Data mining approaches for type 2 diabetes mellitus prediction using anthropometric measurements

Area of Science:

• Endocrinology
• Metabolic Diseases
• Biostatistics

Background:

• Type 2 Diabetes Mellitus (T2DM) is a growing global health concern.
• Identifying reliable predictors for T2DM risk is crucial for early intervention.
• Anthropometric measurements offer a non-invasive approach to risk assessment.

Purpose of the Study:

• To identify anthropometric measurements most strongly associated with T2DM risk.
• To apply machine learning techniques for enhanced predictive accuracy.
• To validate findings using statistical modeling and receiver operating characteristic analysis.

Main Methods:

• Prospective study involving 9354 participants aged 35-65.
• Collection of diverse anthropometric data including Waist Circumference (WC), Hip Circumference (HC), and Body Adiposity Index (BAI).
• Analysis using logistic regression (LR) and decision tree (DT) models, with ROC curve evaluation.

Main Results:

• Logistic regression identified Waist Circumference (WC) and demispan as key predictors in women, and WC and Body Adiposity Index (BAI) in men.
• Decision tree analysis highlighted WC as the most critical factor, followed by HC and BAI.
• All identified associations were statistically significant (p-value < 0.001).

Conclusions:

• Waist Circumference (WC) emerged as the paramount anthropometric predictor for T2DM risk in both sexes.
• These findings underscore the importance of WC in T2DM risk assessment strategies.
• Machine learning approaches effectively identified key anthropometric determinants of T2DM.