摘要
目的探索ARIMA模型在预测医院感染发病率趋势中的应用价值,为进一步防控工作提供指导。方法回顾性采集海南某医院2009-2018年住院患者的医院感染发病率数据,构建ARIMA模型并评价模型精度,预测2019年7月至2020年6月的医院感染月发病率。结果本次研究最终确定ARIMA(0,1,1)模型为最优模型(BIC=-2.068,MAPE=10.574),残差序列Ljung-Box检验为白噪声(Q=9.864,P=0.909),模型拟合精度良好,预测2019年7月-2020年6月的月平均发病率为2.79%。结论ARIMA(0,1,1)模型拟合该院医院感染发病率效果较好,可用于短期预测发病趋势和早期预警,有助于指导医院感染宏观防控决策的制定。
Objective To explore the application value of ARIMA model in predicting the trend of nosocomial infection incidence,and to provide guidance for further prevention and control.Methods The data on nosocomial infection incidence in hospitalized patients in a hospital of Hainan from 2009 to 2018 was retrospectively collected,and an ARIMA model was constructed and the accuracy of this model was evaluated to predict the monthly incidence of nosocomial infections from July2019 to June 2020.Results In this study,the model ARIMA(0,1,1)was finally determined to be the optimal model(BIC=-2.068,MAPE=10.574),and the Ljung-Box test of residual sequence of this model was white noise(Q=9.864,P=0.909)with a good fitting accuracy.The average monthly incidence from July 2019 to June 2020 was predicted to be 2.79%.Conclusion ARIMA(0,1,1)model is effective in forecasting the incidence of nosocomial infections in this hospital.It can be used for short-term prediction of the incidence trend and early warning,and is helpful to guide the formulation of macro prevention and control of nosocomial infection.
作者
樊雯婧
陈健
楼冬洁
欧万秋
卢新
陈海霞
徐海群
鲜于舒铭
FAN Wen-jing;CHEN Jian;LOU Dong-jie;OU Wan-qiu;LU Xin;CHEN Hai-xia;XU Hai-qun;XIAN-YU Shu-ming(Hainan Provincial People's Hospital,Haikou Hainan 570311,China)
出处
《中国消毒学杂志》
CAS
2020年第9期670-673,678,共5页
Chinese Journal of Disinfection
关键词
医院感染
ARIMA模型
流行病学分析
预测
nosocomial infection
ARIMA model
epidemiological analysis
prediction
