Page 89 - 《中国药房》2026年4期
P. 89

·药学服务·


          初治冠心病患者在医院-家庭过渡期的用药偏差预测模型构建与
          验证     Δ



          李玉双    1, 2* ,李 殊 ,张倩影 ,黄 炎 ,刘 坤 ,谷秀林 ,蒋欢欢 (1. 华北理工大学附属医院药学部,河北
                                                     1
                            1
                                    1
                                                                     1 #
                                            1
                                                             3
          唐山 063000;2. 华北理工大学药学院,河北 唐山 063210;3. 河北北方学院附属第二医院药剂科,河北
          张家口 075100)
          中图分类号  R972;R543      文献标志码  A      文章编号  1001-0408(2026)04-0491-06
          DOI  10.6039/j.issn.1001-0408.2026.04.14

          摘  要  目的  开发初治冠心病患者医院-家庭过渡期用药偏差风险预测模型,以助力医务人员快速识别用药偏差高危人群。方法
          纳入2024年1-7月华北理工大学附属医院(以下简称“我院”)的462例初治住院冠心病患者。将患者随机分为建模组与内部验
          证组。建模组患者依据是否发生用药偏差,分为用药偏差组和非用药偏差组。同法收集2025年6-9月我院心血管内科的57例
          初治住院冠心病患者作为外部验证组。采用单因素分析筛选预测因子,进一步通过多因素Logistic回归分析构建预测模型,并采
          用内部验证方法评估模型性能,采用外部验证方法检验模型的泛化能力。结果  462例患者被分为建模组(319例)和内部验证组
         (143例)。在建模组中,用药偏差组有192例(占比60.19%),非用药偏差组有127例(占比39.81%)。多因素Logistic回归分析结果
          显示,年龄、药品种类、服药依从性、合理服药自我效能是初治冠心病患者用药偏差发生的预测因子(P<0.05),预测模型方程为
          logitP=ln[P/(1-P)]=1.321+1.732×年龄+4.091×药品种类-4.360×服药依从性-3.081×合理服药自我效能。模型区分度良
          好,Hosmer-Lemeshow检验拟合的P值为0.439,受试者工作特征曲线下面积(AUC)为0.870,灵敏度为0.970,特异度为0.607;绘制了
          总分为350分、截断值为110分的风险列线图;内部验证组患者的AUC为0.787,预测准确率为77.6%;外部验证组患者的AUC为
          0.802,预测准确率为73.7%。结论  本研究成功构建了初治冠心病患者医院-家庭过渡期用药偏差风险预测模型,此模型具备良好的
          区分度与预测准确率,识别出高龄(>70岁)、药品种类≥5种、服药依从性差、合理服药自我效能差为用药偏差的独立危险因素。
          关键词  冠心病;用药偏差;医院-家庭过渡期;影响因素;预测模型

          Construction  and  validation  of  a  medication  deviation  prediction  model  for  hospital-to-home  transition
          period in coronary heart disease patients with initial treatment
                                               1
                                                            1
          LI Yushuang ,LI Shu ,ZHANG Qianying ,HUANG Yan ,LIU Kun ,GU Xiulin ,JIANG Huanhuan(1. Dept. of
                             1
                     1, 2
                                                                      1
                                                                                 3
                                                                                                  1
          Pharmacy,  the  Affiliated  Hospital  of  North  China  University  of  Science  and  Technology,  Hebei  Tangshan
          063000,  China;2.  School  of  Pharmacy,  North  China  University  of  Science  and  Technology,  Hebei  Tangshan
          063210,  China;3.  Dept.  of  Pharmacy,  the  Second  Affiliated  Hospital  of  Hebei  Northern  University,  Hebei
          Zhangjiakou 075100, China)
          ABSTRACT   OBJECTIVE  To  develope  a  predictive  model  for  medication  deviation  risks  during  the  hospital-to-home  transition
          period in coronary heart disease (CHD) patients with initial treatment, aiming to assist medical staff in rapidly identifying high-risk
          groups  for  medication  deviation.  METHODS  A  total  of  462  CHD  patients  with  initial  treatment  from  the  Affiliated  Hospital  of
          North China University of Science and Technology (hereinafter referred to as “our hospital”) between January and July 2024 were
          enrolled.  The  patients  were  randomly  divided  into  a  modeling  group  and  an  internal  validation  group.  The  modeling  group  was
          further categorized into a medication deviation group and a non-medication deviation group based on whether medication deviations
          occurred.  Similarly,  57  CHD  patients  with  initial  treatment  from  the  cardiology  department  of  our  hospital  between  June  and
          September 2025 were collected as an external validation group. Univariate analysis was used to screen predictive factors, followed
          by  multivariate  Logistic  regression  to  construct  the  predictive  model.  Internal  validation  methods  were  employed  to  evaluate  model
          performance,  while  external  validation  methods  were  used  to  test  the  model’s  generalizability.  RESULTS  The  462  patients  were
          divided  into  a  modeling  group (319  cases)  and  an  internal  validation  group (143  cases).  In  the  modeling  group,  the  medication
          deviation  group (192  cases,  60.19%)  and  the  non-medication  deviation  group (127  cases,  39.81%)  were  identified.  Multivariate
                                                             Logistic  regression  analysis  revealed  that  age,  medication
             Δ 基金项目 河北省医研企联合创新专项课题(No.LH20250181)
             * 第一作者 硕 士 研 究 生 。 研 究 方 向 :临 床 药 学 。 E-mail:   type,  medication  adherence,  and  self-efficacy  in  rational
                                                             medication  use  were  predictive  factors  for  medication
          15776262164@163.com
             # 通信作者 主任药师,硕士生导师。研究方向:临床药学、药物经                 deviations  in  CHD  patients  with  initial  treatment (P<0.05).
          济学、老年慢病管理。E-mail:jianghuan1001@163.com             The predictive model equation was logitP=ln[P/(1-P)]=1.321+


          中国药房  2026年第37卷第4期                                                 China Pharmacy  2026 Vol. 37  No. 4    · 491 ·
   84   85   86   87   88   89   90   91   92   93   94