Development of a multilevel regression model to analyze the association between eye diseases and air pollutants (PM10, PM2.5, NO2, SO2, O3)

초록

Recently, many studies have been conducted on the relationship between air pollutants and respiratory diseases. However, the relationship between eye diseases and air pollutants has not been actively researched compared to the studies with respiratory diseases. Most studies have focused on identifying the presence or absence of associations by analyzing correlations between air pollutants and eye diseases. The objective of developing a multilevel regression model is to provide a more systematic and accurate prediction by taking into account the patient's characteristics (age, sex, etc.) and other environmental characteristics (temperature, humidity, etc.). In regression analysis, unlike correlation analysis, it has the advantage of showing the relation between the incidence of eye disease and other factors as a mathematical prediction model. In multilevel regression, hierarchical structure, the relationship between independent and dependent variables can be better explained than the general regression model. In this study, the relationship between air pollutants and the number of patients tends to change depending on the characteristics of the patients. Therefore, the characteristics of the patient are high-level independent variables that determine the degree of association, and the amount or concentration of pollutants is a low-level independent variable that determines the direct association. Based on data from the National Health Insurance Corporation, this multilevel regression model used the number of patients with various ocular diseases such as keratitis conjunctivitis as a dependent variable based on the data of the eye patients who were treated at the medical institutions in Seoul. And the amount and concentration of air pollutants and patient characteristics and environmental characteristics were used as independent variables.