ANALYSIS OF THE CHANCE OF CUSTOMERS PERFORMING ROUTINE REPEAT MCU USING BINARY LOGISTIC REGRESSION METHOD

Abstract

Routine Medical Check-Up (MCU) plays an important role in maintaining the health of the workforce, especially in high-risk industrial sectors such as oil and gas. However, not all individuals who undergo an initial MCU return for a re-examination according to medical recommendations. This study aims to analyze the influence of medical parameters on patient decisions to conduct routine MCUs using the two-stage binary logistic regression method. The data used are the results of MCU records from an oil and gas company, which include nine medical parameters as independent variables and MCU repeat status as the dependent variable. In the first stage, all variables were analyzed for statistical significance. The results showed that total cholesterol (K_Total), fasting blood glucose (GDP), and creatinine were variables that significantly influenced the decision to repeat MCU (p < 0.05). These three variables were then used to build the second stage regression model, which resulted in a prediction accuracy rate of 75.9 percent, with the highest classification accuracy in the group of patients who repeated the MCU (90.6 percent). This finding suggests that medical indicators indicating high risk can be strong predictors of repeat MCU compliance. This study recommends utilizing medical data to build digital follow-up strategies and risk-based occupational health education.