Comparing ChatGPT and clinical nurses’ performances on tracheostomy care: A cross-sectional study

Comparing ChatGPT and clinical nurses’ performances on tracheostomy care: A cross-sectional study

2024 | Tongyao Wang, Juan Mu, Jialing Chen, Chia-Chin Lin
This study compares the performance of ChatGPT-3.5 and ChatGPT-4.0 with clinical nurses in answering questions about tracheostomy care. The research aims to evaluate the accuracy of ChatGPT responses and determine if different prompts affect their performance. A cross-sectional study was conducted, collecting data from 272 clinical nurses in mainland China and using ChatGPT-3.5 and 4.0. The study used a multiple-choice format with 43 questions related to tracheostomy care. Responses were graded on a 3-point scale (correct, partially correct, incorrect) by two otorhinolaryngology nurses. Results showed that ChatGPT-4.0 had significantly higher accuracy (64.3%) compared to ChatGPT-3.5 (60.5%) and clinical nurses (36.7%). ChatGPT-4.0 consistently outperformed both in all domains except "airway humidification." Prompt engineering did not significantly impact ChatGPT's performance. The study concludes that ChatGPT can serve as a valuable tool for improving knowledge in tracheostomy care, complementing patient and physician education.This study compares the performance of ChatGPT-3.5 and ChatGPT-4.0 with clinical nurses in answering questions about tracheostomy care. The research aims to evaluate the accuracy of ChatGPT responses and determine if different prompts affect their performance. A cross-sectional study was conducted, collecting data from 272 clinical nurses in mainland China and using ChatGPT-3.5 and 4.0. The study used a multiple-choice format with 43 questions related to tracheostomy care. Responses were graded on a 3-point scale (correct, partially correct, incorrect) by two otorhinolaryngology nurses. Results showed that ChatGPT-4.0 had significantly higher accuracy (64.3%) compared to ChatGPT-3.5 (60.5%) and clinical nurses (36.7%). ChatGPT-4.0 consistently outperformed both in all domains except "airway humidification." Prompt engineering did not significantly impact ChatGPT's performance. The study concludes that ChatGPT can serve as a valuable tool for improving knowledge in tracheostomy care, complementing patient and physician education.
Reach us at info@study.space