Related Experiment Video
Updated: Aug 26, 2025
Author Spotlight: Exploring the Impact of Reduced Resistance Exercise Volume on Metabolic Health
Published on: December 1, 2023
Simple Method to Predict Insulin Resistance in Children Aged 6-12 Years by Using Machine Learning
1Department of Pediatrics, Beijing Jishuitan Hospital, Beijing, People's Republic of China.
Insights
A new machine learning model effectively predicts childhood insulin resistance (IR), a key factor for future diabetes and cardiovascular disease. This simple tool aids pediatricians in early identification and intervention for at-risk children.
Area of Science:
- Pediatric Endocrinology
- Biostatistics
- Machine Learning in Healthcare
Background:
- Childhood insulin resistance (IR) is a growing concern, potentially leading to increased rates of diabetes and cardiovascular disease.
- Early identification of IR in children is crucial for timely intervention and preventing long-term health complications.
- Primary care practitioners require accessible tools to screen for pediatric IR.
Purpose of the Study:
- To develop and validate a machine learning model for predicting insulin resistance in children aged 6-12 years.
- To identify key clinical features predictive of IR in the pediatric population.
- To establish a simple, cost-effective method for IR screening in primary care settings.
Main Methods:
- Utilized data from 503 children (aged 6-12), defining IR using HOMA-IR > 3.0.
- Applied data preprocessing techniques including imputation and oversampling, followed by recursive feature elimination.
- Trained and evaluated five machine learning models (LR, SVM, RF, XGBoost, CatBoost) on an external dataset of 133 children.
Main Results:
- The XGBoost model, utilizing 14 selected features including glucose, waist circumference, and age, demonstrated the highest performance.
- XGBoost achieved an Area Under the Curve (AUC) of 0.85, with strong accuracy, sensitivity, precision, and F1 scores.
- XGBoost exhibited superior specificity (0.79) compared to other models, making it the optimal choice for IR prediction.
Conclusions:
- The developed XGBoost model accurately predicts IR in children aged 6-12.
- This model offers a clinically applicable, simple, and inexpensive method for pediatric IR screening.
- Early identification via this model can facilitate timely interventions, mitigating risks of future diabetes and cardiovascular disease.
Background:
Due to the increasing insulin resistance (IR) in childhood, rates of diabetes and cardiovascular disease may rise in the future and seriously threaten the healthy development of children. Finding an easy way to predict IR in children can help pediatricians to identify these children in time and intervene appropriately, which is particularly important for practitioners in primary health care.
Patients And Methods:
Seventeen features from 503 children 6-12 years old were collected. We defined IR by HOMA-IR greater than 3.0, thus classifying children with IR and those without IR. Data were preprocessed by multivariate imputation and oversampling to resolve missing values and data imbalances; then, recursive feature elimination was applied to further select features of interest, and 5 machine learning methods-namely, logistic regression (LR), support vector machine (SVM), random forest (RF), extreme gradient boosting (XGBoost), and gradient boosting with categorical features support (CatBoost)-were used for model training. We tested the trained models on an external test set containing information from 133 children, from which performance metrics were extracted and the optimal model was selected.
Results:
After feature selection, the numbers of chosen features for the LR, SVM, RF, XGBoost, and CatBoost models were 6, 9, 10, 14, and 6, respectively. Among them, glucose, waist circumference, and age were chosen as predictors by most of the models. Finally, all 5 models achieved good performance on the external test set. Both XGBoost and CatBoost had the same AUC (0.85), which was highest among those of all models. Their accuracy, sensitivity, precision, and F1 scores were also close, but the specificity of XGBoost reached 0.79, which was significantly higher than that of CatBoost, so XGBoost was chosen as the optimal model.
Conclusion:
The model developed herein has a good predictive ability for IR in children 6-12 years old and can be clinically applied to help pediatricians identify children with IR in a simple and inexpensive way.
