Information needs of psychiatric nurses and the potential of ChatGPT as an artificial intelligence-based learning tool: a qualitative study

Abstract

Purpose:

This study aims to explore the potential of Chat Generative Pre-trained Transformer (ChatGPT) in meeting the information needs of psychiatric nurses.

Method:

A two-stage qualitative phenomenological design was employed. In Stage 1, in-depth interviews were conducted with 18 psychiatric nurses to identify their theoretical knowledge requirements. In Stage 2, 18 questions, derived from the themes identified, were posed to ChatGPT; responses were then evaluated by 10 psychiatric nursing experts using the Lawshe content validity ratio method.

Results:

In Stage 1, thematic analysis identified Information Needs of Nurses as the main theme, encompassing six sub-themes: pharmacological practices; psychiatric care processes; psychiatric disorders; treatment and care approaches; roles and responsibilities, and medical terminology. In Stage 2, of the 18 ChatGPT responses assessed, 11 were judged ‘appropriate’ and 7 ‘partially appropriate’. ChatGPT showed strong performance in pharmacological knowledge and terminology, but less-satisfactory performance in emotionally engaged care processes.

Conclusions:

ChatGPT demonstrates potential as a supplementary educational resource for psychiatric nurses primarily in supporting factual and evidence-based learning domains, such as pharmacological knowledge and medical terminology. However, its use requires critical oversight and ethical awareness, as it cannot adequately address relational, contextual, or therapeutic aspects of psychiatric nursing care.