Objective To establish an optimal model for demand prediction of drugs with high value and easy fluctuation by ARIMA model,so as to build a prediction platform and continuously capture the demand trend of fluctuated high-value drugs.Methods ABC classification and seasonal coefficient classification were used to screen out fluctuated high value drugs from the drug catalogue of our hospital.According to the drug consumption data from 2014 to 2022,the optimal ARIMA model of each variety was determined by the minimum principle of BIC,and the batch prediction platform was constructed in Excel according to the ARIMA principle.Results 51 high-value and volatile drugs were screened out,accounting for 50.82％of the drug consumption in 2022.ARIMA model fit index stationary R-square and average absolute percentage error get ideal values,residual and parameter pass white noise test and significance test respectively.The model was judged by sustainability,and the batch prediction platform was successfully built.Conclusion ARIMA model can be applied to judge the trend of demand for drugs with high value and easy fluctuation.The batch prediction platform based on the ARIMA model can efficiently,accurately and continuously serve the execution of such drug procurement plans.
维普
万方数据