Development of a Nursing Diagnosis/Record Generative AI System Based on Virtual Patient Data.
Journal:
Studies in health technology and informatics
PMID:
39049378
Abstract
This study investigates how to reduce nurses' repetitive electronic nursing record tasks. We applied generative AI by learning nursing record data practiced with virtual patient data. We aim to evaluate generative AI's usefulness, usability, and availability when applied to nursing record creation tasks. The nursing record data collected through the electronic nursing record system for nursing students without privacy issues is in the form of NANDA, FocusDAR, SOAPIE, and narrative records. We trained 50,000 nursing record data and upgraded the performance through generative AI and fine-tuning. A separate API was used to connect with the practice electronic nursing record system, and 40 experienced nurses from a university hospital conducted tests. The electronic nursing record, through generative AI, is expected to contribute to easing the workload of nurses.