Clarity on ICD10 Dictionary and Concept Matching

[Matthew Vita <ma...@gmail.com>]

Hi Sean,

If you recall earlier this month, I noted that I had to type the full "type
2 diabetes mellitus" to get a match with my ICD10 dictionary:
http://mail-archives.apache.org/mod_mbox/ctakes-dev/201709.mbox/%3CCAOV_6RL8rpg0AirHx4w+UAc6d+OABukfgsOz5ZvaoRF2D8V0hg@mail.gmail.com%3E
(I made the dictionary myself and posted my solution to YouTube:
https://www.youtube.com/watch?v=4aOnafv-NQs)

I appreciate the response you gave, pointing me to the "type 2 diabetes"
direct matches in MEDLINEPLUS, MSH, NCI, CHV, etc.

Those make sense to me, but I worry that users are typically used to just
ICD10 and won't understand the other coding schemas. Can you provide
details on improving the ICD10 dictionary concept matches? In the worst
case scenario, can I update the dictionary by hand to improve certain
common concept matches?

Thanks,

Matthew Vita
www.matthewvita.com

RE: Clarity on ICD10 Dictionary and Concept Matching [EXTERNAL]

["Finan, Sean" <Se...@childrens.harvard.edu>]

Hi Matthew,

It is possible to both
1. Edit the primary hsql dictionary, and
2. Add secondary dictionaries to your ctakes pipeline.

You can find information on the web related to editing an hsql database.  There is some random info on the web regarding the creation and use of bsv files for dictionaries, like here http://mail-archives.apache.org/mod_mbox/ctakes-dev/201510.mbox/%3Cf8bb8fbaf4d64d04814732dd3f8465ca@CHEXMAIL1A.CHBOSTON.ORG%3E

I am in the process of putting something on the wiki describing the best way to create and use a simple bsv dictionary.  What I really need to do is snag a couple of days and make a series of videos.

Sean

-----Original Message-----
From: Matthew Vita [mailto:matthewvita48@gmail.com] 
Sent: Friday, September 29, 2017 1:47 AM
To: dev@ctakes.apache.org
Subject: Clarity on ICD10 Dictionary and Concept Matching [EXTERNAL]

Hi Sean,

If you recall earlier this month, I noted that I had to type the full "type
2 diabetes mellitus" to get a match with my ICD10 dictionary:
https://urldefense.proofpoint.com/v2/url?u=http-3A__mail-2Darchives.apache.org_mod-5Fmbox_ctakes-2Ddev_201709.mbox_-253CCAOV-5F6RL8rpg0AirHx4w-2BUAc6d-2BOABukfgsOz5ZvaoRF2D8V0hg-40mail.gmail.com-253E&d=DwIBaQ&c=qS4goWBT7poplM69zy_3xhKwEW14JZMSdioCoppxeFU&r=fs67GvlGZstTpyIisCYNYmQCP6r0bcpKGd4f7d4gTao&m=8cccokAs_n3WUvM7HdWmV90l8PTJ07yDpKJJaXCYG_0&s=Hwppz_WKUTOKPmNiYGJyBOZdKLfpe4_iK_zBWBdXm00&e=
(I made the dictionary myself and posted my solution to YouTube:
https://urldefense.proofpoint.com/v2/url?u=https-3A__www.youtube.com_watch-3Fv-3D4aOnafv-2DNQs&d=DwIBaQ&c=qS4goWBT7poplM69zy_3xhKwEW14JZMSdioCoppxeFU&r=fs67GvlGZstTpyIisCYNYmQCP6r0bcpKGd4f7d4gTao&m=8cccokAs_n3WUvM7HdWmV90l8PTJ07yDpKJJaXCYG_0&s=ImCKEUl2XqxuodslayVwpP44_bQ6yRXDO0CNcAC6HWA&e= )

I appreciate the response you gave, pointing me to the "type 2 diabetes"
direct matches in MEDLINEPLUS, MSH, NCI, CHV, etc.

Those make sense to me, but I worry that users are typically used to just
ICD10 and won't understand the other coding schemas. Can you provide details on improving the ICD10 dictionary concept matches? In the worst case scenario, can I update the dictionary by hand to improve certain common concept matches?

Thanks,

Matthew Vita
www.matthewvita.com