[Analysis of training effectiveness and feedback in a pollen monitoring training program integrating artificial...

This study aimed to evaluate the effectiveness of the pollen training program at Peking Union Medical College Hospital (PUMCH) in improving participants' knowledge of pollen monitoring, identification, and artificial intelligence (AI)-based identification, as well as to assess the participants' satisfaction with the course design and their needs, in order to provide reference for future course optimization. A pre-post self-controlled design and a post-training cross-sectional survey were used. A total of 120 participants who attended the pollen training program at the Department of Allergy, Peking Union Medical College Hospital, between January 2024 and August 2025 were included. Electronic questionnaires administered before and after training were used to assess participants' knowledge acquisition, while a post-training questionnaire was used to collect course feedback. The exact McNemar test and Mann-Whitney U test were used for statistical analysis. As a result, after systematic training in the pollen course, the trainees showed significant improvement in their accuracy of answering questions related to pollen identification and monitoring (exact McNemar test, All P<0.05). Particularly marked improvements were observed in the identification of Humulus pollen morphology, identification of Ginkgo pollen morphology, and pollen-slide staining methods, with correct response rates increasing from 11.3% (8/71), 14.1% (10/71), and 12.7% (9/71) to 80.3% (57/71), 95.8% (68/71), and 77.5% (55/71), respectively. In the assessment of AI-based pollen identification, participants showed the greatest improvement in their understanding of image-recognition methods, with the correct response rate increasing from 36.1% (13/36) to 80.6% (29/36) (exact McNemar test, P<0.001). Participants reported a high level of perceived mastery of pollen identification and monitoring-related knowledge and skills. The self-assessment scores [M(Q₁,Q₃)] for all course components were ≥79.0, with a score of 80.0 (67.0, 92.5) for mastery of airborne allergenic pollen identification. Satisfaction scores for the overall program, theoretical sessions, and practical sessions [M (Q₁, Q₃)] were all 10 (10, 10). The duration of the theoretical sessions, practical sessions, and question-and-answer sessions was considered appropriate by 83.6% (61/73), 76.7% (56/73), and 78.1% (57/73) of participants, respectively.In conclusion, the pollen training program at Peking Union Medical College Hospital, which incorporated AI-based pollen identification, significantly enhanced allergy-related professionals' knowledge in pollen monitoring, identification, and AI-based pollen recognition, providing a reference for the optimization of related training programs.