Work

Trust in artificial intelligence for diagnosis of dermatologic lesions: a systematic review of qualitative studies

Year: 2026

Type: article

Abstract: Artificial intelligence (AI) has considerable potential to supports dermatological diagnosis, but its adoption depends on understanding the perspectives of stakeholders. This review aimed to synthesise qualitative evidence on the perceptions, barriers, and facilitators of AI among patients and healthcare professionals. We conducted a systematic search of major databases (MEDLINE, Web of Science, Scopus, CINAHL, and PsycINFO) from database inception through May 26, 2025. Data from ten included qualitative studies were analysed using thematic synthesis. The confidence in each finding was rigorously assessed using the GRADE-CERQual approach. The synthesis revealed three main themes: (1) the evolving role of AI across the dermatological diagnostic workflow; (2) significant barriers centred on clinical judgment, trust, and the physician-patient relationship; and (3) essential implementation requirements, such as usability and regulation. Confidence in these findings ranged from moderate to high. Successful AI integration requires transparent, validated, equitable, and usable tools. Stakeholders consistently emphasise that AI should be designed to complement, rather than replace, human clinical judgment and the physician-patient relationship to maintain a human-centred focus in medical practice. PROSPERO database (CRD420251010095)

Source: BMC Medical Informatics and Decision Making

DOI: 10.1186/s12911-026-03793-7