三叉神经痛患者疼痛灾难化风险预测模型的构建与验证*

Construction and verification of pain catastrophizing risk prediction model for trigeminal neuralgia patients

摘要:

目的:分析三叉神经痛(Trigeminal neuralgia, TN)患者疼痛灾难化的危险因素并建立风险预测模型,为临床有效预防TN患者疼痛灾难化的发生提供参考依据。方法:选取2021年1月至2023年3月在江西某三甲医院疼痛科住院的205例TN患者为研究对象,根据是否发生疼痛灾难化分为疼痛灾难化组和无疼痛灾难化组,采用单因素分析和多因素Logistic回析探讨疼痛灾难化的危险因素,通过R软件构建列线图风险预测模型并验证效果。结果: Logistic回归表明年龄、文化程度、疼痛程度、焦虑、抑郁、睡眠质量是TN患者疼痛灾难化的危险因素,Bootstrap法内部验证显示平均ROC曲线下面积(AUC)为0.978、C-Index为0.978,外部验证显示AUC为0.882、模型特异度为0.941、灵敏度为0.792,表明模型区分度较好,Calibrate曲线图显示模型校准度良好,DCA结果显示模型临床获益水平较高。结论:年龄、文化程度、疼痛程度、焦虑、抑郁、睡眠质量是TN患者疼痛灾难化的危险因素,该列线图风险预测模型具有良好的预测效能和临床应用价值。

Abstract:

Objective: To analyze the risk factors of pain catastrophizing in Trigeminal neuralgia (TN) patients and establish a risk prediction model, so as to provide reference for effectively preventing the occurrence of pain catastrophizing in clinic. Methods: A total of 205 TN patients hospitalized in the pain department of a Grade-A hospital in Jiangxi Province from January 2021 to March 2023 were selected as the research objects. According to whether pain catastrophizing occurred, they were divided into the pain catastrophizing group and the group without pain catastrophizing. Univariate analysis and multivariate Logistic regression were used to explore the risk factors of pain catastrophizing. R software was used to construct the line graph risk prediction model and verify the effect. Results: Logistic regression showed that age, education level, pain degree, anxiety, depression and sleep quality were risk factors for pain disaster in TN patients. Internal verification of Bootstrap method showed that the average area under ROC curve (AUC) was 0.978 and C-Index was 0.978. External verification showed that the AUC was 0.882, the model specificity was 0.941, and the sensitivity was 0.792, indicating good model differentiation. Calibrate curve graph showed good model calibration, and DCA results showed high clinical benefit level of the model. Conclusion: Age, education level, pain degree, anxiety, depression and sleep quality are risk factors for pain catastrophizing in TN patients. The risk prediction model of this column graph has good predictive efficacy and clinical application value.