Related Experiment Video
Updated: Aug 5, 2026
06:48
Clinical Anthropometrics and Body Composition from 3-Dimensional Optical Imaging
Published on: June 7, 2024
Estimation of Cardiometabolic Risk in Turkish Adolescents Using Different Anthropometric Techniques: Development and
Meryem Kahriman1, Nihan Çakır Biçer1,2, Murat Baş1,3
1Department of Nutrition and Dietetics, Graduate School of Health Sciences, Acibadem Mehmet Ali Aydinlar University, 34752 Istanbul, Türkiye.
Nutrients
|July 28, 2026
Summary
This study found that machine learning models, particularly logistic regression, can effectively predict cardiometabolic risk in adolescents. The developed formula aids in screening, but positive results require further confirmation.
Area of Science:
- Adolescent Health
- Cardiovascular Disease Prevention
- Machine Learning in Medicine
Background:
- Obesity is a key factor in cardiometabolic risk, but optimal anthropometric assessment methods are debated.
- This study addresses the need for accurate anthropometric techniques to predict cardiometabolic risk in adolescents.
Purpose of the Study:
- To evaluate various anthropometric techniques for predicting cardiometabolic risk in Turkish adolescents.
- To develop and validate a machine learning-based prediction model for cardiometabolic risk.
Main Methods:
- Utilized data from the Türkiye Nutrition and Health Survey (2010, 2017) including 1357 and 561 adolescents, respectively.
- Assessed anthropometric (BMI z-scores, waist-to-hip/height ratios, TMI, VAI, LAP) and biochemical parameters.
- Developed and validated machine learning models (XGBoost, logistic regression) for risk prediction.
Main Results:
- Visceral adiposity index (VAI) showed the highest discrimination among anthropometric measures (AUC=0.747).
- Logistic regression demonstrated the most stable performance with minimal overfitting and good calibration (Brier score=0.107).
- The derived logistic-based formula showed high negative predictive value (NPV=91.7%), useful for ruling out risk.
Conclusions:
- Machine learning models and a simplified formula can aid in estimating adolescent cardiometabolic risk.
- These tools may support a stepwise screening strategy for cardiometabolic risk assessment.
- Further recalibration and prospective validation are recommended for clinical implementation.