Objective: Our article demonstrates the effectiveness of using a validated framework to create a ChatGPT prompt that generates valid nursing care plan suggestions for one hypothetical older patient with lung cancer.
Method: This study describes the methodology for creating ChatGPT prompts that generate consistent care plan suggestions and its application for a lung cancer case scenario. After entering a nursing assessment of the patient's condition into ChatGPT, we asked it to generate care plan suggestions. Subsequently, we assessed the quality of the care plans produced by ChatGPT.
Results: While not all the suggested care plan terms (11 out of 16) utilized standardized nursing terminology, the ChatGPT-generated care plan closely matched the gold standard in scope and nature, correctly prioritizing oxygenation and ventilation needs.
Conclusion: Using a validated framework prompt to generate nursing care plan suggestions with ChatGPT demonstrates its potential value as a decision support tool for optimizing cancer care documentation.
Keywords: artificial intelligence; cancer; large language model; nursing; standardized nursing terminologies.
© The Author(s) 2024. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved. For permissions, please email: journals.permissions@oup.com.