基于中断时间序列设计的分段回归模型评价抗菌药物管理防控院感的效果  

Segmented Regression Model based on Interrupted Time Series Design to Evaluate the Effect of Antimicrobial Management on the Prevention and Control of Nosocomial Infection

作　　者：赵执扬 刘静[1] 史楠 康建邦[1] 段金菊[1] 乔宇超 仇丽霞[1] Zhao Zhiyang;Liu Jing;Shi Nan(Department of Health Statistics,Shanxi Medical university,Taiyuan 030001)

机构地区：[1]山西医科大学卫生统计教研室,030001

出　　处：《中国卫生统计》2024年第6期830-833,839,共5页Chinese Journal of Health Statistics

基　　金：国家自然科学基金资助项目(81973155);山西省科技厅重大专项(201803D31066)。

摘　　要：目的应用中断时间序列设计的分段回归模型分析抗菌药物管理防控院感的长短期效果,为公共卫生干预措施的效果评价提供方法学上的参考。方法收集山西省某三甲医院2017年7月至2020年6月抗菌药物管理实施前后住院患者院内感染肺炎克雷伯菌(klebsiella pneumoniae,KPN)数据,采用调整异常值的中断时间序列的泊松分段回归模型进行分析。结果实施抗菌药物管理后短期干预效果较好,干预之后感染KPN的风险是干预前的81.25%(P<0.05),即感染KPN的风险下降了18.75%;但长期干预效果较差,时间每增加一个月,感染风险会增加1.55%;因新冠疫情期间重病患者占比大,院内KPN感染率升高,感染风险增加了68.81%。结论实施抗菌药物管理,仅在短期内降低了KPN感染的风险,需要长期严格执行才能持久地发挥作用。考虑了异常值的中断时间序列设计的分段回归模型既可以定量的评价短期、长期的干预效果,又可以对异常值导致的风险进行分析,为公共卫生干预措施的效果评价提供方法学上的参考。Objective The segmented regression model designed by interruption time series was used to analyze the short-term and long-term effects of antimicrobial management and prevention of nosocomial infection,so as to provide methodological reference for the effect evaluation of public health interventions.Methods The data of nosocomial infection with Klebsiella pneumoniae(KPN)in a third class hospital in Shanxi Province from July 2017 to June 2020 were collected and analyzed by Poisson piecewise regression model with interrupted time series adjusted for outliers.Results The short-term intervention effect was better after the implementation of antibiotics management.The risk of KPN infection after the intervention was 81.25%before the intervention(P<0.05),and the risk of infection with Klebsiella pneumoniae decreased by 18.75%.However,the effect of long-term intervention is poor.For each additional month,the risk of infection will increase by 1.55%.Because of the large proportion of serious ill patients during the COVID-19 period,the KPN infection rate in the hospital increased,and the infection risk increased by 68.81%.Conclusion The implementation of antimicrobial drug management only reduces the risk of KPN infection in the short term,which needs to be strictly implemented in the long term to play a lasting role.Considering the interruption time series of outliers,the piecewise regression model can not only quantitatively evaluate the short-term and long-term intervention effects,but also analyze the risks caused by outliers,so as to provide methodological reference for the effect evaluation of public health interventions.

关 键 词：中断时间序列设计 分段回归模型 抗菌药物管理 院内感染 克雷伯菌肺炎 

分 类 号：R195.1[医药卫生—卫生统计学]