基于梯度下降法的医院平均住院日内部优化路径研究  

Research on Internal Optimization Path of Average Length of Stay Based on DRGs Data and Gradient Descent Method

作　　者：庚硕 黄鹤妹 刘瑾 杨棋 段文厚 曹慧银 Geng Shuo;Huang Hemei;Liu Jin(The First Affiliated Hospital of Shandong First Medical University,Jinan,Shandong,250014,China)

机构地区：[1]山东第一医科大学第一附属医院,山东济南250014

出　　处：《中国医院管理》2022年第12期47-49,共3页Chinese Hospital Management

摘　　要：目的以某三级甲等医院疾病诊断相关分组(DRG)病种统计数据为研究基础,提出了一种优化平均住院日的新思路和新方法。方法采用机器学习的梯度下降法分析病种分组数据,选择各病种具有显著性的较低平均住院日作为优化目标。结果编程并计算得到各科室各病种平均住院日优化目标及路径,在此方法下分析得到全院综合平均住院日可实现的最优结果。结论机器学习的梯度下降法适合于医院管理中的优化求解问题。Objective A new idea and method for optimizing the average length of stay is proposed based on the statistical data of disease diagnosis-related groups(DRG)in a tertiary care hospital.Methods The gradient descent method of machine learning was used to analyze the patient grouping data and select the lower average hospital stay with significance for each patient group as the optimization target.Results The optimization target and path of average hospitalization days of each department and disease are calculated.Under this method,the optimal results of the comprehensive average length of stay of the whole hospital are obtained.Conclusion Gradient descent method is suitable for solving optimization problems in hospital management.

关 键 词：疾病诊断相关分组 平均住院日 梯度下降法 

分 类 号：R197.323.2[医药卫生—卫生事业管理]