ROLE OF THE PHYSICAL FITNESS TEST IN RISK PREDICTION OF DIABETES AMONG URBAN RESIDENTS IN CHINA

摘要

Objectives: To explore the association between physical fitness parameters and risk prediction of diabetes among urban residents in Guangxi, China. Methods: This was a cross-sectional study conducted from July 2015 to December 2016 with urban adult residents recruited. All subjects received the 8-dormain Physical Fitness Test (PFT) and EZSCAN (ESC) screening for diabetes. Results from these domains were entered into the logistic regression model of the stepwise logistic regression analysis method. The diagnostic accuracies were determined by calculating the area under the receiver-operating characteristics (ROC) curve. Results: Among 5668 subjects (age at 42.9 ± 12.3) were 2984 (52.6%) males and 3998 (70.5%) females. There were 2268 (39.6%) subjects who showed diabetes-indicated ESC value (ESC≤60uS); while other 3400 subjects (60%) were with normal ESC value. It was found that systolic blood pressure, Harvard step test index, bending-forward in sitting, hand grip strength, vertical jump height, time of single-leg-standing with eyes closed, choice reaction time (CRT) playing a game, and body composition minerals were all independently associated with diabetes. From the resulting equation for the prediction of the disease, diabetes had an area under the ROC curve (AUC) of 0.808 with a sensitivity of 85.44% and specificity of 66.09%. In all PFT parameters, CRT demonstrated an independent predictor of the disease. Conclusion: For urban residents the parameter of 8 PFT domains were associated with the ESC screened values of diabetes, particularly the choice reaction time may have most sensitive and specific value in possibly predicting the warning of diabetes.