have had complaints recently.