Evaluating How Timing of AI Support Affects Chest X-ray Interpretation by Healthcare Professionals Using Eye-Tracking

What this Trial Is About

Researchers are evaluating how the timing of artificial intelligence AI support affects chest X-ray interpretation by healthcare professionals. This study explores whether showing AI information before or after a clinician first reviews a chest X-ray changes their visual search behavior, interpretation time, decision-making, confidence, and trust in AI. The study uses a behavioral diagnostic reader design involving eye-tracking technology to better understand these effects in a controlled NHS research setting. Participants will complete two chest X-ray interpretation sessions spaced at least four weeks apart, using the same set of de-identified images. In one session, AI output is shown before the clinician reviews the X-ray, while in the other session, the AI information is shown after the initial review. Each session involves a two-phase interpretation process where either the X-ray or AI output is first viewed, followed by the introduction of the other source before final interpretation submission. Eye-tracking equipment records visual attention and search patterns during the tasks. Participants will review 25 chest X-ray cases per session and complete questionnaires about their experience with AI support. Researchers will analyze interpretation time, diagnostic accuracy, clinical recommendations, confidence, trust, and visual search behavior. Data collected includes eye-tracking metrics, structured interpretation responses, and participant feedback. The study does not involve patients directly, change clinical care, or use AI for actual diagnosis. It aims to provide evidence to guide safer and more effective use of AI in chest X-ray interpretation.

AI Timing in Chest X-ray Interpretation Using Eye-Tracking