Driving patient care

Oakland University: Clinicians seek evidence-based medicine for informed decision-making. Knowledge graphs (KGs) aid in clinical decision support systems, yet current frameworks lack automation and context, diminishing relevance over time. This innovation proposes a fully automated, PICO-enabled KG framework utilizing concept extraction and neural networks. Evaluation demonstrates significant improvement, with 93% accuracy in aneurysm and 82% in COVID datasets for PICO classification. Relationship extraction achieves 96% accuracy, with precision and recall at 92% and 90% respectively.