Clare Rainey, Raymond Bond, Jonathan McConnell, Avneet Gill, Ciara Hughes, Devinder Kumar, Sonyia McFadden
Abstract
Artificial intelligence decision support systems have been proposed to assist a struggling National Health Service (NHS) workforce in the United Kingdom. Its implementation in UK healthcare systems has been identified as a priority for deployment.
Introduction
The current backlog and delay in the reporting of radiographs has driven investigations into the adoption of new technologies that could increase efficiency and “free up clinicians” to spend more time with patients [1,2]. Artificial intelligence (AI) has been proposed as a solution in automating the diagnosis of pathology on radiographic images.
Methods
Ethical permission for this study has been granted by Ulster University Nursing and Health Research Ethics Filter Committee FCNUR-20–035. Online, informed consent was gathered form all participants prior to commencement of the study, by an initial slide presentation.
Results
All data analysis was conducted on SPSS® v 27 [27] and Microsoft® Excel® [28].
Limitations
There were a relatively small number of participants interpreting each examination, however this was intentional to encourage participation in an acceptable time frame, reducing the within-study attrition rate. There were 21 examinations included in this study.
Conclusion
Radiographers’ and student radiographers’ accuracy in diagnosis can be improved with the use of AI, even a poorly functioning system. Participants in this study tended to follow the diagnosis from the system, resulting in decreased accuracy in the diagnostic task in some cases.
Citation: Rainey C, Bond R, McConnell J, Gill A, Hughes C, Kumar D, et al. (2025) The impact of AI feedback on the accuracy of diagnosis, decision switching and trust in radiography. PLoS One 20(5): e0322051. https://doi.org/10.1371/journal.pone.0322051
Editor: Cristiano Miranda de Araujo, Tuiuti University of Parana: Universidade Tuiuti do Parana, BRAZIL
Received: July 24, 2024; Accepted: March 15, 2025; Published: May 9, 2025
Copyright: © 2025 Rainey et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The original Excel and SPSS datasets supporting this publication are openly available from Ulster University’s Research Portal at [https://doi.org/10.21251/50890091-4b54-4644-b980-ea9da646aa0e and https://doi.org/10.21251/9ec9de85-31c3-47c9-90d2-0ededc0a73c3]. The dataset of clinical images cannot be shared publicly because of ethical restrictions on sharing sensitive data. Access to the data can be provided following a successful application to Ulster University’s Nursing and Health Research Ethics Filter Committee. Ulster University’s Research Portal contains metadata on the dataset and instructions on how to request access to this dataset. This information can be accessed at [https://doi.org/10.21251/50890091-4b54-4644-b980-ea9da646aa0e and https://doi.org/10.21251/9ec9de85-31c3-47c9-90d2-0ededc0a73c3]. This paper is accompanied by representative samples of experimental data and the relevant numerical tabulated raw data is available from Ulster University’s Research Portal at 10.21251/794ab086-d855-4e44-805a-5c019868546b or by contacting pure-support@ulster.ac.uk.
Funding: This work has been funded by the College of Radiographers Research Industry Partnership Research awards scheme (CoRIPS) no. 183. The funder played no role in study design, data collection, analysis and interpretation of data, or the writing of this manuscript.
Competing interests: The authors have declared that no competing interests exist.
