Development and validation of a machine learning model for predicting pediatric metabolic syndrome using anthropometric and bioelectrical impedance parameters
Authors' conclusion
The integrated model using both anthropometric and BIA parameters demonstrated strong predictability for pediatric MS, underlining its potential as an effective screening tool for MS in both clinical and general populations.
Study context
- Population
- Youths from Korea National Health and Nutrition Examination Survey (KNHANES) (n=359; 16 metabolic syndrome (MS), 343 normal) and real-world clinical data (n=174; 66 MS, 108 normal)359 KNHANES
- Comparison
- Anthropometric vs bioelectrical impedance analysis (BIA) vs combined prediction models
- Outcome
- AUC for predicting pediatric metabolic syndrome
Methodological detailsmachine learning (eXtreme Gradient Boosting), development and validation, Shapley value analysis
- Machine learning (eXtreme Gradient Boosting)
- Development and validation
- Shapley value analysis
Performance
As reported in the abstract
AUC for pediatric MS
| Test or model | AUC (KNHANES) | AUC (real-world) |
|---|---|---|
| Model 1 | 0.75 | 0.56 |
| Model 2 | 0.66 | 0.61 |
| Model 3 | 0.90 | 0.74 |
AUC (KNHANES)
Model 10.75
Model 20.66
Model 30.90
AUC (real-world)
Model 10.56
Model 20.61
Model 30.74
- Model 3 vs Model 1: P = 0.026 — Model 3 outperformed
- Model 3 vs Model 2: P = 0.033 — Model 3 outperformed
- Model 3 vs Model 1: P = 0.035 — Model 3 outperformed
- Model 3 vs Model 2: P = 0.008 — Model 3 outperformed
Further findings
Contributors to Model 3 (Shapley analysis)
- Body fat mass was the most significant contributor to Model 3
Original publication
Development and validation of a machine learning model for predicting pediatric metabolic syndrome using anthropometric and bioelectrical impedance parameters. Int J Obes (Lond). Published 1 April 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →