A Hybrid Natural Language Processing Platform for Multi-Site RWD Studies...20th World Congress on Medical and Health Informatics (MEDINFO), August 9-13, 2025, Taipei, Taiwan.
Real-world data (RWD) obtained from electronic medical records has become a valuable resource for healthcare research. However, integrating unstructured free-text clinical data remains a significant challenge. Although natural language processing (NLP) offers a promising solution, its implementation...
| Publicado en: | Studies in Health Technology & Informatics Vol. 329; pp. 1230 - 1235 |
|---|---|
| Autores principales: | , , , , , |
| Formato: | proceedings research tables/charts Journal Article |
| Publicado: |
Sage Publications Inc.
2025
|
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=187335053&site=ehost-live header: @attributes: shortDbName: ccm uiTerm: 187335053 longDbName: CINAHL Complete uiTag: AN controlInfo: bkinfo: dissinfo: jinfo: jid: 09269630 U1V jtl: Studies in Health Technology & Informatics issn: 09269630 maglogo: N pubinfo: dt: 2025 vid: 329 pid: 344 pub: Sage Publications Inc. place: Thousand Oaks, California artinfo: ui: 187335053 187335053 187335053 10.3233/SHTI251035 187335053 ppf: 1230 ppct: 5 formats: tig: atl: A Hybrid Natural Language Processing Platform for Multi-Site RWD Studies...20th World Congress on Medical and Health Informatics (MEDINFO), August 9-13, 2025, Taipei, Taiwan. aug: au: Kento SUGIMOTO Yasushi MATSUMURA Shoya WADA Shozo KONISHI Katsuki OKADA Toshihiro TAKEDA affil: Department of Medical Informatics, Osaka University Graduate School of Medicine, Osaka, Japan sug: subj: Natural Language Processing Electronic Health Records Data Security Information Retrieval Human Congresses and Conferences Taiwan Taiwan Deep Learning Workflow Funding Source ab: Real-world data (RWD) obtained from electronic medical records has become a valuable resource for healthcare research. However, integrating unstructured free-text clinical data remains a significant challenge. Although natural language processing (NLP) offers a promising solution, its implementation is frequently hampered by high computational costs. Moreover, privacy concerns complicate data integration in multi-site RWD studies. This study proposes a hybrid platform that integrates centralized NLP processing with robust privacy protection, facilitating effective information extraction from free-text data across various institutions. We performed comparative experiments utilizing 500 sample reports to assess the efficacy of the proposed hybrid platform against a fully distributed method using on-site servers. The results indicated that the central graphics processing units server significantly outperformed the site central processing units, processing reports in 0.12 s compared to an average of 64.23 s. Additionally, the central server exhibited a low and consistent increase in processing time regardless of report lengths, highlighting its efficiency and scalability. Our developed hybrid platform enhances computational efficiency while tackling privacy and data governance issues. pubtype: Academic Journal doctype: proceedings research tables/charts Journal Article ougenre: Article language: English refInfo: holdings: @attributes: islocal: N |
|---|