concept_path is the primary key in the concept dimension, so the path for 540.1 
has to be different from the path for K35.3.

They might differ only in their final segment, though; something like:


\Diagnosis\icd\icd09\001-999.99\520-579.99\540-543.99\540\540.1\

\Diagnosis\icd\icd09\001-999.99\520-579.99\540-543.99\540\K35.3\

presuming WHO says that K35 coincides with 540.

If the upper levels of ICD9 and ICD10 don't coincide that way, we may have to 
have completely independent diagnosis hierarchies. It's no trouble to represent 
in the database, but it makes the I2b2 UI messy.

________________________________
From: Wilson Nathan [[email protected]]
Sent: Thursday, January 16, 2014 2:24 PM
To: Dan Connolly
Cc: '[email protected]'
Subject: RE: PCORI Information Model Comments/Questions

OK.  Thanks for the clarifications.

One more clarification on slide 5, are the CONCEPT_PATHs unique to their source 
vocabulary or not?

For ICD-9 code 540.1 would the CONCEPT_PATH be (for example) 
\Diagnosis\icd\icd09\001-999.99\520-579.99\540-543.99\540\540.1\
For ICD-10 code K35.3 would the CONCEPT_PATH be (for example) 
\Diagnosis\icd\icd10\K00-K95\K35-K38\K35\K35.3
Or are they both the same unique CONCEPT_PATH (i.e. \Diagnosis\abc\def\xyz\)?

No worries about the top posting.

Nathan


From: Dan Connolly [mailto:[email protected]]
Sent: Wednesday, January 15, 2014 5:51 PM
To: Wilson Nathan; Subject: RE: PCORI Information Model Comments/Questions

Yes, Observer Dimension and Provider Dimension are synonymous, in my 
understanding.

On slide 5, the concept_cds are all the same because it's just one 
observation_fact we're looking at. Rather than "combining terms into single 
concepts" I'd say we're relating one concept_cd to multiple concept paths, 
which is fairly normal in i2b2. I'm not sure I can explain this in email. It 
might need a shared screen or face-to-face time at a whiteboard.

17 isn't a count of instances; it's just an identifier/key for an instance.

Thanks for the review comments!

p.s. Please excuse 
top-posting<http://en.wikipedia.org/wiki/Posting_style#Top-posting>. I'd rather 
put my answers with the respective question, but Outlook is the only mailer 
KUMC supports. Sigh.
________________________________
From: Greater Plains Collaborative Software Development 
[[email protected]] on behalf of Wilson Nathan [[email protected]]
Sent: Wednesday, January 15, 2014 4:37 PM
To: [email protected]<mailto:[email protected]>
Subject: PCORI Information Model Comments/Questions
I understand that PCORI/GPC Standard Data 
Model<https://docs.google.com/presentation/d/1sre0KAUOG_FtSioecLhr2SBHcdd8tZQzuaKGBmtVHWs/edit?usp=sharing>
 is a work in progress, but I do have a few questions about the content in a 
few of the slides.

Slide 2:
·         Should the Observer Dimension actually be Provider Dimension?
I looked into the i2b2 documentation and it looks like Observer Dimension is 
from an earlier version of i2b2

Slide 5:
·         Why are the CONCEPT_CDs the same for each of the concept paths?
It appears that you are attempting to combine equivalent terms/codes into 
single concepts (similar to the UMLS), but is that the best solution here?
It works for the slide because each of the underlying terms/codes have a 1 to 1 
relationship to each other, but how are we planning on addressing many to 1 and 
1 to many instances of terms/codes that can and will combine to form a single 
concept?

Slide 7:
·         Why are there 17 instances for each row?
My understanding of how INSTANCE_NUM is used, is to identify each unique 
instance of a row as defined by its ENCOUNTER_NUM, PATIENT_NUM, CONCEPT_CD, 
PROVIDER_ID, and MODIFIER_CD values.  Since the MODIFIER_CDs are different in 
each row shouldn't the INSTANCE_NUM be 1.  17 indicates that the data is 
duplicated 17 times.

Nathan S. Wilson
Ontologist
Health Information Management Center (HIMC)
______________________________________________________________________________________________________
301 S. Westfield Rd - Suite 320
Madison, WI  53717
Phone: (608) 890-5780
eMail: [email protected]<mailto:[email protected]>



Reply via email to