Evaluating the Impact of Workshop-Based Training for AI in Radiology: A Pre- and Postsurvey Analysis of Confidence, Understanding, and Perceptions

Abstract

Abstract Artificial intelligence (AI) fundamentally involves mathematical concepts related to handling and applying data to real-world scenarios. In the current educational scenario, the academic separation between biological sciences and data sciences often begins at an early educational stage (e.g., by 10th grade in many curricula), leading to a significant understanding gap between data scientists and medical professionals. In radiology, where large volumes of reproducible numerical data are routinely stored, AI holds significant transformative potential. To harness its benefits and understand its limitations, radiologists must acquire foundational knowledge of these techniques. This workshop was therefore designed to bridge this knowledge and conceptual divide, in addition to identifying deficiencies in basic concepts. This article evaluates the effectiveness of an AI in radiology workshop by comparing participants' knowledge perception, confidence, and overall perspectives before and after the survey. Prospective survey-based study using pre- and post-workshop questionnaires. The workshop was held on January 26, 2025 during the 77th annual Indian Radiological and Imaging Association (IRIA) and 23rd Asian Oceanian Congress of Radiology (AOCR) in Chennai, Tamil Nadu, India. Radiologists, fellowship trainees, and radiology residents attending the workshop. Forty-five participants completed both surveys; only 11.1% had prior formal AI training, and 20% had used AI in clinical practice. Perceived understanding improved significantly with 26.7% reporting “significant” and 44.4% “moderate” improvement. The mean postworkshop confidence in applying AI-assisted tools was 3.2 on a 5-point Likert scale, indicating a moderate level of confidence following the training. A positive shift in perception was observed in 82.2% of participants. Case studies and practical examples were most valued, alongside requests for more hands-on sessions and improved technical infrastructure. The “AI in Radiology” workshop successfully fostered a more confident and informed perspective on AI in radiology among its participants. The findings strongly advocate for continued educational efforts that are clinically relevant, practical, and highly interactive. Future initiatives should prioritize extended hands-on sessions and advanced case-based discussions to meet the expressed needs for continued, deeper learning.

Ähnliche Arbeiten