ARIMA模型预测医院感染发病状况研究

Application of ARIMA Model in Predicting Incidence Trend of Nosocomial Infection

目的探讨ARIMA季节乘积模型在时间序列资料中的应用,建立金坛市中医院医院感染发病率的预测模型。方法收集本院2005-2012年住院病人病案资料,应用SPSS18.0软件中的ARIMA模型预测模块对数据进行分析建立ARIMA预测模型,并预测2013年医院感染情况。结果 ARIMA(1,0,1)(0,1,1)4能够较好的拟合本院医院感染发病率情况,利用此模型预测2013年本院4个季度的医院感染率分别为2.67%、2.03%、2.68%和1.93%。结论 ARIMA模型能够较好的拟合和预测医院感染的发病情况,可以为医院决策提供科学依据。

Objective To explore the application of multiple ARIMA model in the time series data and establish a predic- tive model on nosocomial infection. Methods Information of inpatients from Traditional Chinese Medicine Hospital of Jintan City was collected, and then analyzed by ARIMA model predication module of SPSS 18.0 software. Nosocomial infection in 2013 was predicted by the established ARIMA predication model. Results ARIMA（ 1,0,1 ） （0,1,1）4 could fit the incidence of nos- ocomial infection, and the predicted incidence rates of nosocomial infection in four quarters in 2013 were 2.67%, 2.03%, 2.68 %, and 1.93 %, respectively. Conclusions ARIMA model can well fit and predict the changing trend of the incidence of nosocomial infection, which provides an effective basis for the hospital decision making.