Abstract
BACKGROUND: Hallux valgus (HV) is a common toe deformity with various contributory factors. The interactions between intrinsic risk factors of HV, such as arch height, sex, age, and body mass index (BMI) should be considered. The present study aimed to establish a predictive model for HV using intrinsic factors, such as sex, age, BMI, and arch height based on decision tree (DT) model. METHODS: This is retrospective study. The study data were based on the fifth Size Korea survey, of the Korea Technology Standard Institute. Among 5,185 patients, 645 were excluded due to unsuitable age or missing data, and 4,540 (males = 2,236 and females = 2,304) were selected for inclusion in the study. Seven variables (i.e., sex, age, BMI, and four normalized arch height variables) were used to develop the prediction model for the presence of HV using a DT model. RESULTS: The DT model correctly classified 68.79% (95% confidence interval [CI] = 67.25-70.29%) of the training data set (3,633 cases). The predicted presence of HV based on the DT was verified against the testing data set (907 cases) and showed an accuracy of 69.57% (95% CI = 66.46-72.55%). CONCLUSIONS: The DT model predicted the presence of HV on the basis of sex, age, and normalized arch height. According to our model, women aged over 50 years and those with lower normalized arch height were at high risk of HV.