Modelling of risk factors that influence malaria infection using Binary Logistic Regression

Abstract

Malaria is a parasitic disease that is currently endemic in many countries, including in Indonesia. Batubara District in North Sumatera is one of the malaria-endemic areas. This study was conducted to make a mathematical model based on risk factors that influence malaria infection. The data used is secondary data from other researchers. The method used to model risk factors that influence malaria infection is the binary logistic regression with a categorical response variable, namely diagnosis by microscopic examination. There are 10 variable predictors; mosquito nets (X 1), quality use of mosquito nets (X 2), condition of the livingenvironment (X 3), knowledge (X 4), attitudes (X 5), actions (X 6), quality of health workers (X 7), use of mosquito coils (X8), use of topical insect repellent (X 9), and access of health workers (X 10). Maximum likelihood estimation is used to findout the parameter estimates. From the 10 predictor variables studied, it was found that only 8 variables had a significant effect on the model. Based on the odds ratio, the most significant variable is the quality of health workers with 16,5. The ability of predictor variables in explaining the response variable in the model is 87, 8%. The classification function in this study is 96, 3%.