Construction and application of combination forecasting model for human resources in a large public hospital
Objective To understand the current status of healthcare human resources(HHR)in a large public hospital,predict the HHR demand aligned with the development of the hospital,and provide reference basis and feasible solutions for HHR planning for high-quality development of the large public hospital.Methods Based on grey model and auto regressive integrated moving average model,a variance reciprocal method for weight allocation was applied to set up the combination forecasting model.Different types of HHR demand of the large public hospital from 2024 to 2026 were predicted and the accuracies of the three different model predictions were compared.Results The numbers of total personnel,health technical personnel,physicians,nurses,and technicians predicted by the combination forecasting model for 2026 were 17 654,13 041,4389,6 198,and 2264,respectively.The corresponding average annual growth rates from 2024 to 2026 were 5.54％,5.55％,5.37％,4.27％,and 5.60％,respectively.Compared with the two single forecasting models,the combination forecasting model had the smallest average absolute errors,mean squared errors,and mean absolute percentage errors for predicting the numbers of total personnel,nurses,and technicians.It also had the smallest average absolute error and mean absolute percentage error for predicting the number of health technical personnel,and the smallest average absolute error for predicting the number of physicians.Conclusions Compared with the single forecasting model,the combination forecasting model shows fewer system errors and better predictive results.The demand for total personnel,health technical personnel,physicians,nurses,and technicians of this large public hospital will continue to increase,so planning and reserving staff in advance is a key to high-quality development of the hospital.
Human resourceslarge public hospitalgrey modelautoregressive integrated moving average modelcombination forecasting
万方数据
维普
