基于Copula理论的医院感染率动态相关性分析

Dynamic correlation analysis of nosocomial infection rate based on Copula theory

目的从宏观管理研究中提升医院感染预防和控制水平。方法文章基于某医院2010年1月-2017年4月感染率数据及其时间连续性的特征,首次利用时变SJC Copula方法对超过全院平均感染率的胸外科、肝胆外科和神经外科感染率两两间的相关性进行研究。结果肝胆外科和胸外科、神经外科和肝胆外科、胸外科和神经外科间均是感染率低的相关性高于感染率高时的相关性,肝胆外科和胸外科,神经外科和肝胆外科、胸外科和神经外科间低感染率间的相关性,其均值分别为0.566,0.704,0.633,在置信度为95%的置信区间分别为(0.563,0.570)、(0.702,0.706)和(0.632,0.634);肝胆外科和胸外科高感染率间相关性的均值为0.418,在置信度为95%的置信区间为(0.403,0.432);总结原因与各科室病种分布特征、环境因素、手术操作、护理情况、抗菌药物管理及感控科干预等有直接关系。结论本研究首次利用Copula方法从宏观上建模分析科室间医院感染率发生的相关性,为医院感染管理提供了科学化的新思路,为医院感染管理的决策和科研提供真实有效的量化证据,对促进医院感染管理学科的发展也具有重要的现实意义。

OBJECTIVE To improve the level of prevention and control of nosocomial infections from macro management research.METHODS Based on the nosocomial infection rate data and the characteristics of time continuity of a hospital from Jan.2010 to Apr.2017,the time-varying SJC Copula method was used to investigate the correlations between department of thoracic surgery,department of hepatobiliary surgery and department of neurosurgery with the infection rate over the average infection rate in the hospital for the first time.RESULTS The correlations between department of hepatobiliary surgery and department of thoracic surgery,department of neurosurgery and department of hepatobiliary surgery,and department of thoracic surgery and department of neurosurgery were higher when the infection rates were low than those when the infection rates were high.The correlations between department of hepatobiliary surgery and department of thoracic surgery,department of neurosurgery and department of hepatobiliary surgery,and department of thoracic surgery and department of neurosurgery when the infection rates were 0.566,0.704,and 0.633,respectively,and confidence intervals of 95% were（0.563,0.570）,（0.702,0.706）and（0.632,0.634）,respectively.The mean correlation between department of hepatobiliary surgery and department of thoracic surgery with high infection rate was 0.418,and confidence interval of 95% was（0.403,0.432）.The reasons were directly related to the distribution of diseases,environmental factors,operation,nursing,antimicrobial agents management and intervention of department of infection control in each department.CONCLUSIONIn this study,the correlations between nosocomial infection rates in different departments are modeled and analyzed using the Copula method macroscopically for the first time,and it provides a new scientific idea for the management of nosocomial infections.In addition,the study provided real and effective quantitative evidence for the decision-making and scientific research of nosocomial infection management,which also has importantpractical significance to promote the development of nosocomial infection management discipline.