分层线性模型在三级公立医院患者安全文化调查研究中的应用与探讨

Application on hierarchical linear model in hospital survey of patient safety culture in tertiary public hospitals

目的:探究影响医院患者安全文化的因素。方法:使用分层线性模型分析三级公立医院患者安全文化问卷调查数据;采用美国AHRQ患者安全文化调查问卷(HSOPSC)对深圳市11所三级医院进行随机调查,共回收1 621份有效问卷;使用SPSS v22和HML v7软件对数据进行统计分析。结果:问卷整体Cronbach'sα为0.873,信度良好;KMO取样适切性量数为0.929,Bartlett球形度检验显著性P<0.05,变量间偏相关性非常强;分层线性模型数据分析显示,影响患者安全文化的变量按重要性排序依次为科室间合作、科室内合作、不良事件报告频率、是否为主管;主管身份与患者安全文化呈负相关。结论:分层线性模型在分析较大样本且具有嵌套结构特征的患者安全文化数据时更加精准;研究结果为医院患者安全管理提供了参考依据。

Objective: To analyze factors that affect patient safety culture in hospitals. Method: A HSOPSC questionnaire was used to have a random investigation in 11 Shenzhen public hospitals. A total of 1,621 questionnaires were returned. Statistical analyses were conducted by SPSS v22 and HML v7 according to the nature of the survey data. Results: The overall Cronbach’s α of this questionnaire was 0.864, KMO was 0.941, Bartlett sphericity test was P<0.05, indicating that questionnaire has good internal validity. There is no significant difference in patient safety culture between Shenzhen tertiary public hospitals. From the Hierarchical Linear Model(HLM) analyses showing that the dimensionality of safety culture variables is ranked in order of importance: Teamwork Across Units > Teamwork Within the Units > Frequency of Error Reporting > Supervisory Identity. This study reveals for the first time that t Supervisory Identity is negatively correlated with the patient safety culture. Conclusions: HLM is more accurate when analyzing large sample like safety culture data with Nested Structure features. It gives evidence for patient safety management in hospital.