Predictive Model of Internal Bleeding in Elderly Aspirin Users Using XGBoost Machine Learning

IF 2 4区医学 Q2 HEALTH CARE SCIENCES & SERVICES Risk Management and Healthcare Policy Pub Date : 2024-09-18 DOI:10.2147/rmhp.s478826

Tenggao Chen, Wanlin Lei, Maofeng Wang

{"title":"Predictive Model of Internal Bleeding in Elderly Aspirin Users Using XGBoost Machine Learning","authors":"Tenggao Chen, Wanlin Lei, Maofeng Wang","doi":"10.2147/rmhp.s478826","DOIUrl":null,"url":null,"abstract":"Objective: This study aimed to develop a predictive model for assessing internal bleeding risk in elderly aspirin users using machine learning. Methods: A total of 26,030 elderly aspirin users (aged over 65) were retrospective included in the study. Data on patient demographics, clinical features, underlying diseases, medical history, and laboratory examinations were collected from Affiliated Dongyang Hospital of Wenzhou Medical University. Patients were randomly divided into two groups, with a 7:3 ratio, for model development and internal validation, respectively. Least absolute shrinkage and selection operator (LASSO) regression, extreme gradient boosting (XGBoost), and multivariate logistic regression were employed to develop prediction models. Model performance was evaluated using area under the curve (AUC), calibration curves, decision curve analysis (DCA), clinical impact curve (CIC), and net reduction curve (NRC). Results: The XGBoost model exhibited the highest AUC among all models. It consisted of six clinical variables: HGB, PLT, previous bleeding, gastric ulcer, cerebral infarction, and tumor. A visual nomogram was developed based on these six variables. In the training dataset, the model achieved an AUC of 0.842 (95% CI: 0.829– 0.855), while in the test dataset, it achieved an AUC of 0.820 (95% CI: 0.800– 0.840), demonstrating good discriminatory performance. The calibration curve analysis revealed that the nomogram model closely approximated the ideal curve. Additionally, the DCA curve, CIC, and NRC demonstrated favorable clinical net benefit for the nomogram model. Conclusion: This study successfully developed a predictive model to estimate the risk of bleeding in elderly aspirin users. This model can serve as a potential useful tool for clinicians to estimate the risk of bleeding in elderly aspirin users and make informed decisions regarding their treatment and management. Keywords: aspirin, bleeding, haemorrhage, predictive model, extreme gradient boosting, nomogram ","PeriodicalId":56009,"journal":{"name":"Risk Management and Healthcare Policy","volume":"3 1","pages":""},"PeriodicalIF":2.0000,"publicationDate":"2024-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Risk Management and Healthcare Policy","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.2147/rmhp.s478826","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"HEALTH CARE SCIENCES & SERVICES","Score":null,"Total":0}

引用次数: 0

Abstract

Objective: This study aimed to develop a predictive model for assessing internal bleeding risk in elderly aspirin users using machine learning.
Methods: A total of 26,030 elderly aspirin users (aged over 65) were retrospective included in the study. Data on patient demographics, clinical features, underlying diseases, medical history, and laboratory examinations were collected from Affiliated Dongyang Hospital of Wenzhou Medical University. Patients were randomly divided into two groups, with a 7:3 ratio, for model development and internal validation, respectively. Least absolute shrinkage and selection operator (LASSO) regression, extreme gradient boosting (XGBoost), and multivariate logistic regression were employed to develop prediction models. Model performance was evaluated using area under the curve (AUC), calibration curves, decision curve analysis (DCA), clinical impact curve (CIC), and net reduction curve (NRC).
Results: The XGBoost model exhibited the highest AUC among all models. It consisted of six clinical variables: HGB, PLT, previous bleeding, gastric ulcer, cerebral infarction, and tumor. A visual nomogram was developed based on these six variables. In the training dataset, the model achieved an AUC of 0.842 (95% CI: 0.829– 0.855), while in the test dataset, it achieved an AUC of 0.820 (95% CI: 0.800– 0.840), demonstrating good discriminatory performance. The calibration curve analysis revealed that the nomogram model closely approximated the ideal curve. Additionally, the DCA curve, CIC, and NRC demonstrated favorable clinical net benefit for the nomogram model.
Conclusion: This study successfully developed a predictive model to estimate the risk of bleeding in elderly aspirin users. This model can serve as a potential useful tool for clinicians to estimate the risk of bleeding in elderly aspirin users and make informed decisions regarding their treatment and management.

Keywords: aspirin, bleeding, haemorrhage, predictive model, extreme gradient boosting, nomogram

查看原文

微信好友朋友圈 QQ好友复制链接

本刊更多论文

利用 XGBoost 机器学习建立阿司匹林老年患者内出血预测模型

研究目的本研究旨在利用机器学习技术开发一种用于评估老年阿司匹林使用者体内出血风险的预测模型：研究回顾性纳入了 26,030 名老年阿司匹林使用者（65 岁以上）。患者的人口统计学、临床特征、基础疾病、病史和实验室检查数据均来自温州医科大学附属东阳医院。按 7:3 的比例将患者随机分为两组，分别用于模型开发和内部验证。采用最小绝对收缩和选择算子（LASSO）回归、极梯度提升（XGBoost）和多元逻辑回归建立预测模型。使用曲线下面积（AUC）、校准曲线、决策曲线分析（DCA）、临床影响曲线（CIC）和净减少曲线（NRC）对模型性能进行评估：在所有模型中，XGBoost 模型的 AUC 值最高。该模型由六个临床变量组成：HGB、PLT、既往出血、胃溃疡、脑梗塞和肿瘤。根据这六个变量绘制了直观的提名图。在训练数据集中，该模型的 AUC 为 0.842（95% CI：0.829- 0.855），而在测试数据集中，该模型的 AUC 为 0.820（95% CI：0.800- 0.840），显示出良好的判别性能。校准曲线分析表明，提名图模型非常接近理想曲线。此外，DCA 曲线、CIC 和 NRC 均显示提名图模型具有良好的临床净效益：本研究成功建立了一个预测模型，用于估计老年阿司匹林使用者的出血风险。该模型可作为临床医生估计老年阿司匹林使用者出血风险的潜在有用工具，并就其治疗和管理做出明智决策。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文去求助

来源期刊

Risk Management and Healthcare Policy Medicine-Public Health, Environmental and Occupational Health

CiteScore

6.20

自引率

2.90%

发文量

242

审稿时长

16 weeks

期刊介绍： Risk Management and Healthcare Policy is an international, peer-reviewed, open access journal focusing on all aspects of public health, policy and preventative measures to promote good health and improve morbidity and mortality in the population. Specific topics covered in the journal include: Public and community health Policy and law Preventative and predictive healthcare Risk and hazard management Epidemiology, detection and screening Lifestyle and diet modification Vaccination and disease transmission/modification programs Health and safety and occupational health Healthcare services provision Health literacy and education Advertising and promotion of health issues Health economic evaluations and resource management Risk Management and Healthcare Policy focuses on human interventional and observational research. The journal welcomes submitted papers covering original research, clinical and epidemiological studies, reviews and evaluations, guidelines, expert opinion and commentary, and extended reports. Case reports will only be considered if they make a valuable and original contribution to the literature. The journal does not accept study protocols, animal-based or cell line-based studies.