Novel statistical classification model of type 2 diabetes mellitus patients for tailor-made prevention using data...

Area of Science:

• Medical Informatics
• Data Mining
• Diabetology

Background:

• Diabetic vascular complications pose a significant health burden.
• Predicting these complications aids in early intervention and management.
• Identifying risk factors in order of importance is crucial for tailored prevention strategies.

Purpose of the Study:

• To evaluate the utility of data mining algorithms for identifying and ordering risk predictors of diabetic vascular complications.
• To apply the Classification and Regression Trees (CART) method to predict macroangiopathy and microangiopathy in type 2 diabetes.

Main Methods:

• Utilized Classification and Regression Trees (CART) analysis on prevalence data from 165 type 2 diabetic outpatients.
• Analyzed 6 categorical and 15 continuous risk factors to identify predictors for macroangiopathy and microangiopathy.
• Determined optimal cutoff points for significant risk factors within different age groups.

Main Results:

• Age was the primary predictor for both macroangiopathy (cutoff: 65.4 years) and microangiopathy (cutoff: 64.8 years).
• In older patients, body weight predicted macroangiopathy, while BMI predicted microangiopathy.
• Systolic blood pressure and morbidity term were significant predictors in younger patients for macroangiopathy and microangiopathy, respectively.

Conclusions:

• The CART method successfully identified clinically relevant risk factors and their cutoff points for diabetic vascular complications.
• This data mining approach shows potential for prioritizing predictors to enable personalized prevention of diabetic vascular complications.