{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,4]],"date-time":"2026-08-04T14:49:26Z","timestamp":1785854966123,"version":"3.56.0"},"reference-count":37,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["JMIR Med Inform"],"abstract":"\n Background<\/jats:title>\n Prediction models based on machine learning (ML) methods are being increasingly developed and adopted in health care. However, these models may be prone to bias and considered unfair if they demonstrate variable performance in population subgroups. An unfair model is of particular concern in bladder cancer, where disparities have been identified in sex and racial subgroups.<\/jats:p>\n <\/jats:sec>\n \n Objective<\/jats:title>\n This study aims (1) to develop a ML model to predict survival after radical cystectomy for bladder cancer and evaluate for potential model bias in sex and racial subgroups; and (2) to compare algorithm unfairness mitigation techniques to improve model fairness.<\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n We trained and compared various ML classification algorithms to predict 5-year survival after radical cystectomy using the National Cancer Database. The primary model performance metric was the F1-score. The primary metric for model fairness was the equalized odds ratio (eOR). We compared 3 algorithm unfairness mitigation techniques to improve eOR.<\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n We identified 16,481 patients; 23.1% (n=3800) were female, and 91.5% (n=15,080) were \u201cWhite,\u201d 5% (n=832) were \u201cBlack,\u201d 2.3% (n=373) were \u201cHispanic,\u201d and 1.2% (n=196) were \u201cAsian.\u201d The 5-year mortality rate was 75% (n=12,290). The best naive model was extreme gradient boosting (XGBoost), which had an F1-score of 0.860 and eOR of 0.619. All unfairness mitigation techniques increased the eOR, with correlation remover showing the highest increase and resulting in a final eOR of 0.750. This mitigated model had F1-scores of 0.86, 0.904, and 0.824 in the full, Black male, and Asian female test sets, respectively.<\/jats:p>\n <\/jats:sec>\n \n Conclusions<\/jats:title>\n The ML model predicting survival after radical cystectomy exhibited bias across sex and racial subgroups. By using algorithm unfairness mitigation techniques, we improved algorithmic fairness as measured by the eOR. Our study highlights the role of not only evaluating for model bias but also actively mitigating such disparities to ensure equitable health care delivery. We also deployed the first web-based fair ML model for predicting survival after radical cystectomy.<\/jats:p>\n <\/jats:sec>","DOI":"10.2196\/63289","type":"journal-article","created":{"date-parts":[[2024,10,6]],"date-time":"2024-10-06T11:33:39Z","timestamp":1728214419000},"page":"e63289","source":"Crossref","is-referenced-by-count":3,"title":["Survival After Radical Cystectomy for Bladder Cancer: Development of a Fair Machine Learning Model"],"prefix":"10.2196","volume":"12","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5297-3611","authenticated-orcid":false,"given":"Samuel","family":"Carbunaru","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-5928-7133","authenticated-orcid":false,"given":"Yassamin","family":"Neshatvar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5317-6809","authenticated-orcid":false,"given":"Hyungrok","family":"Do","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-1707-954X","authenticated-orcid":false,"given":"Katie","family":"Murray","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2854-3376","authenticated-orcid":false,"given":"Rajesh","family":"Ranganath","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7502-3475","authenticated-orcid":false,"given":"Madhur","family":"Nayan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2024,12,13]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1016\/j.eururo.2021.08.001"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pmed.1002758"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1016\/j.urolonc.2021.08.007"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1002\/pros.24233"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1126\/science.aax2342"},{"key":"ref6","unstructured":"AngwinJLarsonJMattuSKirchnerLMachine bias: there's software used across the country to predict future criminals. And it\u2019s biased against blacksProPublica20162024-11-07https:\/\/www.propublica.org\/article\/machine-bias-risk-assessments -in-criminal-sentencing"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1145\/3457607"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-022-07939-1"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1016\/j.urolonc.2017.12.015"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.3390\/cancers15174372"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2021.692728"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0210976"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.3233\/BLC-200303"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1097\/COC.0000000000000471"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1016\/j.juro.2017.04.086"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2303.16626"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1245\/s10434-007-9747-3"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.7759\/cureus.41253"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1001\/jamanetworkopen.2022.26327"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1002\/cncr.23733"},{"key":"ref21","doi-asserted-by":"crossref","unstructured":"JeniLACohnJFde la TorreFFacing imbalanced data\u2014recommendations for the use of performance metrics20132013 Humaine Association Conference on Affective Computing and Intelligent Interaction (ACII 2013)September 2-5, 2013Geneva, Switzerland","DOI":"10.1109\/ACII.2013.47"},{"key":"ref22","unstructured":"HardtMPriceESrebroNEquality of opportunity in supervised learning2016Proceedings of the 30th International Conference on Neural Information Processing SystemsDecember 5-10, 2016Barcelona, Spain"},{"key":"ref23","unstructured":"LeeJBuYRajanDSattigeriPPandaRDasSFair selective classification via sufficiency2021Proceedings of the 38th International Conference on Machine Learning (ICML 2021)July 18-24, 2021Virtual event"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1093\/qje\/qjx032"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1707.00046"},{"key":"ref26","author":"Barocas, S","year":"2023","journal-title":"Fairness and Machine Learning: Limitations and Opportunities"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1002\/cam4.4917"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.2196\/36388"},{"key":"ref29","unstructured":"AgarwalABeygelzimerADud\u00edkMLangfordJWallachHA reductions approach to fair classification2018Proceedings of the 35th International Conference on Machine Learning (ICML 2018)July 10-15, 2018Stockholm, Sweden"},{"key":"ref30","unstructured":"AgarwalADud\u00edkMWuZFair regression: quantitative definitions and reduction-based algorithms2019Proceedings of the 36th International Conference on Machine Learning (ICML 2019)June 9-15, 2019Long Beach, CA"},{"key":"ref31","unstructured":"BirdSDud\u00edkMEdgarRHornBLutzRMilanVFairlearn: a toolkit for assessing and improving fairness in AIMicrosoft20202024-11-07https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2020\/05\/Fairlearn_WhitePaper-2020-09-22.pdf"},{"key":"ref32","unstructured":"AngwinJLarsonJBias in criminal risk scores is mathematically inevitable, researchers sayProPublica20162024-11-07https:\/\/www.propublica.org\/article\/bias-in-criminal-risk-scores-is-mathematically-inevitable-researchers-say"},{"key":"ref33","unstructured":"Corbett-DaviesSPiersonEFellerAGoelSA computer program used for bail and sentencing decisions was labeled biased against Blacks. It?s actually not that clearThe Washington Post2016-10-17https:\/\/www.washingtonpost.com\/news\/monkey-cage\/wp\/2016\/10\/17\/can-an-algorithm-be-racist-our-analysis-is-more-cautious-than-propublicas\/"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.3322\/caac.21492"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1002\/jmri.26777"},{"key":"ref36","first-page":"1","volume":"140","author":"Tang, Z","year":"2022","journal-title":"Proc Mach Learn Res"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1136\/bmjhci-2021-100459"}],"container-title":["JMIR Medical Informatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/medinform.jmir.org\/2024\/1\/e63289","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,4]],"date-time":"2026-08-04T13:56:00Z","timestamp":1785851760000},"score":1,"resource":{"primary":{"URL":"https:\/\/medinform.jmir.org\/2024\/1\/e63289"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,12,13]]},"references-count":37,"URL":"https:\/\/doi.org\/10.2196\/63289","relation":{"has-preprint":[{"id-type":"doi","id":"10.2196\/preprints.63289","asserted-by":"object"}]},"ISSN":["2291-9694"],"issn-type":[{"value":"2291-9694","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,12,13]]}}}