Machine Learning-Based predictive model for adolescent metabolic syndrome: Utilizing data from NHANES 2007–2016
Abstract Metabolic syndrome (Mets) in adolescents is a growing public health issue linked to obesity, hypertension, and insulin resistance, increasing risks of cardiovascular disease and mental health problems. Early detection and intervention are crucial but often hindered by complex diagnostic req...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-88156-4 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832585869073055744 |
|---|---|
| author | Yu-zhen Zhang Hai-ying Wu Run-wei Ma Bo Feng Rui Yang Xiao-gang Chen Min-xiao Li Li-ming Cheng |
| author_facet | Yu-zhen Zhang Hai-ying Wu Run-wei Ma Bo Feng Rui Yang Xiao-gang Chen Min-xiao Li Li-ming Cheng |
| author_sort | Yu-zhen Zhang |
| collection | DOAJ |
| description | Abstract Metabolic syndrome (Mets) in adolescents is a growing public health issue linked to obesity, hypertension, and insulin resistance, increasing risks of cardiovascular disease and mental health problems. Early detection and intervention are crucial but often hindered by complex diagnostic requirements. This study aims to develop a predictive model using NHANES data, excluding biochemical indicators, to provide a simple, cost-effective tool for large-scale, non-medical screening and early prevention of adolescent MetS. After excluding adolescents with missing diagnostic variables, the dataset included 2,459 adolescents via NHANES data from 2007–2016. We used LASSO regression and 20-fold cross-validation to screen for the variables with the greatest predictive value. The dataset was divided into training and validation sets in a 7:3 ratio, and SMOTE was used to expand the training set with a ratio of 1:1. Based on the training set, we built eight machine learning models and a multifactor logistic regression model, evaluating nine predictive models in total. After evaluating all models using the confusion matrix, calibration curves and decision curves, the LGB model had the best predictive performance, with an AUC of 0.969, a Youden index of 0.923, accuracy of 0.978, F1 score of 0.989, and Kappa value of 0.800. We further interpreted the LGB model using SHAP, the SHAP hive plot showed that the predictor variables were, in descending order of importance, BMI age sex-specific percentage, weight, upper arm circumference, thigh length, and race. Finally, we deployed it online for broader accessibility. The predictive models we developed and validated demonstrated high performance, making them suitable for large-scale, non-medical primary screening and early warning of adolescent Metabolic syndrome. The online deployment of the model allows for practical use in community and school settings, promoting early intervention and public health improvement. |
| format | Article |
| id | doaj-art-b56bf828d871436491e97663eaf008a9 |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-b56bf828d871436491e97663eaf008a92025-01-26T12:26:01ZengNature PortfolioScientific Reports2045-23222025-01-0115111310.1038/s41598-025-88156-4Machine Learning-Based predictive model for adolescent metabolic syndrome: Utilizing data from NHANES 2007–2016Yu-zhen Zhang0Hai-ying Wu1Run-wei Ma2Bo Feng3Rui Yang4Xiao-gang Chen5Min-xiao Li6Li-ming Cheng7Department of Anesthesiology and Surgical Intensive Care Unit, Kunming Children’s HospitalDepartment of Emergency, The First Affiliated Hospital of Kunming Medical UniversityDepartment of Cardiac Surgery, Fuwai Yunnan Hospital, Chinese Academy of Medical Sciences/Affiliated Cardiovascular Hospital of Kunming Medical UniversityDepartment of Anesthesiology and Surgical Intensive Care Unit, Kunming Children’s HospitalDepartment of Clinical Laboratory, The Third Affiliated Hospital of Kunming Medical UniversityDepartment of Anesthesiology and Surgical Intensive Care Unit, Kunming Children’s HospitalDepartment of Anesthesiology and Surgical Intensive Care Unit, Kunming Children’s HospitalDepartment of Anesthesiology and Surgical Intensive Care Unit, Kunming Children’s HospitalAbstract Metabolic syndrome (Mets) in adolescents is a growing public health issue linked to obesity, hypertension, and insulin resistance, increasing risks of cardiovascular disease and mental health problems. Early detection and intervention are crucial but often hindered by complex diagnostic requirements. This study aims to develop a predictive model using NHANES data, excluding biochemical indicators, to provide a simple, cost-effective tool for large-scale, non-medical screening and early prevention of adolescent MetS. After excluding adolescents with missing diagnostic variables, the dataset included 2,459 adolescents via NHANES data from 2007–2016. We used LASSO regression and 20-fold cross-validation to screen for the variables with the greatest predictive value. The dataset was divided into training and validation sets in a 7:3 ratio, and SMOTE was used to expand the training set with a ratio of 1:1. Based on the training set, we built eight machine learning models and a multifactor logistic regression model, evaluating nine predictive models in total. After evaluating all models using the confusion matrix, calibration curves and decision curves, the LGB model had the best predictive performance, with an AUC of 0.969, a Youden index of 0.923, accuracy of 0.978, F1 score of 0.989, and Kappa value of 0.800. We further interpreted the LGB model using SHAP, the SHAP hive plot showed that the predictor variables were, in descending order of importance, BMI age sex-specific percentage, weight, upper arm circumference, thigh length, and race. Finally, we deployed it online for broader accessibility. The predictive models we developed and validated demonstrated high performance, making them suitable for large-scale, non-medical primary screening and early warning of adolescent Metabolic syndrome. The online deployment of the model allows for practical use in community and school settings, promoting early intervention and public health improvement.https://doi.org/10.1038/s41598-025-88156-4Machine learningMetabolic syndromePredictive modelAdolescentsNHANES |
| spellingShingle | Yu-zhen Zhang Hai-ying Wu Run-wei Ma Bo Feng Rui Yang Xiao-gang Chen Min-xiao Li Li-ming Cheng Machine Learning-Based predictive model for adolescent metabolic syndrome: Utilizing data from NHANES 2007–2016 Scientific Reports Machine learning Metabolic syndrome Predictive model Adolescents NHANES |
| title | Machine Learning-Based predictive model for adolescent metabolic syndrome: Utilizing data from NHANES 2007–2016 |
| title_full | Machine Learning-Based predictive model for adolescent metabolic syndrome: Utilizing data from NHANES 2007–2016 |
| title_fullStr | Machine Learning-Based predictive model for adolescent metabolic syndrome: Utilizing data from NHANES 2007–2016 |
| title_full_unstemmed | Machine Learning-Based predictive model for adolescent metabolic syndrome: Utilizing data from NHANES 2007–2016 |
| title_short | Machine Learning-Based predictive model for adolescent metabolic syndrome: Utilizing data from NHANES 2007–2016 |
| title_sort | machine learning based predictive model for adolescent metabolic syndrome utilizing data from nhanes 2007 2016 |
| topic | Machine learning Metabolic syndrome Predictive model Adolescents NHANES |
| url | https://doi.org/10.1038/s41598-025-88156-4 |
| work_keys_str_mv | AT yuzhenzhang machinelearningbasedpredictivemodelforadolescentmetabolicsyndromeutilizingdatafromnhanes20072016 AT haiyingwu machinelearningbasedpredictivemodelforadolescentmetabolicsyndromeutilizingdatafromnhanes20072016 AT runweima machinelearningbasedpredictivemodelforadolescentmetabolicsyndromeutilizingdatafromnhanes20072016 AT bofeng machinelearningbasedpredictivemodelforadolescentmetabolicsyndromeutilizingdatafromnhanes20072016 AT ruiyang machinelearningbasedpredictivemodelforadolescentmetabolicsyndromeutilizingdatafromnhanes20072016 AT xiaogangchen machinelearningbasedpredictivemodelforadolescentmetabolicsyndromeutilizingdatafromnhanes20072016 AT minxiaoli machinelearningbasedpredictivemodelforadolescentmetabolicsyndromeutilizingdatafromnhanes20072016 AT limingcheng machinelearningbasedpredictivemodelforadolescentmetabolicsyndromeutilizingdatafromnhanes20072016 |