Analysis of factors influencing hospitalization cost based on back propagation neural network model

Abstract

Objective To analyze the influencing factors of hospitalization cost by back propagation(BP)neural network. Methods Inpatients with a total of 50 611 were collected from the hospital information system in a tertiary hospital. BP Neural network modeling was developed by SPSS 22.0 software to study the factors which influence the hospitalization cost. Results The hospitalization expenses prediction accuracies of training sample and test sample were 81.1% and 79.0% respectively. The top three factors were pharmaceutical cost(100.0%), general medical operation cost(88.8%)and disposable medical materials cost(60.7%). Conclusions The results show that BP model is suitable for dependent variable with the hospitalization expenses. The pharmaceutical cost, disposable medical materials cost and general medical operation cost were greater influencing factors than others. We should take actions by improving the treatment level and measures, so as to reduce the economic burden of patients and society Key words: Back propagation neural network; Models; Hospitalization cost; Influencing factors