Evaluating the Accuracy of AI-Generated Nursing Care Plans for Impaired Gas Exchange Hasby Pri Choiruna (a*) & Indah Ramadhan (a)
a) Department of Nursing, Universitas Jenderal Soedirman, Indonesia
*hasby.pri[at]unsoed.ac.id
Abstract
Background : Accurate and timely nursing care planning is fundamental to patient safety and effective outcomes, particularly for complex conditions like impaired gas exchange. The emergence of generative Artificial Intelligence (AI) presents a novel opportunity to support nurses in this critical task. However, the accuracy and clinical validity of care plans generated by these models remain unevaluated.
Purpose : The purpose of this study was to evaluate the accuracy of AI-generated nursing care plans for impaired gas exchange.
Methods : The design of this study was a comparative analysis. The study was conducted by giving prompts in the form of patient case scenarios to nine AI models to generate a comprehensive nursing care plans. The accuracy of each generated plan was independently scored using a rubric based on the Indonesian Nursing Standards. The rubric evaluated accuracy across domains: nursing diagnosis, main nursing diagnosis, main nursing outcome, and main nursing interventions.
Results : A total of 9 AI-generated care plans were evaluated. The results showed that there were variations in accuracy of AI-generated nursing care plans for impaired gas exchange. Three of nine AI care plans had accuracy scores less than 100.
Conclusion : Generative AI models demonstrate promising potential in formulating foundational nursing care plans for impaired gas exchange, but their accuracy varies significantly between models. These technologies can serve as valuable assistive tools but are not yet a substitute for the critical appraisal and clinical judgment of professional nurses. Further research is imperative to refine AI models for clinical reliability and to develop guidelines for their ethical and effective integration into nursing practice.
Keywords: Artificial Intelligence, Nursing Care Plan, Impaired Gas Exchange, Clinical Decision Support, Nursing Informatics.
