摘要
目的 应用自回归移动平均模型(ARIMA)建立神经外科重症监护室(ICU)医院感染发病率的预测模型,为医院感染防控工作提供参考依据。方法 收集2012年1月-2021年2月上海交通大学医学院附属仁济医院神经外科ICU感染发病率数据,利用2012年1月-2019年12月的数据建立ARIMA模型,选取2020年1月-2021年2月的发病率数据评价模型的预测性能,并进行短期预测。结果 2012年1月-2021年2月神经外科ICU共有12 708例住院患者,总住院天数为52 397 d,医院感染人数为756例,日感染发病率为14.43‰。逐月的日感染发病率时间序列呈下降趋势,且有周期性。根据2012年1月-2019年12月医院发病数据,ARIMA(3,1,1)(1, 1,0)_(12)模型的拟合效果相对较优,2020年1月-2021年2月实际观测值均落在预测值95%可信区间内。利用全部数据重新拟合模型进行预测,结果提示该神经外科ICU的医院感染发病率将保持在较平稳的水平。结论 ARIMA模型对神经外科ICU患者日感染发病率的拟合效果较好,有助于制定医院感染管理相关防控措施。
OBJECTIVE To establish the model for prediction of incidence of nosocomial infection in intensive care unit(ICU) of neurosurgery department based on autoregressive integrated moving average(ARIMA) model so as to provide guidance for prevention and control of nosocomial infection. METHODS The data of incidence rates of infection were collected from the ICUs of neurosurgery department of Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine between Jan 2012 and Feb 2021, the ARIMA model was established based on the data from Jan 2012 to Dec 2019, and the data of incidence rates from Jan 2020 to Feb 2021 were collected to evaluate the estimated performance of the model and carry out a short-term prediction. RESULTS A total of 12 708 patients were hospitalized in the ICUs of neurosurgery department from Jan 2012 to Feb 2021, with the total length of hospital stay 52397 days, the number of cases of nosocomial infection 756, the daily incidence of infection 14.43‰. The time series of monthly daily infection incidence showed a downward trend and periodic change. According to the data from Jan 2012 to Dec 2019, the imitative effect of the ARIMA(3,1,1)(1,1,0)_(12)model was relatively good, the actual observed values from Jan 2020 to Feb 2021 fell within 95% confidence interval of the predicted values. All of the data were used to establish the fitting model for prediction, and the result indicated that the incidence of nosocomial infection in ICU of neurosurgery department maintained a stable level. CONCLUSION ARIMA model has an imitative effect on the daily incidence of infection in the ICU of neurosurgery department and may facilitate the development of relative prevention measures for nosocomial infection.
作者
杨亚
蔡冰超
沈瑞红
袁雅赟
傅小芳
班海群
YANG Ya;CAI Bing-chao;SHEN Rui-hong;YUAN Ya-yun;FU Xiao-fang;BAN Hai-qun(Renji Hospital Affiliated to Shanghai JiaotongUniversity School of Medicine,Shanghai 200127,China)
出处
《中华医院感染学杂志》
CAS
CSCD
北大核心
2022年第10期1500-1503,共4页
Chinese Journal of Nosocomiology
关键词
神经外科
重症监护室
医院感染
自回归移动平均模型
预测
Neurosurgery department
ICU
Nosocomial infection
Autoregressive integrated moving average
Prediction
