Log-Linear Applications To The Relationship Of Maternal Age And Parity To Preterm Birth

The log-linear model is a special model of the general linear model of Poisson distributed data, and also the development of cross-tabulation analysis of two or more categorical variables. The purpose of this study was to determine the relationship or interaction between the variables of maternal age, parity, and preterm birth. This study uses medical record data at the Haji General Hospital in East Java Province, namely patient data for pregnant women who gave birth between January 1st 2020 to 31st December 2021. The number of samples was 147 respondents. The data analysis method used is log-linear regression analysis. The log-linear model is used as an alternative solution to show if there is a relationship between several variables in a multidimensional contingency table, with the ability to modify the interaction between two or more variables. The resulting log-linear model is: logμ ijk = 4.083 − 0.693 (X) − 0.638 (Y) − 3.795 (Z) + 2.143 (YZ) . The resulting model states that there is no simultaneous interaction between preterm birth, maternal age, and parity, but there is a partial interaction between maternal age and parity where preterm birth is significant in the model (YZ, X).