Predictive Analysis of Diabetes using Logistic Regression (A Case Study from Sahiwal)

Abstract

Diabetes mellitus is a long lasting metabolic disorder that is characterized by increased blood glucose levels, leading to serious complications if not diagnosed and managed at earliest. The prevalence of diabetes is increasing in Pakistan that highlights the need for effective prediction models to identify the individuals which are at high risk. The objective of this study is to predict the likelihood of diabetes among patients and highlight its significant factors. For this purpose, secondary data set of 809 patients collected by the DHQ Teaching Hospital, Sahiwal in 2024 were used. Descriptive statistics, graphical representation and binary logistic regression were applied to determine significant predictors. The analysis revealed that age, gender, hemoglobin A1c, dipsia and nephropathy were the statistically significant factors associated with diabetes. The findings demonstrate that logistic regression is an effective tool for predicting diabetes outcomes and identifying key risk factors. This study will help to highlight the importance of data-driven approaches in improving early diagnosis and supporting medical decision making in healthcare.