Discovering Related Clinical Concepts Using Large Amounts of Clinical Notes

Abstract

The ability to find highly related clinical concepts is essential for many applications such as for hypothesis generation, query expansion for medical literature search, search results filtering, ICD-10 code filtering and many other applications. While manually constructed medical terminologies such as SNOMED CT can surface certain related concepts, these terminologies are inadequate as they depend on expertise of several subject matter experts making the terminology curation process open to geographic and language bias. In addition, these terminologies also provide no quantifiable evidence on how related the concepts are. In this work, we explore an unsupervised graphical approach to mine related concepts by leveraging the volume within large amounts of clinical notes. Our evaluation shows that we are able to use a data driven approach to discovering highly related concepts for various search terms including medications, symptoms and diseases.

DeepSuggest: Using Neural Networks to Suggest Related Keywords for a Comprehensive Search of Clinical Notes

“…Also, creating and maintaining ontologies up to date is a timeconsuming manual process. The unsupervised neural networks, such as Word2Vec, capture a larger number of related concepts in the clinical note corpus that are beyond the static ontology terms (e.g., generic and user defined abbreviations and common misspellings), which supports the work of Ganesan et al 10 While the intelligent assistance of DeepSuggest is rudimentary, it is surprisingly effective. DeepSuggest recommends a larger number of relevant words due to its much larger dictionary derived from the local corpus (►Table 4).…”

“…The residents conducted 11 predefined queries and marked the top suggested 60 terms as relevant or not relevant. This set of 11 queries were primarily inspired from the previous work of Ganesan et al, 10 and modified by the authors to accommodate a broad category of medical words in pediatric setting (e.g., drugs, devices, procedures, diseases, and symptoms). For instance, "iPhone" (Apple, United States) was added to test the capability of the system to handle nonmedical terms.…”

“…1 Ontologydriven approaches also lack timely updates due to the high curation costs for such efforts. 10 Unsupervised shallow neural networks, such as Word2Vec 11 and GloVe, 12 have been effective at embedding related concepts from unstructured and unannotated texts when a large corpus is available. [13][14][15][16][17][18] Ye and Fabbri 19 extensively evaluated different approaches to identify similar terms using semantic embeddings.…”

Extracting similar terms from multiple EMR-based semantic embeddings to support chart reviews

“…Third, the constructions of user types and task types can be formalized and made more fine-grained, for example, by categorizing MD users by discipline or skill. Fourth, we utilized semantic embeddings to identify similar words, while other methods could be used find related terms like graphical models [33], Fifth, we only included unigrams when training EMR-based embeddings in our current study. We did try word embeddings based on bi-grams or tri-grams.…”

Finding answers to COVID-19-specific questions: An information retrieval system based on latent keywords and adapted TF-IDF

“…Removing stop words avoids generating style indicator keywords that are not useful in the desired context. We have used the list of clinical stop words provided by Ganesan et al [ 36 ].…”