EUVIMEDEUVIMED entdecken
CrossrefInfokarte

Bias in Medical AI in Relation to Patient Empowerment and Digital Equity: Scoping Review

Lisa-Maria Norz, Elizaveta Beeck, Elske Ammenwerth

Online Journal of Public Health Informatics · 2026-11-09 · 18 · e94439-e94439

DOI: 10.2196/94439 ↗
AbstractBibliografische Angaben

Abstract

Abstract Background AI is increasingly embedded in health care and digital health applications, shaping clinical processes as well as how patients access information and participate in care. Research on bias in medical AI has largely focused on technical and performance-related disparities. Less attention has been paid to how such biases may affect patients’ access, understanding, trust, and opportunities for meaningful involvement in care. Objective This scoping review aimed to map how studies of bias in medical AI describe patient-related consequences relevant to patient empowerment and the digital divide and to identify gaps across technical, social, structural, and patient-centered perspectives. Methods This scoping review followed JBI (Joanna Briggs Institute) methodology and PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) reporting guidance. PubMed/MEDLINE and EBSCOHost databases were searched in August 2025 and supplemented by Google Scholar searches. Peer-reviewed studies published from 2015 onward were included if they addressed bias in medical AI and described patient-related consequences relevant to patient empowerment or digital inequality. Data charting focused on AI application types, bias dimensions, patient empowerment domains, digital divide–related factors, intersectionality, and debiasing strategies. Findings were synthesized descriptively and thematically. Results The search identified 497 records, of which 22 studies met the inclusion criteria. Bias was represented as a multidimensional sociotechnical phenomenon, with social bias identified most frequently (21/22 studies), followed by algorithmic bias (16/22 studies), structural bias (14/22 studies), and design-related bias (12/22 studies). Patient empowerment was most frequently addressed through involvement in care (n=13), followed by health literacy (n=9) and trust and assurance (n=7), whereas broader psychosocial and everyday-life domains were addressed less frequently. Digital divide–related factors were identified in 19 of the 22 studies, with digital literacy and usability barriers reported most frequently. Seven studies considered 2 or more social or structural dimensions of disadvantage. Although many studies discussed approaches that could mitigate bias-related barriers, explicit empirical evaluation of debiasing strategies in relation to patient empowerment or digital equity was largely absent. Conclusions Patient-related consequences of bias in medical AI extend beyond algorithmic performance and are reflected in access, understanding, trust, and involvement in care. Broader empowerment dimensions, multidimensional disadvantage, and empirically evaluated mitigation strategies remain comparatively underexplored.

Wissen, das zu Ihrer Arbeit passt.

Vom kostenlosen Einstieg über die persönliche Recherche bis zur vollständigen Research Suite.

Tarife ansehen →

Sind alle Verlagsvolltexte enthalten?

Nein. Bibliografische Nachweise, Abstracts und Volltexte sind unterschiedliche Inhalte. Ob ein Volltext verfügbar ist, hängt von der Quelle, Open Access und Ihren vorhandenen Lizenzen ab. Ein EUVIMED-Tarif ersetzt keine zusätzlichen Verlagsrechte.

Tarife ansehen →