Presentation Information
[2J1-GS-10a-02]Proposal of an AI-driven Platform for Automated Systematic Review in Clinical Practice
〇Kenichi Inoue1 (1. Breast Cancer Center, Minamiyamato Clinic)
Keywords:
systematic review
[Background] Modern medicine relies on Evidence-Based Medicine (EBM). While systematic reviews (SRs) inform clinical guidelines, the rapid expansion of medical knowledge necessitates faster evidence synthesis. We developed a system using generative AI to automate SR creation.
[Methods] Using the PubMed API, the system automatically retrieves abstracts. Upon receiving a clinical question (CQ), the AI generates search queries, selects relevant literature, identifies outcomes, and synthesizes a comprehensive SR. A web-based platform was established to share these results via a public database.
[Results] The system automates SR generation, achieving a volume and update frequency impossible with traditional printed media. It easily adapts to new information, such as the emergence of new drugs, through real-time updates.
[Discussion] Generative AI facilitates precise evidence extraction, realizing "dynamic guidelines" that evolve in real-time to keep pace with medical advancements.
[Methods] Using the PubMed API, the system automatically retrieves abstracts. Upon receiving a clinical question (CQ), the AI generates search queries, selects relevant literature, identifies outcomes, and synthesizes a comprehensive SR. A web-based platform was established to share these results via a public database.
[Results] The system automates SR generation, achieving a volume and update frequency impossible with traditional printed media. It easily adapts to new information, such as the emergence of new drugs, through real-time updates.
[Discussion] Generative AI facilitates precise evidence extraction, realizing "dynamic guidelines" that evolve in real-time to keep pace with medical advancements.