CREAITED AI timing and eye tracking in chest Xray interpretation v1.0
Research type
Research Study
Full title
A within-subject eye-tracking study examining how the timing of AI decision support influences visual search behaviour, diagnostic accuracy, and trust during chest X-ray interpretation
IRAS ID
347920
Contact name
Elizabeth Davies
Contact email
Sponsor organisation
University Hospitals of Leicester NHS Trust
Duration of Study in the UK
0 years, 7 months, 31 days
Research summary
Chest X-rays (CXRs) are one of the most widely used imaging tests in the NHS and help identify infections, nodules and other chest abnormalities. Artificial intelligence (AI) tools are increasingly used to support clinicians when interpreting CXRs, but there is limited evidence about how the timing of AI information affects interpretation. This study, CREAITED (Chest Radiograph Evaluation with AI Timing, Eye-tracking, and Diagnostic decision making), investigates whether showing AI outputs before versus after a clinician first reviews a CXR changes visual attention, diagnostic accuracy and confidence.
The study will take place in NHS research settings and will involve healthcare professionals who interpret or act on CXR findings. Participants will complete a computer-based task interpreting CXRs while wearing eye-tracking equipment. Eye-tracking records where participants look on the image and how long they spend viewing different areas, providing objective information about visual search behaviour. Each participant will complete two sessions, one in each timing condition, with the order counterbalanced across participants.
Outcomes include visual attention, diagnostic accuracy, time spent reviewing images and confidence. Participants will also complete a short questionnaire about their experience using AI support. The main eye-tracking reader study does not involve patients and does not affect clinical care, using de-identified CXR images from existing examinations. Alongside this, a separate online supplementary survey will capture clinician, patient and public views on AI use in CXR interpretation. This collects no identifiable information and does not influence the eye-tracking task. Risks are minimal and mainly relate to time commitment and possible minor discomfort or fatigue from wearing eye-tracking equipment.
This research forms part of the NHS England Higher Specialist Scientific Training (HSST) doctorate programme and is not commercially funded. Findings will help inform safer and more effective use of AI in CXR interpretation to support evidence-based clinical practice and patient care.
REC name
East of England - Essex Research Ethics Committee
REC reference
26/EE/0111
Date of REC Opinion
30 Apr 2026
REC opinion
Further Information Favourable Opinion