Researchers from SUNY Polytechnic Institute’s College of Health Sciences are studying how artificial intelligence could support nurses by helping organize clinical documentation. A recent study by Dr. Jerome Niyirora, Dr. Lynne Longtin, Dr. Cynthia Grabski, David Patrishkoff, and Andriana Semko compared machine learning models with human nursing experts to see how accurately each could classify nursing notes into standardized Nursing Interventions Classification (NIC) categories. 

The research found that experienced nurses were still more accurate overall, especially when interpreting notes that required clinical judgment or contextual understanding. However, AI models such as GPT and Bio-Clinical BERT showed strong potential when nursing notes used clear and direct clinical language. Medication-related notes, for example, were often classified correctly because the wording was easier for AI systems to interpret. 

The study also found that AI struggled with abbreviations, numerical values, unclear wording, and interventions involving patient education, mental health assessment, or catheter care. Researchers concluded that AI may become a valuable support tool for healthcare documentation, but human expertise remains essential in clinical decision-making. 

Read more