摘要
目的:探讨珠海市结肠癌患者住院费用的影响因素,为合理控制结肠癌患者住院费用提供参考。方法:收集并分析珠海市2017年1月1日至2022年12月31日期间结肠癌患者的住院数据,采用梯度提升机模型识别结肠癌患者住院费用的影响因素,数据按7∶3随机分为训练集和测试集,基于一致性指数(C-index)、灵敏度、特异度和准确率等指标衡量模型拟合的效果,绘制受试者工作特征(ROC)曲线,以评估模型的区分能力。结果:共纳入3475例结肠癌住院患者,住院总费用的中位数为44899.7元。梯度提升机模型结果显示:结肠癌患者住院费用影响因素前5的权重:手术级别(0.599)、住院天数(0.196)、医院级别(0.075)、术前等候天数(0.006)、年龄(0.006)。模型准确率为81.87%,C-index值为0.9064,ROC曲线下面积值为0.9064,模型拟合效果良好。结论:梯度提升机在预测结肠癌住院费用方面表现优越,手术级别是最关键因素,其次为住院天数、医院级别等。建议医疗机构提升结肠癌患者临床路径入径率,优化结肠癌的单病种付费标准,通过推广日间手术、日间化疗等举措,简化诊疗流程,提高医疗资源利用效率,在保证医疗质量的同时合理控制医疗费用。未来仍需从更全面的变量选取、更广泛的地域验证、更高效的模型解释性及智能化分析方法等方面进一步拓展,以提升研究的普适性和政策参考价值。
Objective:To explore the influencing factors of the hospitalization costs of colon cancer patients in Zhuhai city,and to provide reference for controlling hospitalization costs of colon cancer patients.Methods:The hospitalization information of colon cancer patients in Zhuhai City from January 1,2017 to December 31,2022 was collected and analyzed.A gradient boosting machine model was used to identify the influencing factors of hospitalization costs of colon cancer patients,the data were randomly divided into training set and test set according to 7∶3,and the fitting effect of the model was measured by C-index,sensitivity,specificity,accuracy and receiver operating characteristic(ROC)curve to assess the discriminative ability of the model.Results:A total of 3475 colon cancer inpatients were included,with a median total hospitalization cost of 44899.7 yuan.The GBM model revealed the top five influencing factors by weight:surgery level(0.599),hospitalization days(0.196),hospital level(0.075),preoperative waiting days(0.006),and age(0.006).The model achieved 81.87%accuracy,a C-index of 0.9064,and an ROC area under the curve(AUC)of 0.9064,indicating excellent performance.Conclusion:The gradient boosting machine demonstrated superior predictive capability for colon cancer hospitalization costs,with surgery level being the most critical factor,followed by hospitalization duration and hospital level.Recommendations include enhancing clinical pathway adherence,optimizing single-disease payment standards,promoting daytime surgeryand daytime chemotherapyto streamline care processes,and improving resource efficiency while maintaining quality.Future research should expand variable selection,validate findings across broader regions,and enhance model interpretability for greater generalizability and policy relevance.
作者
刘砾雯
韩立丰
张文洁
王钰瑜
张璐
肖颖
刘军卫
江鸿
LIU Liwen(School of Business,Macao University of Science and Technology,Macao,999078,China)
出处
《医学与社会》
北大核心
2025年第3期25-31,共7页
Medicine and Society
基金
中央财政医疗服务和保障能力提升(公立医院综合改革)项目,编号为YC202306
广东省科学技术厅国际科技合作项目,编号为112175605105。
关键词
住院费用
结肠癌
影响因素
梯度提升机
珠海
Hospitalization Cost
Colon Cancer
Influencing Factor
Gradient Lifting Machine
Zhuhai
