運用生成式人工智慧協作建構實證照護實務之PICO提問與檢索詞彙

Leveraging Generative Artificial Intelligence to Construct PICO Questions and Search Terms for Evidence-Based Practice

提出一個臨床問題是實證照護實務的首要步驟，包括PICO架構及發展關鍵字、醫學主題詞及同義字的檢索詞彙。然而，檢索詞彙稀少或未使用醫學主題詞，是實證護理新手的障礙之一。生成式人工智慧具有協作提問實證主題、優化PICO架構及擴展檢索詞彙的功能。本文先詳述手動操作流程，以建構實作者完備的核實能力，再使用ChatGPT、Gemini及NotebookLM說明協作建構PICO提問及檢索詞彙。最後，特別強調實作者應具備實證照護實務概念及專業領域知能，以確保對生成式人工智慧的產出結果進行嚴謹的核實與專業評判。

Asking a clinical question is the foundational step of evidence-based practice, encompassing the PICO framework and the development of search terms that include keywords, Medical Subject Headings (MeSH), and synonyms. However, limited search terms and the failure to utilize MeSH terms are among the major barriers for novices in evidence-based nursing. Generative AI possesses the capabilities to collaborate on formulating evidence-based topics, optimize PICO frameworks, and expand search terms. This article first details the manual workflow to establish the practitioner’s comprehensive verification competency, followed by utilizing ChatGPT, Gemini, and NotebookLM to collaboratively construct PICO questions and search terms. Finally, it is emphasized that practitioners should possess concepts of evidence-based practice and domain-specific expertise to rigorously verify and professionally judge generative AI outputs.