Machine Learning for Screening Microvascular Complications in Type 2 Diabetic Patients Using Demographic, Clinical,...

Area of Science:

• * Diabetes Mellitus Research
• * Medical Informatics
• * Computational Biology

Background:

• * Microvascular complications are a major cause of mortality in type 2 diabetes.
• * Early detection and prediction of these complications are crucial for patient outcomes.

Purpose of the Study:

• * To investigate a novel machine learning approach for predicting microvascular complications in type 2 diabetes.
• * To utilize patient demographic, clinical, and laboratory profiles for prediction.

Main Methods:

• * Recruitment of 96 Bangladeshi type 2 diabetes patients.
• * Statistical analysis using chi-squared (χ2) test to identify key predictive markers.
• * Development of machine learning models (logistic regression, random forest, SVM) for automated prediction.

Main Results:

• * Random Forest (RF) achieved high prediction accuracies: 98.67% for cardiac autonomic neuropathy (CAN), 67.78% for diabetic peripheral neuropathy (DPN), and 84.38% for diabetic retinopathy (RET).
• * Key predictors identified: diastolic blood pressure, albumin-creatinine ratio, and gender for CAN; microalbuminuria, smoking history, and hemoglobin A1C for DPN and RET.

Conclusions:

• * Machine learning demonstrates significant promise as an automated tool for predicting microvascular complications in diabetic patients.
• * Utilizing patient profiles can aid in preventing severe complications and reducing mortality in type 2 diabetes.