A Decision-Theoretic Perspective on Fairness in Clinical Predictive Models

A Decision-Theoretic Perspective on Fairness in Clinical Predictive Models | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article A Decision-Theoretic Perspective on Fairness in Clinical Predictive Models Joshua W. Anderson, Nader Shaikh, Gregory F. Cooper, Shyam Visweswaran This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9644545/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Fairness is an important concern in statistical models, especially in clinical prediction models. Most fairness methods focus on model predictions, aiming for parity in model performance across relevant groups. However, this approach overlooks the broader implications of fairness when these models are used in clinical decision-making. We argue that prediction-based fairness frameworks, while valuable, are inherently limited when patient outcomes are equally, if not more, important concerning fairness. We analyze a deployed clinical prediction model, UTICalc, which was revised to improve fairness across racial groups and showed improved performance on a prediction-based fairness metric, namely, equal opportunity (equal true positive rate). We developed a decision-theoretic framework to assess the fairness of UTICalc by integrating patient outcome utilities with model predictions. To this end, we constructed a decision tree to model the clinical decision-making process for assessing and treating urinary tract infection (UTI) in young children, for which UTICalc was developed. Our results show that the revised UTICalc model did not improve an outcome-based fairness metric, namely, expected utility parity. This suggests that prediction-based and outcome-based fairness may diverge, with implications for clinical settings. Furthermore, we suggest that fairness in clinical prediction models should be evaluated based on patient outcomes as well as model predictions. Biological sciences/Computational biology and bioinformatics Health sciences/Diseases Physical sciences/Mathematics and computing Health sciences/Medical research Health sciences/Risk factors Full Text Additional Declarations No competing interests reported. Supplementary Files Supplementaryinformationsubmission.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9644545","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":636320949,"identity":"c86934d2-3af3-4bf3-8e52-81d7370c64d9","order_by":0,"name":"Joshua W. Anderson","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAtElEQVRIiWNgGAWjYBACxgYGBuMfFQw8YB4PsVqKGc4w8PCwEasFBD4ztgFVE62FeUbyw82F8+7I2Ms3MD5420aMw2akGRvP3PYM5DBmw7lEaZmdYGbAu+0wSAubNC9xWtK//+CdA9bC/ptILTkGxrwNEFuYidMy/02B4YxjQL8cS2yWnHOOCC2GPcc3GHyouWPP3nz44Ic3ZcRoaQBTBxggsUoMkGeAaxkFo2AUjIJRgAMAAOTGM6vXQBpiAAAAAElFTkSuQmCC","orcid":"","institution":"University of Pittsburgh","correspondingAuthor":true,"prefix":"","firstName":"Joshua","middleName":"W.","lastName":"Anderson","suffix":""},{"id":636320951,"identity":"94bc3b94-c48b-479a-b47f-bf0dd88cf80a","order_by":1,"name":"Nader Shaikh","email":"","orcid":"","institution":"Children's Hospital of Pittsburgh","correspondingAuthor":false,"prefix":"","firstName":"Nader","middleName":"","lastName":"Shaikh","suffix":""},{"id":636320953,"identity":"7e4aa6d2-ceb7-48da-a5e7-1a65fb4c85d2","order_by":2,"name":"Gregory F. Cooper","email":"","orcid":"","institution":"University of Pittsburgh","correspondingAuthor":false,"prefix":"","firstName":"Gregory","middleName":"F.","lastName":"Cooper","suffix":""},{"id":636320954,"identity":"50fb5d08-0fd7-41e8-a932-ba91ce92959b","order_by":3,"name":"Shyam Visweswaran","email":"","orcid":"","institution":"University of Pittsburgh","correspondingAuthor":false,"prefix":"","firstName":"Shyam","middleName":"","lastName":"Visweswaran","suffix":""}],"badges":[],"createdAt":"2026-05-07 15:24:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9644545/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9644545/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108977741,"identity":"853c5def-1baf-4e37-855a-290276615a82","added_by":"auto","created_at":"2026-05-11 11:32:46","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":518628,"visible":true,"origin":"","legend":"","description":"","filename":"Fairnessmanuscriptsumbission.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9644545/v1_covered_3dd3e995-52d4-4d4b-8199-f3a8ddfdb78b.pdf"},{"id":108785007,"identity":"8a43a5a9-53d1-464c-9ce5-a2936668eb84","added_by":"auto","created_at":"2026-05-08 11:06:58","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":112095,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementaryinformationsubmission.docx","url":"https://assets-eu.researchsquare.com/files/rs-9644545/v1/70e336718d911b9a3e4ab4a6.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Decision-Theoretic Perspective on Fairness in Clinical Predictive Models","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9644545/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9644545/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eFairness is an important concern in statistical models, especially in clinical prediction models. Most fairness methods focus on model predictions, aiming for parity in model performance across relevant groups. However, this approach overlooks the broader implications of fairness when these models are used in clinical decision-making. We argue that prediction-based fairness frameworks, while valuable, are inherently limited when patient outcomes are equally, if not more, important concerning fairness. We analyze a deployed clinical prediction model, UTICalc, which was revised to improve fairness across racial groups and showed improved performance on a prediction-based fairness metric, namely, equal opportunity (equal true positive rate). We developed a decision-theoretic framework to assess the fairness of UTICalc by integrating patient outcome utilities with model predictions. To this end, we constructed a decision tree to model the clinical decision-making process for assessing and treating urinary tract infection (UTI) in young children, for which UTICalc was developed. Our results show that the revised UTICalc model did not improve an outcome-based fairness metric, namely, expected utility parity. This suggests that prediction-based and outcome-based fairness may diverge, with implications for clinical settings. Furthermore, we suggest that fairness in clinical prediction models should be evaluated based on patient outcomes as well as model predictions.\u003c/p\u003e","manuscriptTitle":"A Decision-Theoretic Perspective on Fairness in Clinical Predictive Models","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-08 11:06:40","doi":"10.21203/rs.3.rs-9644545/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f49e06cb-d076-484b-bad0-5f363bb1d119","owner":[],"postedDate":"May 8th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Rejected","date":"2026-05-09T19:14:26+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-05-09T19:13:57+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-05-08T09:22:02+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Digital Medicine","date":"2026-05-07T15:08:58+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":67728636,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":67728637,"name":"Health sciences/Diseases"},{"id":67728638,"name":"Physical sciences/Mathematics and computing"},{"id":67728639,"name":"Health sciences/Medical research"},{"id":67728640,"name":"Health sciences/Risk factors"}],"tags":[],"updatedAt":"2026-05-09T19:24:18+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-08 11:06:40","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9644545","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9644545","identity":"rs-9644545","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}