Development and Application of Multimodal Knowledge Graph in Precision Nursing

Abstract

Objectives: This study aims to implement and evaluate a multimodal knowledge graph technology to enhance precision nursing, thereby offering more efficient, accurate, and personalized patient care services.

Methods: Data from various sources including clinical databases, nursing training textbooks, and online platforms were collated to create a comprehensive multimodal dataset in the nursing field. Utilizing natural language processing, data mining algorithms, and graph database technology, we constructed a multimodal knowledge graph integrating textual, image, and video data. Visualization tools were employed to facilitate the interaction with and validation of the constructed graph.

Results: The developed multimodal knowledge graph encompasses 62,909 entities and 330,285 relationships, covering a wide spectrum of nursing-related aspects such as patient care, disease symptoms, and nursing techniques. This graph has been successfully applied in precision nursing to generate personalized nursing profiles, perform clinical nursing semantic searches, support real-time question-answering, and assist in personalized nursing decision-making. The system demonstrated high accuracy and reliability in clinical nursing applications, particularly in enhancing decision support.

Conclusions: The deployment of multimodal knowledge graph technology in precision nursing significantly improves the delivery of personalized care, augments clinical decision-making, and promotes educational excellence. The integration of this technology into nursing practices holds great promise for advancing patient outcomes and fostering more precise and informed healthcare interventions.