Precision Nursing Research Based on Multimodal Knowledge Graph

Abstract

Background: Precision nursing seeks to tailor care to individual patient needs, and knowledge graphs offer a promising way to integrate diverse data for enhanced precision. However, the application of knowledge graphs in nursing remains relatively unexplored, motivating this study. Objective: This study aims to explore and apply multimodal knowledge graph technology to facilitate the development of precision nursing, providing patients with more efficient, accurate, and personalized care services. Methods: Firstly, we collected and integrated data sources, including clinical databases, nursing training textbooks, and internet data, to form a multimodal dataset in the field of nursing. Then, we used natural language processing techniques, data mining algorithms, and graph database technology to extract and represent knowledge from different data sources, constructing a nursing multimodal knowledge graph containing textual, image, and video data. After completing the graph construction, we used visualization tools to display and interactively query the graph to validate its accuracy and utility. Results: We have built a multimodal knowledge graph in the nursing domain, focusing on patients and diseases, and highlighting nursing issues, nursing techniques, nursing assessments, and disease symptoms. This comprehensive multimodal knowledge graph encompasses a total of 62,909 entities and 330,285 relationships. We have effectively applied this graph in precision nursing research, yielding favorable outcomes in the domains of personalized nursing profiles generation, clinical nursing semantic search, real-time nursing question-answering, and personalized nursing decision-making. Conclusions: This study demonstrates the value and potential applications of multimodal knowledge graph in precision nursing research. The graph provides comprehensive and precise knowledge support for nursing education, clinical practice, and decision-making, and holds the promise of further advancing and innovating nursing informatization and intelligence. And our code and databases can be accessed through the link: https://github.com/XiongLP208/NursingKnowledgePN .