某三甲中医医院ICU感染发生率时间序列分析及趋势预测
Time Series Analysis of ICU Infection Incidence Rates in a Tertiary TCM Hospital and the Trend Prediction
杨丽萍 1程立军 2李潇 3杨雳畯 3丁淑玉 4王靖研 4黄文莉 4毛宝宏5
作者信息
- 1. 甘肃省中医院公共卫生与医院感染管理处,甘肃 兰州 730050;甘肃中医药大学公共卫生学院,甘肃 兰州 730000
- 2. 甘肃省康泰医院,甘肃 兰州 730046
- 3. 甘肃省中医院公共卫生与医院感染管理处,甘肃 兰州 730050
- 4. 甘肃中医药大学公共卫生学院,甘肃 兰州 730000
- 5. 甘肃中医药大学公共卫生学院,甘肃 兰州 730000;甘肃省妇幼保健院/甘肃省中心医院医学教育部,甘肃 兰州 730050
- 折叠
摘要
目的:了解某三甲中医医院ICU感染发生率的时序分布特征,预测其发生规律和趋势,为中医医院ICU感染监测提供数据支持.方法:收集某三甲中医医院2019年1月至2024年2月ICU医院感染数据.利用求和自回归滑动平均模型(Autoregressive integrated moving average,ARIMA)对ICU感染发生趋势进行预测并评价其预测效果.结果:2019年1月至2024年2月某三甲中医医院ICU医院感染发生率为2.61%(232/8895);时间序列分析显示,ICU医院感染发生率波动较大且存在一定周期性,总体呈下降趋势.根据赤池信息准则和贝叶斯信息准则拟合,ARIMA(0,1,1)为最优预测模型.经参数估计与效果评价,感染发生率实际值均在预测值95%可信区间内,模型预测效果较好.结论:运用ARIMA对某三甲中医医院ICU医院感染发生率的预测结果良好,可显示其长期发生规律与趋势,能为医院感染监测提供科学依据.
Abstract
Objective:To investigate the temporal distribution characteristics of ICU infection incidences in a tertiary TCM hospital,and to predict the law and trend of its occurrence,so as to provide data support for the monitoring of ICU infection in TCM hospital.Methods:The data of nosocomial infection in ICU were collected from January 2019 to February 2024 in a tertiary TCM hospital,ARIMA model was utilized to predict the trend of ICU infection and predict its predicitive effects.Results:The incidence of nosocomial infection in ICU between January 2019 and February 2024 was 2.61%(232/8895)in a tertiary TCM hospital;time series analysis showed that the ICU hospital infection incidence rate fluctuated and had a certain periodicity with an overall decreasing trend.ARIMA(0,1,1)was the optimal prediction model according to the Akaike informativeness criterion(AIC)and Bayesian information criterion(BIC)fitting.After parameter estimation and effectiveness evaluation,the actual values of infection incidence were within the 95%confidence interval of the predicted values,and the model has a good predictive effect.Conclusion:ARIMA can be used to predict the incidence of hospital infection in ICU of a tertiary hospital,and it can show the long-term pattern and trend of the occurrence,which could provide scientific basis for the monitoring of hospital infection.
关键词
医院感染/重症监护病房/求和自回归滑动平均模型/时间序列/趋势预测Key words
nosocomial infection/ICU/ARIMA model/time series/trend prediction引用本文复制引用
基金项目
甘肃省自然科学基金(22JR5RA633)
陇原青年创新创业人才项目(2021LQTD10)
出版年
2024
国家科技期刊平台
NSTL
万方数据