Understanding what matters most to patients
A new systematic review by PhD student Preksha Machaiya Kuppanda, and supervisors Prof Monika Janda, and Prof Liam Caffery from The University of Queensland has identified the key factors that influence how patients perceive and accept artificial intelligence (AI) used to analyse medical images for screening and diagnosis.
Published in the Interactive Journal of Medical Research, the review examined evidence from 59 studies across a range of medical imaging specialties to better understand what shapes patient acceptance of AI. The findings highlight that acceptance is influenced by multiple interacting factors rather than any single issue.
Human oversight remains the highest priority
The review found that the strongest and most consistently reported factor influencing patient acceptance was the role of clinicians in the diagnostic process. Across 48 studies, patients preferred AI to be used as a tool that supports healthcare professionals rather than replacing them as an independent decision-maker.
Other important factors included the accuracy and performance of AI systems, trust, transparency around AI use, clear accountability for AI-assisted decisions, and ethical considerations such as privacy and fairness. The review also found that individual characteristics, including demographics and previous healthcare experiences, can influence how patients perceive these factors.
Supporting patient-centered implementation of AI
Using evidence from the 59 included studies, the researchers developed a conceptual model illustrating how the different factors interact to shape patient acceptance of AI in medical image analysis.
Guiding future research in melanoma screening
The review provides a foundation for future research exploring patient acceptance of AI. The next stage of this work will involve validating the identified factors directly with patients and applying the findings to AI-supported skin cancer and melanoma screening.
Ultimately, the research aims to develop a patient-centered framework for assessing AI acceptability in skin cancer and melanoma screening. Such a framework could help guide the implementation of AI technologies in ways that align with patient expectations and support their successful integration into clinical practice. Additionally, the framework could form a baseline in the development of an instrument to capture and measure patients’ AI acceptance at an early implementation stage.
Link to publication: https://doi.org/10.2196/92969
