Description
An analysis of User feedback of different interfaces and designs for clinically relevant decision support when using AI: what do imaging professionals’ prefer?Short Title: User Friendly AI study
Abstract (250 words maximum):
Background and Aim
Radiographers and radiologists are currently faced with the professional adjustment that Artificial Intelligence (AI) assistive technologies introduces to clinical practice (RCR, 2018; SoR 2021). These technological advancements are reflected in the Health and Care Professions Council (HCPC) (2023) recent documentation for radiographers, which now requires registrants to understand the principles and apply AI in practice. As the professions navigate through this change, the importance of engaging clinicians in AI research exploring the human computer interaction is essential, particularly the explanation user interface (EUI) which is the method the system provides to demonstrate its decision focus to the user. Further research is required around user opinions into interface preference to ease the transition to acceptance of the technology in clinical practice and investigate the impact on perceived user accuracy to build on existing research in the area (Bernstein et al, 2023).
Methods
An online international questionnaire surveying opinions on common chest radiograph EUIs including, heat “salience” map, textual reports and area of interest boxes was undertaken. The recognised System Usability Scale (SUS) inspired questions to explore radiographer, radiologist and student opinion on perceived ease of use in context of the AI technology. Participants were recruited through purposive convenience sampling. A mix of statistical and thematic analysis allows examination of the quantitative and qualitative data gathered.
Results
The results will be used to understand the current opinion surrounding preferred user interface on Chest radiographs.
Discussion and Conclusion
The lack of research in the area indicates the need for an exploration into EUI preference of imaging professionals and students. In understanding optimal user preference and investigation into the factors contributing to reluctance to trust the AI, could prove essential in the next step towards responsible technology acceptance in radiology.
| Period | 10 Oct 2024 |
|---|---|
| Event title | European Society of Medical Imaging Informatics |
| Event type | Conference |
| Location | Pisa, ItalyShow on map |