ARIMA和VARMA模型在医院管理中的应用

Applying ARIMA and VARMA Models to Hospital Management

目的通过对医院门诊量、急诊量和住院量三项指标的预测,对医院运营状况进行早期预警,评价医院的经营策略。方法通过对某三级医院1999—2008年月门、急诊量、住院量分析,分别为门诊量、急诊量和住院量建立ARIMA模型,以整体数据建立VARMA模型,然后用上述模型对某医院2008年1~12月的门诊量、急诊量和住院量预测并和实际值比较。结果 AR IMA模型预测急诊量、住院量数据可靠,平均相对误差分别为5.63%和3.11%,在预测门诊量时不太理想,平均相对误差达到13%;VARMA模型在预测整体数据时良好,预测门诊量改善尤其明显,平均相对误差降低到8.97%,在预测急诊量、住院量,平均相对误差有所增加。结论单用某一预测方法时,总存在着一定的缺陷,综合应用AR IMA和VARMA模型可为医院管理者提供合理的决策。

Objective The amount of outpatient,inpatient and emergency treatment is seasonal time series,and they are strong correlated.Methods The author built up the ARIMA models for the amount of outpatient,inpatient and emergency treatment respectively,and VARMA model to treat 3 group data as a whole.The forecasting amount of outpatient,inpatient and emergency treatment in 2008,and compared it with the real data.Results ARIMA models can stably and reliably forecast the inpatient and emergency treatment,the relative average error is about 5.63% and 3.11% respectively.Conclusion VARMA model performed well in predicting the whole data,and the improvement in forecasting the amount of outpatient was obvious.And it was useful to use the two models for hospital management.