首页|基于分位数回归和随机森林的前列腺癌患者住院费用影响因素分析

基于分位数回归和随机森林的前列腺癌患者住院费用影响因素分析

扫码查看
目的:了解前列腺癌患者住院费用的影响因素,为有效降低因开展前列腺癌根治术承受的直接经济疾病负担提供参考依据.方法:申请获取湖北省武汉市某顶级三甲医院从2019年1月1日至2024年6月30日间开展前列腺癌根治术的住院患者病案数据893条.采用Stata 18.0软件构建数据库,使用SPSS 27.0和R 4.4.1软件进行Mann-Whitney U检验和Kruskal-Wallis H检验,并采用多重线性回归、分位数回归和随机森林算法对影响因素进行分析.结果:前列腺癌患者个人自付金额占住院总费用的64.40%,平均住院费用为60 794.9元,其中人均治疗费(2 0631.6元)、人均药品费(15 046.2元)和人均耗材费(12 469.6元)为排名前3的费用类型,分别占33.94%、24.75%和20.51%.多重线性回归分析结果显示,住院天数(β=0.027)、合并症数量(β=0.008)、入院途径(以门诊为对照,急诊β=0.128,其他医疗机构转入3=0.251)和手术方式(以腹腔镜前列腺癌根治术为参照,机器人辅助腹腔镜前列腺癌根治术β=0.157)对次均住院费用的影响均有统计学意义(P<0.05);分位数回归分析结果显示,住院天数和手术方式在不同分位点均产生了影响(P<0.001);随机森林回归分析结果显示,重要性排名前2的影响因素为住院天数和手术方式.结论:该研究中前列腺癌患者住院负担较重,且费用构成有待优化,应推动医疗耗材进入医保和加快仿制药研发,实施手术机器人的国产化替代,合理缩短患者住院时间以减少住院费用.
Factors influencing the hospitalization expenses of prostate cancer patients:An analysis based on quantile regression and random forest
Objective:To investigate the factors influencing the hospitalization expenses of PCa patients and provide some evi-dence for effectively reducing the direct economic burden of disease caused by laparoscopic radical prostatectomy(LRP).Methods:The author obtained the hospitalization data on 893 cases of PCa treated by LRP in a top tertiary hospital in Wuhan between January 1,2019 and June 30,2024.A database was constructed using the Stata 18.0 software,the Mann-Whitney U and Kruskal-Wallis H tests performed using the SPSS 27.0 and R 4.4.1 softwares,and the influencing factors for the hospitalization expenses of the patients ana-lyzed by multiple linear regression,quantile regression,and random forest algorithm.Results:Out-of-pocket payment accounted for 64.40%of the total hospitalization costs,averaging ¥60 794.9,with the per capita treatment cost(¥20 631.6),drug cost(¥ 15 046.2)and consumable cost(¥ 12 469.6)as the top 3 items of expenditure,accounting for 33.94%,24.75%and 20.51%,respectively.The results of multiple linear regression analysis showed that the number of hospital days(β=0.027),the number of comorbidities(β=0.008),the mode of admission(with the outpatient department as reference,the emergency depart-ment[β=0.128]and transfer from other institutions[β=0.251]),and the mode of surgery(with LRP[β=0.157]as refer-ence)all had statistically significant impacts on the sub-per capita hospitalization costs(P<0.05).Quantile regression analysis re-vealed dramatic influence of the number of hospital days and surgical mode in different quantiles(P<0.001),and random forest re-gression analysis showed the two influencing factors ranking the top two in terms of importance.Conclusion:PCa patients bear a heavy burden of hospitalization costs.The composition of the costs needs to be optimized,and efforts should be made to reduce the hos-pitalization expenses of the PCa patients by promoting the inclusion of medical consumables in medical insurance,facilitating the devel-opment of drugs and domestication of surgical robots,and reasonably shortening the hospital stay of the patients.

prostate cancerhospitalization expensesinfluencing factorsquantile regressionrandom forest

黄万宇

展开 >

武汉大学公共卫生学院,湖北武汉 430071

前列腺癌 住院费用 影响因素 分位数回归 随机森林

2024

中华男科学杂志
南京军区南京总医院

中华男科学杂志

CSTPCD
影响因子:1.052
ISSN:1009-3591
年,卷(期):2024.30(11)