Automated clinical knowledge graph generation framework for evidence based medicine

Abstract

To practice the evidence-based medicine, clinicians are interested to find the most suitable research for the clinical decision making. The use of knowledge graphs (KGs) in evidence-based clinical decision support systems is becoming increasingly popular. However, existing KG construction frameworks are not fully automated and contextualized, thus unable to adapt to new domains and incorporate constantly changing information into their knowledge base, resulting in loss of relevance over time. Furthermore, existing KGs construction frameworks don't generate KG that provide relevant information within an acceptable response time for evidence-based practitioners because the organization of constructed subgraphs is neither topic-specific nor evidence-based PICO (Participants/Problem P, Intervention-I, Comparison C, Outcome O) query-friendly. By employing concept extraction, semantic enrichment, optimized clustering, and state of art Recurrent Neural Networks (RNNs) with BioBERT based encoded representation to categorize PICO elements and predict relationships between concepts using huge corpus of publicly available literature on COVID-19 and cerebral aneurysm, this paper proposes a topic specific, PICO enabled, and fully automated framework to curate information and create KG of different clinical domains. The evaluation shows that the proposed framework achieves significant improvement over baseline models and has 93 %, and 82 % accuracy on aneurysm and COVID data set respectively for PICO classification. Also, the relationship extraction module has an accuracy of 96 % with precision and recall being 92 %, and 90 % respectively.