Day 1: Tuesday, 13 September-
Extracting the structure in medical texts
Session convener: Julia Kasmire, University of Manchester
Language is a useful medium for transmitting information - but it also contains context-based information that is very valuable. We propose looking at person-first or identity-first language in medical texts to identify patterns that shed light on how people talk about themselves and others. As an example, a person-first description would be " a person with autism" while an identity-first description would be " an autistic person". While individuals may prefer one or the other for various reasons, there are likely to be patterns in how communities of researchers use this language. For example, researchers may refer to children with one description but refer to adults with the other or there may be a clear switch from one description to the other over time. These patterns can be made clear through a systematic study of medical texts through the use of natural language processing methods. Only once the patterns are extracted and quantified, can the meaning behind the patterns be understood. This session describes a novel research project around medical language and seeks audience feedback on the research questions, methods, challenges, and potential benefits.