Modeling The Number of Toddler Pneumonia Sufferers in DKI Jakarta using Negative Binomial Regression

Abstract

Acute lung tissue infection caused by various microorganisms, including fungi, viruses, and bacteria, is known as pneumonia. Pneumonia is the highest cause of child death worldwide. In Indonesia, pneumonia remains the leading cause of death among toddler (12-59 months old). By 2021, the national coverage of pneumonias among toddler was 34.8%, and the provinces with the highest coverage for toddler pneumonia were DKI Jakarta (53.0%), Banten (46.0%), and West Papua (45,7%). To find out the pattern of the relationship between the number of young people with pneumonia and the variables that affect it, a custom mathematical model is needed. The number of cases of toddler pneumonia in DKI Jakarta is a data count distributed by Poisson. Poisson regression is perfectly suitable for analyzing data that qualifies equidispersion. However, on the data, the number of toddler pneumonia cases in DKI Jakarta does not meet the equidispersion condition because the variance value is greater than the average or is called overdispersion. One of the methods developed to deal with overdispersion is negative binomial regression. The analysis showed that the average case of toddler pneumonia in Jakarta DKI was 454, Duren Sawit district recorded the highest case of 1329 cases and Sawah Besar district recorded the lowest case as 50 cases. The AIC criteria indicate that the Negative Binomial Regression model is a suitable model for modeling the number of cases of toddler pneumonia in Jakarta DKI with the smallest AIC value of 592,57. The best modeling results using the negative binomial regression method show two significant variables, they are the numbers of toddlers given exclusive breastfeeding and the numbers toddlers that were affected by covid-19.