Analysis of ROC curves for the effect of predicting hand grip strength based on morphological indicators in high school girls
10.16835/j.cnki.1000-9817.2025024
- VernacularTitle:高中女生身体形态指标对握力预测作用的ROC曲线分析
- Author:
WANG Kai, ZHAO Guanxiang
1
Author Information
1. College of Physical Education, Jiangsu Normal University, Xuzhou (221116) , Jiangsu Province, China
- Publication Type:Journal Article
- Keywords:
Body weights and measures；Hand strength；ROC curve；Regression analysis；Students；Female
- From:
Chinese Journal of School Health
2025;46(1):111-114
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To evaluate and compare the accuracy of body mass index (BMI), waist circumference (WC), waistcorrected BMI (wBMI), and waist to height ratio (WHtR) in predicting hand grip strength (HGS) in high school girls, so as to provide the reference for evaluating their overall physical fitness and health status.
Methods:From May to December in 2022, a total of 1 600 girls aged 15 to 18 years old were selected by cluster random sampling from 12 high schools in Xuzhou, Jiangsu Province. Measures of HGS, waist circumference, height and weight among high school girls were taken, and their BMI, wBMI and WHtR were calculated. Pearson correlation analysis and Logistic regression were used to analyze the association between BMI, wBMI, WC, WHtR and HGS Receiver operating characteristic (ROC) curves were constructed to evaluate the accuracy of different indicators in predicting HGS
Results:Pearson correlation analysis showed that the BMI, wBMI, WHtR and HGS expression of high school girls were positively correlated ( r =0.31, 0.27, 0.32, 0.14, P < 0.01 ). Logistic regression analysis showed that BMI, WC and wBMI were correlated with HGS expression ( OR=2.00, 0.47, 1.22, P <0.05), while the association of WHtR and HGS expression had no statistical significance ( OR=0.51, P >0.05). The area under the curve (AUC) of BMI for predicting HGS expression in high school girls was the largest (AUC=0.65, 95% CI =0.61-0.69), followed by the AUC of wBMI for predicting HGS expression (AUC=0.65, 95% CI =0.57-0.66)( P <0.05). The most sensitivity was wBMI (71.5%), followed by BMI (69.4%), and the least sensitivity was WHtR (52.3%); and WC had the best specificity (61.1%), whereas wBMI had a relatively low specificity (52.5%).
Conclusion:For high school girls, BMI, WC and wBMI could predict the strenth of HGS expression, but BMI and wBMI have the highest accuracy and sensitivity, while wBMI has the lowest specificity in predicting the strenth of HGS.
