冠心病病人二级预防服药依从性预测模型的构建  

Construction of prediction model for secondary prevention medication adherence in patients with coronary heart disease

作　　者：周艺 贾立敏[2] 霍奇文 冯翠娜[3] 杜少英[1] 娄士宁 石奇松[5] 王彦[1] ZHOU Yi;JIA Limin;HUO Qiwen;FENG Cuina;DU Shaoying;LOU Shining;SHI Qisong;WANG Yan(College of Nursing,Hebei University,Hebei 071000 China)

机构地区：[1]河北大学护理学院,河北071000 [2]保定市第五医院 [3]河北大学附属医院 [4]中国乐凯集团有限公司职工医院 [5]保定市第二医院

出　　处：《护理研究》2022年第11期1925-1930,共6页Chinese Nursing Research

基　　金：河北省重点研发计划项目,编号:18277735D。

摘　　要：目的:构建冠心病病人二级预防服药依从性预测模型。方法:选取2017年9月—12月在某市4所医院住院的冠心病病人356例,采用自制问卷调查冠心病病人服药依从性的影响因素。应用多元线性回归及回归树构建预测模型,采用平均绝对误差、均方误差和标准化后的平均绝对误差验证及比较预测模型的预测性能。结果:冠心病病人服药依从性得分为(14.51±3.31)分。多元线性回归预测模型拟合优度及预测性能优于回归树预测模型。模型为:服药依从性=-1.004+0.298×合理用药自我效能+0.754×对疾病的重视程度-0.903×担心药物副作用+0.527×服药管理+1.261×家庭所在地+0.257×疾病危险因素认知+0.970×工作状态+0.774×经皮冠状动脉介入术+1.363×冠状动脉旁路移植术。结论:构建的冠心病病人二级预防服药依从性预测模型具有较好的预测性能,但尚需进一步加大数据量,结合机器学习,开发手机APP,使临床预测更精准、应用更便捷。Objective:To construct the prediction model for secondary prevention medication adherence in patients with coronary heart disease(CHD). Methods:From September to December in 2017,a total of 356 CHD patients who were hospitalized in 4 hospitals in a city were selected as the research subjects. The influencing factors of medication compliance in CHD patients were investigated by using self-made questionnaire. Multiple linear regression and regression tree were used to construct prediction models. Mean absolute error,mean square error and normalized mean squared error were used to validate and compare the prediction performance of prediction models.Results:The mean score of medication adherence in CHD patients was 14. 51±3. 31. The goodness of fit of the multiple linear regression prediction model and prediction performance were better than the regression tree prediction model. The prediction model was that the medication adherence=-1. 004+0. 298×(self-efficacy for appropriate medication use)+0. 754×(the degree of attention on CHD)-0. 903×(worry about the side effect on medication)+0. 527×(medication management)+1. 261×(the place of residence)+0. 257×(the cognition on CHD risk factors)+0. 970×(work status)+0. 774×(PCI)+1. 363×(CABG). Conclusions:The prediction model for secondary prevention medication adherence in patients with coronary heart disease had good predictive performance. However,it is still necessary to further expand sample size,combine machine learning,and develop mobile Apps,in order to provide more accurate prediction and convenient application for clinical professionals.

关 键 词：冠心病 二级预防 服药依从性 预测模型 影响因素 

分 类 号：R473.5[医药卫生—护理学]