Patients prefer ChatGPT to institutional websites for questions on radiation-based imaging exams: international mixed-methods study.
Authors
Affiliations (9)
Affiliations (9)
- Department of Diagnostic and Interventional Radiology, Lausanne University Hospital (CHUV), Lausanne, 1011, Switzerland.
- Faculty of Biology and Medicine (FBM), University of Lausanne (UNIL), Lausanne, 1011, Switzerland.
- Department of Diagnostic and Interventional Radiology, Lebanese Hospital Geitaoui-UMC, Achrafieh, Geitaoui, 1100, Lebanon.
- Institute of Radiation Physics (IRA), University of Lausanne (UNIL), Lausanne, 1011, Switzerland.
- CIBM Center for Biomedical Imaging, Lausanne, 1011, Switzerland.
- Department of Radiologic Medical Imaging Technology, School of Health Sciences (HESAV), University of Applied Sciences and Arts Western Switzerland (HES-SO), Delémont, 1011, Switzerland.
- Clinical Imaging Physics Group, Department of Radiology, Duke University Medical Center, Durham, NC 27710, United States.
- Carl E. Ravin Advanced Imaging Laboratories, Department of Radiology, Duke University Medical Center, Durham, NC 27710, United States.
- Center for Virtual Imaging Trials, Department of Radiology, Duke University Medical Center, Durham, NC 27710, United States.
Abstract
Radiology-risk communication affects multiple clinical specialties that use ionizing radiation, and many patients seek related information online. Prior expert evaluations found comparable performance between ChatGPT-generated and radiology-risk answers from official institutions, but patient perspectives have not been assessed. To assess patients' perceptions of ChatGPT versus human-generated radiology-risk information. From December 2024 to March 2025, patients at 3 hospitals in the United States, Switzerland, and Lebanon were randomly assigned to 1 of 5 common radiology-risk questions. Participants, blinded to source, provided subjective ratings of both ChatGPT‑3.5 and human-generated institutional responses on 7-point Likert scales for satisfaction (primary outcome), comprehensibility, trust, and reassurance. Quantitative comparisons were performed with Inverse Normalizing Transformation, and free-text comments were analyzed using thematic coding. A total of 328 patients participated (34% aged 18-39 years, 33% aged 40-59 years, 31% aged 60-79 years, 3% aged ≥80 years; 188 female). ChatGPT responses were rated significantly higher than human responses for satisfaction (0.70-point advantage; <i>P</i> < .001), comprehensibility (0.27 points; <i>P</i> < .01), trust (0.65 points; <i>P</i> < .001), and reassurance (0.51 points; <i>P</i> < .001). Findings converged with qualitative written comments (r = 0.91, <i>P</i> < .05), in which ChatGPT attracted 2.3× more positive comments while human-generated responses received 1.7× more negative comments. Unlike experts, patients preferred ChatGPT-generated responses to institutional materials for radiology risk questions. This divergence highlights the need for patient-centered communication and suggests that large language model-based styles, implemented with expert oversight, may improve the perceived clarity and trustworthiness of educational materials in medical specialties that use ionizing radiation.