A rapid review on current and potential uses of large language models in nursing

Hobensack et al. (2024) focused on bibliometric trends rather than synthesising implementation strategies.

“Hobensack et al. (2024) focused on bibliometric trends rather than synthesising implementation strategies.”

Most existing reviews of GenAI in nursing education lack structured analytical frameworks for systematically mapping evidence into actionable insights.

“A few reviews have begun to examine GenAI in nursing education, but most lack structured analytical frameworks for systematically mapping evidence into actionable insights.”

The 18 included studies collectively explored diverse applications of AI in nursing across research, practice, and education.

“The 18 included studies in this umbrella review collectively explored the diverse applications of AI in nursing across research, practice, and education [32-49].”

This finding aligns with a previous review of ChatGPT in pharmacy practice, which highlighted a lack of standardized evaluation methods, and with broader healthcare reviews calling for standardized LLM evaluation approaches.

“This finding aligns with a previous review of ChatGPT in pharmacy practice, which highlighted a lack of standardized evaluation methods, 20 and with broader healthcare reviews calling for standardized LLM evaluation approaches.14,15”

A coherent synthesis of emerging evidence on GenAI in nursing education is lacking.

“However, a coherent synthesis of this emerging evidence is lacking.”

Studies report technical issues, training data biases, and AI system opacity.

“Studies reporting technical issues, training data biases, and AI system opacity (Hobensack et al., 2024) underscore the need for empirical validation to support evidence-based implementation.”

Randomised controlled trials represented approximately 2.5% of the corpus (n = 12).

“Randomised controlled trials represented approximately 2.5% of the corpus (n = 12), confirming at scale the observation by Hobensack et al. (2024) that the field is empirically thin.”

Hobensack et al. (2024) identified that most literature on LLMs in nursing is descriptive, based on commentaries and editorials, with limited empirical evidence to inform practice, education, and research.

“Similarly, Hobensack et al. (2024) identified that most literature on LLMs in nursing is descriptive, based on commentaries and editorials, with limited empirical evidence to inform practice, education, and research.”

In this rapid review, most included studies focused on nursing education.

“Our review also indicated that most studies focused on nursing education, consistent with Hobensack et al. (2024) and Yalcinkaya and Cinar Yucel's (2024) studies.”

The earliest systematic effort, a rapid review by Hobensack et al. (2024) synthesising 30 studies, found that 57% of included articles were editorials or letters to the editor and only seven reported original research findings.

“The earliest systematic effort, a rapid review by Hobensack et al. (2024) synthesising 30 studies across nursing education, clinical practice, and J o u r n a l P r e -p r o o f research, delivered a critical methodological assessment: 57% of included articles were editorials or letters to the editor, and only seven reported original research findings.”

The field is empirically thin, as confirmed at scale by the low proportion of RCTs.

“confirming at scale the observation by Hobensack et al. (2024) that the field is empirically thin.”

The field is dominated by opinion and editorial commentary, as confirmed and quantified at scale by the Evidence Gap Map.

“Hobensack et al.'s (2024) observation that the field is dominated by opinion and editorial commentary is confirmed and quantified at scale by the Evidence Gap Map.”