基于GA-BP神经网络的单病种成本预测研究

目的：针对单病种成本管理难度大、管理成本高等问题,通过对单病种成本的快速准确估算,实现病种成本的动态跟踪,为合理确定病种支付标准、有效降低医院运营成本提供重要保障。方法：利用矩阵形式表示单病种成本构成,构建单病种成本结构模型,把GA和BP算法相结合应用到单病种成本预测领域,并以天津市一所三甲医院阑尾炎病种进行分析。结果：建立GA-BP网络的单病种成本预测模型,表明该模型具有计算精度高、适用范围广、运算速度快的特点。结论：为单病种成本预测提供可靠依据,具有广泛的应用前景。

Method of single disease cost prediction based on GA-BP neural network

Objective To dynamically tracking down disease cost and to lay a solid foundation for defining payment standards of diseases and effectively reducing operating costs of hospitals by rapid and accurate estimation of single disease cost.Methods Using matrix to express the single disease cost composition,we established a single disease cost composition model,combined the genetic algorithm(GA) with the back-propagation(BP) algorithm and applied it to the single disease cost estimation,then takes data of appendicitis disease in one of Tianjin 3A hospitals as an example.Results The estimation model for single disease cost based on the GA-BP neural network had the advantages of high estimation accuracy,extensive applications and high operational speed.Conclusion Analysis with GA-BP method can provide an effective basis and great prospect for the single disease cost prediction.