Segmented Regression Model based on Interrupted Time Series Design to Evaluate the Effect of Antimicrobial Management on the Prevention and Control of Nosocomial Infection
10.11783/j.issn.1002-3674.2024.06.005
- VernacularTitle: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
- Author:
Zhiyang ZHAO
1
;
Jing LIU
1
;
Nan SHI
1
Author Information
1. 山西医科大学卫生统计教研室(030001)
- Publication Type:Journal Article
- Keywords:
Interrupt time series analysis;
Piecewise regression model;
Antibiotics management;
Nosocomial infection;
Klebsiella pneumonia
- From:
Chinese Journal of Health Statistics
2024;41(6):830-833,839
- CountryChina
- Language:Chinese
-
Abstract:
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.
