Generative AI-Assisted Construction of a PBL Case Library for Acute and Critical Care in the Emergency Department and Its Evaluation in Clinical Teaching

Abstract

<title>Abstract</title> Background：As medical education reform deepens, student-centered PBL is vital in clinical teaching. Yet, ED diseases’ acute, critical nature and traditional PBL’s inability to simulate dynamic changes, lack multidisciplinary scenarios, and misaligned difficulty grading limit its suitability for ED teaching. Objective：This study focuses on applying AI in ED PBL case generation and teaching. It aims to build an AI-enabled PBL system integrating "ED PBL cases – hierarchical question banks – intelligent assessment" via generative AI, to enhance students’ acute illness identification, emergency management, and multidisciplinary collaboration, addressing ED teaching pain points. Methods：A quasi-experimental design with mixed methods was used. 120 Zhenjiang hospital students were randomized into experimental (AI-enabled PBL) and control (traditional PBL) groups. Data included theoretical exams, Clinical Thinking Ability Scale scores, teacher case-design time, and student satisfaction. The Chinese Clinical Thinking Ability Assessment Scale and supervisor evaluations were used. Thematic analysis handled qualitative data; independent samples t-tests and ANOVA, quantitative data. Results：The AI-enabled model outperformed traditional PBL. Experimental group’s theoretical exam score (84.3 ± 5.2) was higher than control’s (76.8 ± 6.1, t = 6.21, P &lt; 0.001). Its CCTS score (78.5 ± 7.3) also exceeded control’s (69.2 ± 8.1, t = 5.87, P = 0.003), boosting clinical thinking. Teacher case-design time reduced by 63% (45 ± 12 vs. 122 ± 18 minutes, P &lt; 0.001). Qualitative data revealed positive student experience, teacher recognition of AI question bank difficulty (with optimization needs), and suggestions for improvement plus concerns about AI answer accuracy. Conclusion: The AI-enabled PBL model enhances students’ academic performance, clinical learning and reasoning abilities, and teachers’ instruction quality. It serves as a practical example for medical education digital transformation and is worth wide promotion.

Ähnliche Arbeiten