Risk of bias assessment of post-stroke mortality machine learning predictive models: Systematic review

IF 2 4区医学 Q3 NEUROSCIENCES

Journal of Stroke & Cerebrovascular Diseases Pub Date : 2025-03-13 DOI:10.1016/j.jstrokecerebrovasdis.2025.108291

Nicole Maria Radley , Ian Soh , Abdelrahman M. Saad , Milindu Wickramarachchi , Amelia Dawson , Jeremy Ng Chieng Hin , Asad Ali , Abhrajit Giri , Alicia Kwan , Osama Elzankaly , Mariam Tarek Desouki , Mohamed S Jabal , Abdelrahman M Hamouda , Sherief Gozy , David F Kallmes

{"title":"Risk of bias assessment of post-stroke mortality machine learning predictive models: Systematic review","authors":"Nicole Maria Radley , Ian Soh , Abdelrahman M. Saad , Milindu Wickramarachchi , Amelia Dawson , Jeremy Ng Chieng Hin , Asad Ali , Abhrajit Giri , Alicia Kwan , Osama Elzankaly , Mariam Tarek Desouki , Mohamed S Jabal , Abdelrahman M Hamouda , Sherief Gozy , David F Kallmes","doi":"10.1016/j.jstrokecerebrovasdis.2025.108291","DOIUrl":null,"url":null,"abstract":"<div><h3>Background</h3><div>Stroke is a major cause of mortality and permanent disability worldwide. Precise prediction of post-stroke mortality is essential for guiding treatment decisions and rehabilitation planning. The ability of Machine learning models to process large amounts of data, offer a promising alternative for improving mortality prediction in stroke patients. In this review, we aim to evaluate the risk of bias in different machine learning models used for predicting post-stroke mortality.</div></div><div><h3>Methods</h3><div>This review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS). Relevant articles were retrieved from Cochrane Library, Scopus, PubMed, and Web of Science databases.</div></div><div><h3>Results</h3><div>A total of 9 studies were included, with an aggregate patient population of 669,424. Six studies used publicly available datasets, and four used hospital data with a follow up duration ranging from 7 days to 18 months. The range of area under the curve (AUC) for mortality prediction across the studies ranged from 0.81 to 0.95. All studies were determined to have a high overall risk of bias.</div></div><div><h3>Conclusion</h3><div>Machine learning models demonstrated great potential in predicting post-stroke mortality. However, implementation of these models in clinical practice is limited by high risk of bias. Future studies should focus on reducing this bias and enhancing the applicability of these models to improve the reliability of stroke mortality predictions.</div></div>","PeriodicalId":54368,"journal":{"name":"Journal of Stroke & Cerebrovascular Diseases","volume":"34 6","pages":"Article 108291"},"PeriodicalIF":2.0000,"publicationDate":"2025-03-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Stroke & Cerebrovascular Diseases","FirstCategoryId":"3","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1052305725000709","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"NEUROSCIENCES","Score":null,"Total":0}

引用次数: 0

Abstract

Background

Stroke is a major cause of mortality and permanent disability worldwide. Precise prediction of post-stroke mortality is essential for guiding treatment decisions and rehabilitation planning. The ability of Machine learning models to process large amounts of data, offer a promising alternative for improving mortality prediction in stroke patients. In this review, we aim to evaluate the risk of bias in different machine learning models used for predicting post-stroke mortality.

Methods

This review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS). Relevant articles were retrieved from Cochrane Library, Scopus, PubMed, and Web of Science databases.

Results

A total of 9 studies were included, with an aggregate patient population of 669,424. Six studies used publicly available datasets, and four used hospital data with a follow up duration ranging from 7 days to 18 months. The range of area under the curve (AUC) for mortality prediction across the studies ranged from 0.81 to 0.95. All studies were determined to have a high overall risk of bias.

Conclusion

Machine learning models demonstrated great potential in predicting post-stroke mortality. However, implementation of these models in clinical practice is limited by high risk of bias. Future studies should focus on reducing this bias and enhancing the applicability of these models to improve the reliability of stroke mortality predictions.

查看原文本刊更多论文

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of Stroke & Cerebrovascular Diseases Medicine-Surgery

CiteScore

5.00

自引率

4.00%

发文量

583

审稿时长

62 days

期刊介绍： The Journal of Stroke & Cerebrovascular Diseases publishes original papers on basic and clinical science related to the fields of stroke and cerebrovascular diseases. The Journal also features review articles, controversies, methods and technical notes, selected case reports and other original articles of special nature. Its editorial mission is to focus on prevention and repair of cerebrovascular disease. Clinical papers emphasize medical and surgical aspects of stroke, clinical trials and design, epidemiology, stroke care delivery systems and outcomes, imaging sciences and rehabilitation of stroke. The Journal will be of special interest to specialists involved in caring for patients with cerebrovascular disease, including neurologists, neurosurgeons and cardiologists.