I'm sure I'm not the first one to struggle to understand this, but I was unable to find an explanation in the mailing list archives. I am wondering about how to represent the values of vital signs (such as temperature, pulse, blood pressure, etc) when annotating some text.
For example, if I have the following text: Temperature: 98.8, pulse: 65, bp 120/80 and I'd like to identify the values associated with each so I can pull them out later into structured data fields somewhere else. Right now running this through the cTAKES clinical pipeline I get the following relevant annotations: NumToken: 98.8 NumToken: 65 NumToken: 120 NumToken: 80 MeasurementAnnotation: 120/80 SignSymptomMention (with UMLS CUI C0005823): bp I'm disappointed that it didn't annotate "temperature" or "pulse", but that's probably a configuration problem on my side. So what I'd like to do now is to have the 120/80 value be annotated as as blood pressure vital sign. Do I do that using SignSymptom type? If so, where does the value 120/80 go? Does it go into the "mentions" array? That doesn't seem quite right for some reason. My idea was to create a new annotator that looks at the combination of SignSymptomMention and NumToken (or even better MeasurementAnnotation) and creates a representation of the sign with its value and other pertinent attributes in the CAS (probably as a SignSymptom type). How do others do this? Thanks, Dave
