Establishment and validation of a predictive model for the treatment effect of temporomandibular joint dysfunction syndrome among students in a college of Xuzhou

Objective To establish a model for predicting the therapeutic effect of temporomandibular joint disorder syndrome (TMJDS) and validate its efficiency.Methods A total of 346 patients with TMJDS who were admitted to the Health Center of Xuzhou College of Industrial Technology from March 2019 to March 2023 were included.They were divided into in a training set (n=259) and a verification set (n=87) at a ratio of 3:1. According to the changes of clinical symptoms and TCM syndrome scores of patients in the training set before and after treatment, their therapeutic effect was evaluated, and the clinical data of patients with effective and ineffective treatment were compared, including age, gender, duration of illness, and mouth opening degree, etc. Logistic multivariate regression model analysis was used to analyze the risk factors affecting treatment effect.A nomogram model was established based on the risk factors, and internal validation, efficiency analysis and decision curve analysis (DCA) were conducted.Results The total effective rate in the training set and verification set was 88.03% and 88.51%, respectively, without statistically difference between the groups (P>0.05). Differences in duration of illness, mouth opening, unilateral mastication and treatment plan between the effective and ineffective patients in the training set were statistically significant (P<0.05). Logistic multivariate regression analysis showed that the increase of illness duration by one month, the decrease of mouth opening degree by 0.1 cm, and unilateral mastication were the independent risk factors for ineffective treatment in TMJDS patients, while active external application of Huoluo powder, infrared radiation and psychological counseling were protective factors (P<0.05). The internal validation results showed that the AUC of the training set and the verification set was 0.967 and 0.844, respectively. Hosmer-Lemeshow test showed that the calibration of the model was good (χ²=5.343, P=0.720). DCA results showed that the decision curves for the training and validation sets were located above both intervention and non-intervention lines within the threshold range of 5%~60%.Conclusions The ineffective treatment of patients with TMJDS may be related to illness duration, mouth opening, chewing habit and treatment plans. The nomogram model established on the risk factors demonstrates good efficacy and high net yield, providing reliable reference for predicting treatment effect.