Abstract:

Objective This study aims to develop and validate a dynamic web-based nomogram for predicting kinetophobia in patients following percutaneous coronary intervention（PCI）. Methods A prospective design was employed to selectively enroll 330 PCI patients admitted to a hospital in Hangzhou from December 2022 to July 2023. Single-factor analysis and Lasso regression were utilized to identify independent risk factors for kinesophobia post-PCI. Logistic regression was performed using R software，and a nomogram was constructed. The model was assessed through the area under the receiver operating characteristic curve（AUC） and Hosmer-Lemeshow tests. Results There were 206 cases of kinesiophobia in 330 patients after PCI，and the incidence was 62.4%. Logistic regression analysis identified combined heart failure，emergency surgery，NYHA cardiac function grade，ADL level，sedentary behavior，Chinese version of PROMIS Physical Function Summary Table score，and Chinese version of Perceptive Social Support Scale score as independent influencing factors for kinesophobia after PCI（P<0.05）. The AUC value of the model was 0.821，with a sensitivity of 70.4% and specificity of 82.0%. The Hosmer-Lemeshow fit test yielded a non-significant result（χ²=9.350，P=0.314）. Calibration and decision curves demonstrated the model’s favorable calibration and clinical practicability. The C-index of the nomogram prediction model was 0.778，0.774，and 0.800，respectively，by 5-fold cross-validation，10-fold cross-validation，and the Bootstrap method. Conclusion The dynamic nomogram model developed in this study effectively predicts kinesophobia in patients after PCI. It provides valuable references and support for clinical staff in early identification of high-risk patients，enabling the formulation of individualized health education strategies and exercise rehabilitation plans.

Key words: Percutaneous Coronary Intervention, Kinesiophobia, Nomogram, Prediction Model, Lasso Regression, Nursing Care

摘要：

目的 构建和验证经皮冠状动脉介入治疗（percutaneous coronary intervention，PCI）术后患者恐动症的预测模型。方法 采用前瞻性研究设计，便利选取2022年12月—2023年7月杭州市某三级甲等综合医院收治的330例PCI术后患者作为调查对象。采用单因素分析和Lasso回归筛选PCI术后患者恐动症发生的影响因素，基于R软件进行Logistic回归分析并构建列线图。应用受试者操作特征曲线和Hosmer-Lemeshow检验验证模型的预测效果。结果 330例PCI术后患者中有206例出现恐动症，发生率为62.4%。Logistic回归分析显示，合并心力衰竭、急诊手术、纽约心脏协会心功能分级、自理能力分级、久坐行为、中文版患者报告结局测量信息系统身体功能简表6b评分及中文版领悟社会支持量表评分是PCI术后患者恐动症发生的影响因素（P<0.05）。构建的列线图模型受试者操作特征曲线下面积为0.821，灵敏度为70.4%，特异度为82.0%，Hosmer-Lemeshow检验χ²=9.350，P=0.314，校准曲线和决策分析曲线表明模型具有良好的校准度和临床实用性。采用5折交叉验证、10折交叉验证和Boostrop法进行内部验证，模型的平均C指数分别为0.778、0.774和0.800。结论 构建的预测模型可有效预测PCI术后患者的恐动症，对于临床医护人员早期识别高风险患者，制订个体化的健康教育策略和运动康复计划具有指导价值。

关键词: 经皮冠状动脉介入治疗, 恐动症, 列线图, 预测模型, Lasso回归, 护理