肺部病理期刊俱乐部二十四年经验:我们学到了什么?
该研究评估了一个每月举行的肺部病理期刊俱乐部24年来是否达成帮助成员跟进诊断性肺病理相关文献的目标。结果显示讨论文章的平均引用次数显著高于仅记录文章,成员对俱乐部结构和管理总体满意,并提出了减少扫描期刊数量、简化摘要细节及使用生成式人工智能辅助摘要生成等改进机会。
要点
- 讨论文章的平均引用次数显著高于仅记录文章(103.0 vs 64.9,P < .001)。
- 成员最看重文章摘要总结,其次是现场讨论。
- 生成式人工智能可减少准备时间,但人工编辑摘要提供更细致的分析。
核心结论
- 期刊俱乐部达成了最初帮助成员跟进文献的目标。
- 已识别出包括使用生成式人工智能辅助文章总结在内的改进机会。
研究设计与样本
- 对现任教师进行问卷调查以评估俱乐部是否满足期望并识别改进机会。
主要发现
- 讨论文章的平均(标准差)引用次数为103.0(409.80),仅记录文章为64.9(259.77)(P < .001)。
- 引用次数明显右偏,高引用文章抬高了均值。
- 成员对期刊俱乐部的结构和管理大多满意。
- 成员最看重文章摘要总结,其次是现场讨论。
关键数值与推荐
- 改进机会包括减少扫描期刊数量、简化摘要细节、使用生成式人工智能促进摘要生成。
- 生成式人工智能试点 anecdotal 减少准备时间,但人工编辑摘要包含更具体的背景、批判性评论和增强的要点信息。
临床怎么用
- 可考虑使用生成式人工智能辅助文章总结以减少准备时间,但需人工编辑以确保分析深度。
- 根据成员反馈优化期刊俱乐部结构,如减少扫描期刊数量和简化摘要细节。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
CONTEXT.—: A monthly pathology journal club has met for 24 years. It was established to help members stay apprised of the literature relevant to diagnostic pulmonary pathology. OBJECTIVE.—: To assess whether the journal club met its goal and to report on opportunities identified for improvement. DESIGN.—: To determine whether articles chosen for discussion as opposed to notation were more significant, Scopus citation indices for article types reviewed from January 2007 to November 2023 were compared. A survey of current faculty was undertaken to determine if the club was meeting expectations and to identify improvement opportunities. RESULTS.—: Articles from January 2007 to November 2023 included 858 discussed and 3385 noted. Mean (SD) citation count was 103.0 (409.80) for discussion and 64.9 (259.77) for notation articles (P < .001). The citation count was noticeably right skewed, as articles with high citation counts inflated the mean. Members were mostly satisfied with the way the journal club was structured and managed. Members most valued the summary of the articles, followed by the live discussion. Opportunities for improvement included decreasing the number of journals scanned, decreasing detail in summaries, and using generative artificial intelligence (AI) to facilitate summary generation. A pilot using AI anecdotally reduced preparatory time, but the human-edited summary included more specific context, critical commentary, and enhanced take-home messages, providing a more nuanced analysis. CONCLUSIONS.—: The journal club met its initial goal. Opportunities for improvement have been identified including the use of generative AI to facilitate article summarization.
引用
- Tazelaar HD, Aubry MC, Roden AC, et al. Twenty-Four Years' Experience With a Pulmonary Pathology Journal Club: What Have We Learned? Arch Pathol Lab Med. 2025. PMID: 39943856. doi:10.5858/arpa.2024-0331-OA
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。