The Sublanguage of Clinical Problem Lists: A Corpus Analysis

Summary-level clinical text is an important part of the overall clinical record as it provides a condensed and efficient view into the issues pertinent to the patient, or their "problem list." These problem lists contain a wealth of information pertaining to the patient's history as well as current state and well-being. In this study, we explore the structure of these problem list entries both grammatically and semantically in an attempt to learn the specialized rules, or "sublanguage" that governs them. Our methods focus on a large-scale corpus analysis of problem list entries. Using Resource Description Framework (RDF), we incorporate inferencing and reasoning via domain-specific ontologies into our analysis to elicit common semantic patterns. We also explore how these methods can be applied dynamically to learn specific sublanguage features of interest for a particular concept or topic within the domain.