Research On Forecast Model Of Medical Low-value Consumables Demand

As for the development of medicine,every major breakthrough is closely related to the clinical application of new consumables.With the development of technology,many new consumables are widely used in clinical medicine,leading to a variety of hospital medical consumables and the use of large quantities.As a result,consumables account for a big proportion of hospital spending.This requires the hospital to fully understand the use of medical consumables in various departments,and purchase consumables according to its needs to avoid waste of resources.Therefore,scientific methods are used to accurately predict the demand for medical consumables,so as to develop a sound medical consumable management system to guide the rational use of consumables.According to the relevant literatures,the research on medical consumables management mainly focus on the management of high-value consumables,such as the use and traceability of high-value consumables,while the research literature on low-value consumables management is less.Based on this,the paper takes the prediction of medical low-value consumables as the research object,establishes a more accurate prediction model of medical low-value consumables demand,provides scientific basis for the procurement and inventory management of medical consumables,and promotes information-based and refined management in hospital.The main research contents are as follows: First,by comparing the advantages and disadvantages and applicability of various prediction methods,this paper verified several different types of low-value consumables and four single-item models with higher accuracy were selected,namely LAM,DES and improved GM(1,1)and ARIMA;Secondly,on the basis of the selected single prediction model construction method,the weight of the single prediction model is calculated by using the principle of “minimum square of relative error” to realize the real time adjustment of weight coefficients.The optimal combination prediction model is called linear variable weight combination forecasting model.Finally,an verification(MAPE=4.1%)was performed using the safety indwelling needle consumable data.In addition,the monthly tracking and quarterly forecast analysis of the consumables were further verified(MAPE=4.3%),indicating that the combined forecasting model has better robustness.