Construction of a prediction model for sarcopenic obesity based on machine learning

Area of Science:

• Gerontology
• Medical Informatics
• Public Health

Background:

• Sarcopenic obesity (SO) is a growing concern in aging populations, linked to increased disability and mortality.
• Early identification of SO is crucial for managing health risks in older adults.

Purpose of the Study:

• To develop and validate machine learning (ML) models for predicting sarcopenic obesity (SO) in older adults.
• To identify key clinical factors associated with SO.

Main Methods:

• Utilized data from 386 participants, splitting into training (80%) and testing (20%) sets.
• Employed univariate and multivariate logistic regression to identify independent predictors of SO.
• Developed and cross-validated five ML models (RF, NB, LightGBM, KNN, XGBoost) to predict SO.

Main Results:

• Identified BMI, Barthel Index score, grip strength, and calf circumference as independent predictors of SO.
• The Random Forest (RF) model achieved the highest Area Under the Curve (AUC) of 0.839 for predicting SO.
• Calf circumference emerged as a significant factor in assessing SO risk.

Conclusions:

• A predictive model combining BMI, Barthel Index score, grip strength, and calf circumference, based on the RF model, can reliably predict SO.
• This ML-driven approach offers a promising tool for early SO detection in clinical practice.