国家中毒数据系统®:基于语言模型的文献计量趋势与主题建模综述
本研究检索PubMed、Embase和Scopus,纳入2003年1月1日至2025年4月14日明确引用国家中毒数据系统®的344篇文章,按期刊主题分类并对摘要进行文本嵌入、降维和聚类以识别前10大主题。国家中毒数据系统®相关发表从2007年1篇稳步增长至2023年34篇峰值;主题建模识别出儿童中毒(n=187)、阿片类中毒(n=51)、用药错误(n=38)和青少年自伤(n=34)等主要主题。发表期刊以Clinical Toxicology(n=122)、Pediatrics(n=19)和Journal of Medical Toxicology(n=16)最常见,期刊类别包括医学毒理学(47.1%)、儿科学(14.5%)、药理学(10.8%)、急诊医学(9.0%)和公共卫生(7.3%)。作者认为改善国家中毒数据系统®基础设施和可及性可能支持进一步创新并扩大研究者队伍。
要点
- 国家中毒数据系统®相关文献从2007年1篇增至2023年34篇,共纳入344篇。
- 主题建模识别主要主题:儿童中毒n=187、阿片类中毒n=51、用药错误n=38、青少年自伤n=34。
- 发表期刊以Clinical Toxicology(n=122)、Pediatrics(n=19)、Journal of Medical Toxicology(n=16)最常见。
- 期刊类别包括医学毒理学47.1%、儿科学14.5%、药理学10.8%、急诊医学9.0%、公共卫生7.3%。
- 作者认为改善国家中毒数据系统®基础设施和可及性可能支持进一步创新并扩大研究者队伍。
适用人群与场景
- 2003年1月1日至2025年4月14日发表的、明确引用国家中毒数据系统®的同行评议生物医学文献。
研究设计与关键结果
- 研究类型为文献计量学综述,检索PubMed、Embase和Scopus,经去重和筛选后344篇文章符合纳入标准。
- 国家中毒数据系统®相关发表从2007年1篇稳步增长,2023年达峰值34篇。
- 主题建模识别主要主题:儿童中毒n=187、阿片类中毒n=51、用药错误n=38、青少年自伤n=34。
- 最常见发表期刊:Clinical Toxicology n=122、Pediatrics n=19、Journal of Medical Toxicology n=16。
- 期刊类别:医学毒理学47.1%、儿科学14.5%、药理学10.8%、急诊医学9.0%、公共卫生7.3%。
证据类型、等级与局限
- 证据类型:文献计量学综述。
- 局限:原文未报告局限性。
对床旁实践的改变
- 作者认为改善国家中毒数据系统®基础设施和可及性可能支持进一步创新并扩大研究者队伍。
提炼说明
- 本条依据论文摘要整理;摘要未涉及「诊断标准与关键阈值」「处置流程与药物剂量」「红旗征与不能漏的情况」「指南或共识出处」,故未列出这些栏目;全文请点「阅读原文」。
英文原文(摘要)· 点击展开
INTRODUCTION: The National Poison Data System® is a key resource for toxicology research and public health surveillance. Despite widespread use, no recent comprehensive analysis of National Poison Data System® research in peer-reviewed biomedical literature exists. This study evaluates bibliographic trends in publications using National Poison Data System® data, identifies dominant research themes, and characterizes the journals disseminating this work.
METHODS: We searched PubMed, Embase, and Scopus for articles published between January 1, 2003, and April 14, 2025, that explicitly referenced the National Poison Data System®. Abstracts were reviewed for relevance and duplicates, categorized by journal topic using SCImago Journal & Country Rank categories, and publications by year aggregated. Abstracts underwent text embedding, dimensionality reduction, and clustering to identify the top 10 topics. Topic labels were generated through consensus voting, and topics were identified within individual abstracts using gpt-4.1-nano.
RESULTS: After duplicate removal and screening, 344 articles met the inclusion criteria. National Poison Data System® publications increased steadily from one in 2007, peaking at 34 in 2023. Topic modeling identified major themes: pediatric poisoning (n = 187), opioid poisoning (n = 51), medication errors (n = 38), and adolescent self-harm (n = 34). The most common journals included Clinical Toxicology (n = 122), Pediatrics (n = 19), and Journal of Medical Toxicology (n = 16). Journal categories included medical toxicology (47.1%), pediatrics (14.5%), pharmacology (10.8%), emergency medicine (9.0%), and public health (7.3%).
DISCUSSION: This study highlights the growing use of National Poison Data System® data in research, especially in pediatrics and opioid-related topics. The breadth of journals in which the data are published reflects its multidisciplinary value. Improving the National Poison Data System® infrastructure and accessibility may support further innovation and broaden the investigator pool.
CONCLUSION: National Poison Data System® data support a growing and diverse body of research with implications for toxicology, emergency medicine, pediatrics, and public health.
参考来源
- Weigel B, McCrary Q, Zweerink K, et al. The National Poison data System®: a bibliographic review of publication trends and topic modeling using language models. Clin Toxicol (Phila). 2025. PMID: 41432768. doi:10.1080/15563650.2025.2586077
供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。