从预测到理解：机器学习在住院费用预测中的应用进展
摘要
DRG/DIP支付方式改革的全面推开对医院精细化管理提出更高要求，住院费用预测将成为医疗成本管控、医保基金科学调控与资源配置优化的关键环节。机器学习方法因其强大的模式识别与泛化能力而受到广泛关注，本文系统梳理机器学习在住院费用预测领域的研究进展，阐述相关主流算法的应用场景与性能表现。研究表明，集成学习算法在预测精度上普遍优于单一模型，SHAP等解释性框架的引入有效提升了模型的临床可信度。未来应着力于多中心数据验证、深度学习方法及预测模型与DRG/DIP分组机制的深度融合，为住院费用精细化管理提供更坚实的理论支撑与决策依据。
关键词
参考
国家医疗保障局关于印发DRG/DIP 支付方式改革三年行动计划的通知_国务院部门文件_中国政府网[EB/OL]. [2026-03-19]. https://www.gov.cn/zhengce/zhengceku/2021-11/28/content_5653858.htm.
刘瑶瑶. DRG支付背景下临床医生诊疗行为影响机理及对策研究[D]. 安徽医科大学.
TAO Z, HUILING L, WENWEN W, 等. GA-SVM based feature selection and parameter optimization in hospitalization expense modeling[J/OL]. Applied Soft Computing, 2019, 75: 323-332. DOI:10.1016/j.asoc.2018.11.001.
石菊, 王小倩, 王熙. 微观医疗费用预测模型:从线性回归到机器学习[J]. 经济学（季刊）, 2023, 23(6): 2247-2263.
王倩, 邹全. 基于机器学习的心脑血管疾病带病体个人自付医疗费用和保险赔付预测[J]. 保险研究, 2024(7): 51-63.
孙冲. 多种机器学习算法预测呼吸系统疾病住院费用效果比较分析[D]. 吉林大学.
ORJI U, UKWANDU E. Machine learning for an explainable cost prediction of medical insurance[J]. Machine Learning with Applications, 2024, 15(c): 100516.
ARNOLD M, LIOU L, BOLAND M R. Development, evaluation and comparison of machine learning algorithms for predicting in-hospital patient charges for congestive heart failure exacerbations, chronic obstructive pulmonary disease exacerbations and diabetic ketoacidosis[J/OL]. BioData Mining, 2024, 17(1): 35. DOI:10.1186/s13040-024-00387-9.
OZAIR A, MUKHERJEE D. P10.19.A MACHINE LEARNING-BASED PREDICTION OF INPATIENT CARE COSTS FOR PATIENTS WITH BRAIN METASTASES: A REAL-WORLD ANALYSIS[J]. Neuro-Oncology, 2024, 26(Sup5): 2.
ZHEN Z. Prediction model and scoring tool for hospitalization costs in adult patients with lung cancer[J/OL]. Journal of University of Electronic Science and Technology of China, 2025, 54(2): 305-310. DOI:10.12178/1001-0548.2024054.
KIREMIT B Y, ŞAHIN D Ö. Comparison of machine learning algorithms for predicting length of stay in chronic kidney disease patients[J/OL]. Computers in Biology and Medicine, 2025, 196: 110825. DOI:10.1016/j.compbiomed.2025.110825.
MALHOTRA P, JHA R, MALIK K. MSR58 Comparing Predictive Models: Traditional Versus Modern Statistical Approaches for Economic Outcomes in Major Depressive Disorder[J]. Value in Health, 2024, 27(6-Sup): 1.
DHULIPALA S, MARAM B, RAO R, 等. A Comparative Analysis of Deep Learning and Ensemble Methods for Time Series Forecasting[J]. 2025 6th International Conference on Mobile Computing and Sustainable Informatics (ICMCSI), 2025: 1713-1716.
GONG K, XUE Y, KONG L, 等. Cost prediction for ischemic heart disease hospitalization: Interpretable feature extraction using network analysis[J/OL]. Journal of Biomedical Informatics, 2024, 154: 104652. DOI:10.1016/j.jbi.2024.104652.
夏泳, 陈任. 基于BP神经网络模型分析的老年人慢性病住院费用及影响因素[J]. 山东第一医科大学(山东省医学科学院)学报, 2022(005): 043.
ZAFAR S, HAJIANI M A, AFFAN M, 等. Enhancing Accuracy in Pediatric Inpatient Hospital Cost Estimation: A Machine Learning-Based Approach[J]. 2023.
陈默, 蔡苗, 黄阿红, 等. 基于K-means聚类与支持向量机的大病患者住院费用影响因素与控制策略研究[J]. 中国医院管理, 2019, 39(5): 4.
MAYYA V, KING C, VU G T, 等. Empirical Study of Feature Selection Methods in Regression for Large-Scale Healthcare Data: A Case Study on Estimating Dental Expenditures[J/OL]. IEEE Access, 2024, 12: 153564-153579. DOI:10.1109/ACCESS.2024.3482192.
TAKEFUJI Y. Limitations of XGBoost-SHAP integration for interpretable machine learning in antimicrobial resistance prediction[J]. Journal of Infection, 2025, 91(3).
LUNDBERG S, LEE S I. A Unified Approach to Interpreting Model Predictions[J]. 2017.
AAS K, JULLUM M, LLAND A. Explaining individual predictions when features are dependent: More accurate approximations to Shapley values[J]. 2021.
涂朗, 李树蕾, 樊若曦, 等. 缺血性心脏病的住院费用预测:基于SHAP理论的机器学习[J]. 2025(8): 881-887.
Refbacks
- 当前没有refback。