Development and Evaluation of a Large Language Model - Based Training Program for Nurses in Public Health Emergencies

Summary

The goal of this randomized controlled trial is to evaluate the immediate efficacy of a Large Language Model (LLM)-assisted training program in enhancing nurses' emergency response capabilities in 204 practicing nurses with ≤5 years of experience from tertiary hospitals in Guiyang, China, focusing on public health emergencies (PHEs). The main questions it aims to answer are:

1. Does LLM-assisted training improve nurses' comprehensive emergency response capabilities in PHEs?
2. Does it specifically enhance rescue skills and occupational protection abilities? Researchers will compare the experimental group (receiving routine PHE training + LLM-assisted learning) to the control group (receiving routine PHE training only) to see if LLM supplementation leads to significantly greater improvements in targeted emergency competencies.

Participants will:

Complete pre- and post-training assessments (Nurse Self-Assessment Scale for Emergency Response Ability, Nurse's Emergency Response Capacity Scale for PHEs).

Undergo a one-month PHE training program. (Experimental Group Only): Use LLMs for knowledge review, question answering, and exploring unfamiliar concepts during the training period.

Conditions

• The Emergency Response Capabilities of Nurses (Including Occupational Protection, Critical Thinking, Communication Skills and Humanistic Care, Etc.)

Interventions

OTHER

LLM-Assisted Public Health Emergency Training Program

A hybrid training program integrating the hospital's standard public health emergency (PHE) curriculum with Large Language Model (LLM) technology as an auxiliary learning tool. Participants receive: * Standardized PHE training (online lectures + offline simulations) covering professional knowledge, skills, and emergency drills (e.g., infectious disease response, trauma management). * LLM-enabled interactive support: Structured guidance to use LLMs for: Reviewing session content Resolving knowledge uncertainties via Exploring unfamiliar PHE concepts • Duration: 1 month, with 20-minute sessions. Distinguishing feature: Uses LLMs to dynamically adapt to individual learning needs, enabling on-demand knowledge reinforcement and overcoming spatiotemporal limitations of traditional training.

OTHER

Standard Public Health Emergency Training Program

The hospital's existing public health emergency (PHE) training program without AI augmentation. Participants receive: * Identical core content as the experimental group: Professional knowledge, skills training, and emergency drills for PHE response (e.g., disaster protocols, infection control). * Explicit restriction: Prohibited from using LLMs or any AI tools for learning support. * Delivery: Hybrid format (online + offline), 1-month duration, 20-minute sessions. Distinguishing feature: Represents traditional training methods reliant on instructor-led content without personalized, on-demand AI-driven reinforcement.

Sponsors & Collaborators

• The Affiliated Hospital Of Guizhou Medical University

  lead OTHER

Study Design

Allocation

RANDOMIZED

Purpose

SUPPORTIVE_CARE

Masking

DOUBLE

Model

PARALLEL

Eligibility

Min Age

18 Months

Max Age

35 Months

Sex

ALL

Healthy Volunteers

Yes

Timeline & Regulatory

Start

2024-10-01

Primary Completion

2024-12-31

Completion

2024-12-31

Countries

• China

Study Locations