Prediction of Type 2 Diabetes Mellitus using a Logistic Regression Model

Foundation: type 2 diabetes mellitus constitutes a growing epidemic and represents a substantial economic burden for health systems. Detecting the disease at an early stage helps reduce medical costs and the risk of patients having more complicated health problems. Objective: to design a mathematical model to predict the type 2 diabetes mellitus probability of existence in patients treated at a hospital in Guayaquil, Ecuador. Method: a descriptive and cross-sectional study was carried out. The population was made up of 324 patients. The statistical procedure was based on the binary logistic regression application. To evaluate the predictive capacity of the model, Cohen's Kappa test was used. Results: high blood pressure was a positive risk factor for type 2 diabetes mellitus, with a probability coefficient of 1.415. Positive family history influenced the increased probability. Alcohol consumption was a positive risk factor and the coefficient of 0.790 indicated how much it contributed to the increased probability. The Kappa coefficient had a value of 0.434; with approximate T of 7.809 and p < 0.001, it indicated greater prevalence than bias and greater agreement between what was predicted in the model and what was observed. Conclusions: the presence of high blood pressure, positive family history and alcohol consumption were significant factors that increased the probability of developing type 2 diabetes mellitus. Early detection and management of these risk factors is important in the prevention and management of the illness.