Survival Analysis Using Cox Proportional Hazard Regression Approach in Dengue Hemorrhagic Fever (DHF) Case in Abdul Moeloek Hospital Bandar Lampung in 2019

Abstract

Survival analysis is one of the statistical procedures analyzing data in the form of survival time and variables that affect survival time, namely survival time data starting from the beginning of the study (time origin/start point) until the time an event or endpoint occurs. In the field of health data is obtained from observations of patients who were observed and recorded the time event of each individual. The event in question can be in the form of death, recurrence of new diseases, or recovery. This study will discuss the application of cox proportional hazard regression to determine the cox proportional hazard model on DHF patients. Determine the factors that affect the recovery rate of DHF patients and determine the hazard ratio value of DHF patients at Abdul Moeloek Hospital in Bandar Lampung in 2019. Regression Cox proportional hazard is used because the cox proportional hazard model does not depend on the assumption of the distribution of the time of occurrence; the results are almost the same as the parametric model. They can estimate the hazard ratio without knowing the baseline hazard. Based on the selection of the best model with backward elimination and the Akaike Information Criterion (AIC), the best model is obtained with a four-variable model, namely Leukocyte, Hemoglobin, Hematocrit, and Thromobocyte. These four variables are factors that have a significant effect on the patient's length of stay. Then this study also looked at the value of the hazard ratio which thrombocyte are the variables with the largest hazard ratio value, which means the thrombocyte variable has the highest risk level.