CPRS 30A, ICD-10 and SNOMED codes... why not automatically mapped?

[Kevin Toppenberg]

I have been planning on using the problem list in CPRS to store ICD-10 codes for patients.  On prior versions, this worked for ICD-9 codes.  I was unhappily surprised to find that this seems to have changed in CPRS-30A.  it seems that now, the problem lists is SNOWMED oriented.

Nancy sent me this information from the CPRS manual:

Are ICD-10 Codes Mapped to SNOMED Codes?

No, the VA does not use a mapping between ICD-10-CM terms and codes and SNOMED terms and codes. Providers must select the appropriate codes. Any updates to codes come from terminology updates and happen automatically. On the Problem List, providers define problems using SNOMED and those problems are automatically assigned an R69 ICD-10 code (an undefined diagnosis). Assigning an ICD-10 term and code along with the SNOMED term and code happens from the Encounter’s Diagnosis tab. On the Encounter’s Diagnosis tab, providers assign one or more diagnoses for the encounter using the Lexicon search tool or prepopulated sources such as the Problem List Items or encounter forms assigned by the site. If the user selects an item from the Problem List Items that has a SNOMED code, but is undefined in ICD-10 (has an R69 code), CPRS will prompt the user for a more specific ICD-10 code because an encounter cannot be completed using an R69 code. The user can then select the Add to Problem List check box to associate the SNOMED and ICD-10 codes for this specific instance only and make the ICD-10 code show up in the problem’s detailed display. *Note:* Assigning an ICD-10 code to a SNOMED term on the Problem List does not mean the terms are further connected. To link the SNOMED term and ICD-10 term, the provider will have to manually define the relationship each time.

...

To further help providers find the terms they need, CPRS allows sites to create “pick lists” that sites can use to create a list of frequently used terms which will be readily available in the left pane of the Add a New Problem dialog. Providers who want “pick lists” created will need to request this from a Clinical Application Coordinator (CAC). Users cannot create the lists themselves. A CAC must create the list and assign it.

CPRS User Guide: GUI Version 11th entry from the bottom in the VDL for CPRS

--

I have never had a reason to use SNOMED codes, and I wasn't exactly sure what they even are.  

Wikipedia says:

SNOMED CT or SNOMED Clinical Terms is a systematically organized computer processable collection of medical terms providing codes, terms, synonyms and definitions used in clinical documentation and reporting. SNOMED CT is considered to be the most comprehensive, multilingual clinical healthcare terminology in the world ...

 

...

SNOMED CT consists of four primary core components:

1. Concept Codes - numerical codes that identify clinical terms, primitive or defined, organized in hierarchies
2. Descriptions - textual descriptions of Concept Codes
3. Relationships - relationships between Concept Codes that have a related meaning
4. Reference Sets - used to group Concepts or Descriptions into sets, including reference sets and cross-maps to other classifications and standards

 ...

 

SNOMED CT cross maps to other terminologies, such as: ICD-9-CM, ICD-10, ICD-O-3, ICD-10-AM, Laboratory LOINC and OPCS-4. It supports ANSI, DICOM, HL7, and ISO standards. 

So it seems that instead of just having to make the difficult transition from ICD-9 to ICD-10, CPRS-30A also has thrown SNOMED in for added look-up pleasure.

Can anyone explain the following:

• Why is SNOMED not being automatically mapped to ICD-10?  This is apparently a standard thing: https://www.nlm.nih.gov/research/umls/mapping_projects/snomedct_to_icd10cm.html
• If the user has to select both SNOMED code and also an ICD-10 code, then why does this have to be done ONLY on the encounter form?  Why can't it be done from the problem tab?

Thanks

Kevin

[Kevin Toppenberg]

In ICD-10, the code R69 is "undefined diagnosis".  In researching this issue, I tried adding a problem of "gout".   The server returns the following information for me to pick between.

ORQQPL4 LEX

Called at: 6:52:29 PM

 

Params ------------------------------------------------------------------

literal GOUT

literal PLS

literal 3151009.174

literal 1

 

Results -----------------------------------------------------------------

7161849^Gout^R69.^521774^SNOMED CT^90560007^150085018^ICD-10-CM

7701780^Gout, NOS^R69.^521774^SNOMED CT^90560007^^ICD-10-CM^1

7236155^H/O: gout^R69.^521774^SNOMED CT^161451004^251601017^ICD-10-CM

8054983^History of- Gout (Situation)^R69.^521774^SNOMED CT^161451004^^ICD-10-CM^3

8172723^History of - gout^R69.^521774^SNOMED CT^161451004^^ICD-10-CM^3

8208753^History of gout (situation)^R69.^521774^SNOMED CT^161451004^2606884019^ICD-10-CM^3

8231642^History of gout^R69.^521774^SNOMED CT^161451004^2986691018^ICD-10-CM^3

7579147^Gouty tophus^R69.^521774^SNOMED CT^402469004^1781601013^ICD-10-CM

7808027^Tophus^R69.^521774^SNOMED CT^402469004^7521017^ICD-10-CM^8

7280389^Gouty arthropathy^R69.^521774^SNOMED CT^190828008^293392012^ICD-10-CM

7026031^Gouty tophi of ear^R69.^521774^SNOMED CT^14763005^25080011^ICD-10-CM

7623293^Gouty tophus of pinna^R69.^521774^SNOMED CT^14763005^476698015^ICD-10-CM^11

7131969^Chronic tophaceous gout^R69.^521774^SNOMED CT^73877009^122686012^ICD-10-CM

7280391^Chronic gouty nephropathy^R69.^521774^SNOMED CT^190829000^293393019^ICD-10-CM

7733930^Chronic urate nephropathy^R69.^521774^SNOMED CT^190829000^293395014^ICD-10-CM^14

7733931^Gouty nephropathy^R69.^521774^SNOMED CT^190829000^293394013^ICD-10-CM^14

7460039^Drug for the treatment of gout adverse reaction^R69.^521774^SNOMED CT^292671009^432818012^ICD-10-CM

7124843^Rheumatoid arthritis^R69.^521774^SNOMED CT^69896004^116082011^ICD-10-CM

7680561^Atrophic arthritis^R69.^521774^SNOMED CT^69896004^116083018^ICD-10-CM^18

7680562^Chronic rheumatic arthritis^R69.^521774^SNOMED CT^69896004^116084012^ICD-10-CM^18

7680563^Proliferative arthritis^R69.^521774^SNOMED CT^69896004^1788127016^ICD-10-CM^18

7680564^RA - Rheumatoid arthritis^R69.^521774^SNOMED CT^69896004^1233202011^ICD-10-CM^18

7680565^RhA - Rheumatoid arthritis^R69.^521774^SNOMED CT^69896004^1233204012^ICD-10-CM^18

7680566^Rheumatic gout^R69.^521774^SNOMED CT^69896004^116085013^ICD-10-CM^18

7680567^Rheumatoid disease^R69.^521774^SNOMED CT^69896004^1233203018^ICD-10-CM^18

25 matches found

Notice that piece 3 of each of these options is R69.

If I would to add SNOMED CT -> ICD10 mapping, it seems it could be done in this RPC call.

I'm still wondering why the VA hasn't done this.  What concept am I missing or what business logic am I not understanding?

Kevin

[Steven McPhelan]

The LEX APIs do allow for specifying a specific coding system like ICD-10 (if memory serves me correctly).  The CPRS developers had design specifications which content experts had input to.  What you are asking for is an upgrade perhaps to the CPRS RPCs that might allow the end user to select a specific coding system.  However, the primary question is whether this behavior is the desired behavior that the VA wants.  Apparently not.  The VA may not want any provider to select any coding system the provider wishes.  One purpose of these codes is for billing.  Should the coders do the mapping of SNOMED to ICD-10?   That is a business issue and a business process issue.  No two businesses are alike.

Steve

We have no government armed with the power capable of contending with human passions, unbridled by morality and true religion. Our constitution was made only for a moral and religious people. It is wholly inadequate to the government of any other. - John Adams

[Stephen...@va.gov]

Apparently, there was a high level taskforce formed to address this proactively and NLM did develop a tool as you note (with high level VA participation). I inquired to local ICD10 transition team in my VISN and was told tool was never implemented in VA systems.

[Nancy Anthracite]

The picking of an ICD10 code to go with the SNOMED codes is done entirely manually and is only done on the encounter form. There is no helpful mapping. The question in my mind is why not when you go to add a problem on the problem list tab?

 

My guess is that the were hard up against a deadline. I understood that at a certain point the contract was over with the main contractor for the ICD10 project and if they added something else, they would have to start all over with a new contractor which would have a steep learning curve and price. I don't know if that issue was ever fixed, but it wasn't, that probably meant what was done was all that would get done and these additional logical steps just never got done.

 

--

Nancy Anthracite

[David Carter]

SNOMED was ushered in with Meaningful Use.  It is terminology based problem system just as Nancy and Kevin have pointed out whereas ICD10 is a more granular and robust Dx coding system with more clinical value that 9 had.  10 is used as a billing code, but so are other code sets also like CPT.  It just depends on who they are doing the billing to and what company accepts what.  

As for the linking of the ICD10 and SNOMED there isn't a 1 to 1 matching and as stated above there is a clinical decision involved since there are so many to pick from that can be used in various conditions.  

Side note: It really gets exciting when it comes to mapping the taxonomies in the reminders since all code sets have to be mapped in.

There are several new tools in VistA with more coming. There also several online that are helpful for providers.  The Vehu campus has some good material to help get through the curve too.  I will try to get a few of those online outside source links and reply back to this thread again. There is a software tool I have seen in house, but I don't have it.  I have seen it in conference calls though.  Without knowing a name of the one mentioned above I can't say whether its the same one.  My brain isn't wired to Dx and my brain was already spinning, so I didn't get it.  I have gotten a even higher respect for the knowledge and skills of our providers just listening in and helping present.  All that Latin just makes my head spin when they start jumping into the weeds with all the different cases.  I really do appreciate what they do.  

The Project Managers for both have done absolutely awesome too.  The enormity of it all was amazing and they were all over it and still continue to ensure a pretty smooth transition.  

There's going to be some moaning and groaning along with some gnashing of teeth for a little while, but the users will adapt.    

--
--
http://groups.google.com/group/Hardhats
To unsubscribe, send email to Hardhats+u...@googlegroups.com

---
You received this message because you are subscribed to the Google Groups "Hardhats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hardhats+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[David Carter]

http://www.icd10charts.com/chartbuilder

https://vsac.nlm.nih.gov/

https://uts.nlm.nih.gov/home.html

http://www.ihtsdo.org/snomed-ct/

One lookup is in the taxonomy menu.  After typing in a term and enter it goes into another screen with all the code sets listed.  Just hit enter on the code set you want to see the results of the search on.  There is another in the works for the reminder package to lookup term sets.

[David Carter]

Forgot the campus... www.myvehucampus.com

[Kevin Toppenberg]

David, Nancy (and others),

Thanks for your replies.  See below. 

On Friday, October 9, 2015 at 10:31:38 PM UTC-4, davidc wrote:

SNOMED was ushered in with Meaningful Use.  It is terminology based problem system just as Nancy and Kevin have pointed out whereas ICD10 is a more granular and robust Dx coding system with more clinical value that 9 had.  10 is used as a billing code, but so are other code sets also like CPT.  It just depends on who they are doing the billing to and what company accepts what.  

As for the linking of the ICD10 and SNOMED there isn't a 1 to 1 matching and as stated above there is a clinical decision involved since there are so many to pick from that can be used in various conditions.  

Mapping from one code set to another is always a problem.  As to 1:1 mapping, a Dr. Kin Wah Fung with UMLS has tried to make this as close as possible with an I-MAGIC tool (http://imagic.nlm.nih.gov/imagic/code/map) that did a good job with the few sample entries that I tried.

 

Side note: It really gets exciting when it comes to mapping the taxonomies in the reminders since all code sets have to be mapped in.

There are several new tools in VistA with more coming. There also several online that are helpful for providers.  The Vehu campus has some good material to help get through the curve too.  I will try to get a few of those online outside source links and reply back to this thread again. There is a software tool I have seen in house, but I don't have it.  I have seen it in conference calls though.  Without knowing a name of the one mentioned above I can't say whether its the same one.  My brain isn't wired to Dx and my brain was already spinning, so I didn't get it.  I have gotten a even higher respect for the knowledge and skills of our providers just listening in and helping present.  All that Latin just makes my head spin when they start jumping into the weeds with all the different cases.  I really do appreciate what they do.  

I started looking at this presentation about mapping: https://www.nlm.nih.gov/research/umls/mapping_projects/mapping_2012himss.pptx any my head was spinning as much as yours, and I am a provider...

 

The Project Managers for both have done absolutely awesome too.  The enormity of it all was amazing and they were all over it and still continue to ensure a pretty smooth transition.  

There's going to be some moaning and groaning along with some gnashing of teeth for a little while, but the users will adapt.    

My concern is not so much about adapting and more about having to do double work: first looking up a problem with SNOMED, and then later having to look up the problem again with ICD-10.

I have put in a request to download the mapping file.  They will have to approve my request and get back with me in 3 days.

One BIG factor is that it has a very complicated license, including restrictions that I will ensure that it is not distributed outside the United States.  Since I am making open source software, I think I will probably not be able to use that.

Kevin 

[Nancy Anthracite]

I would think that UMLS might have it as David said. One problem with that for open source software is the SNOMED license, which will not be a problem only in the US.

 

I would think narrowing the choice, not one to one mapping, would be helpful. ICD9 was mapped to SNOMED before the transition on October 1. I did not check, but I suspect that was one to one so taking advantage of that code will not be a likely solution, but taking advantage of the Lexicon utilities to look up codes might be something that could be leveraged. It may be that it is too much to ask in a huge hospital with innumerable queries to do this, but probably not in a small practice.

 

I am almost positive that the US extension to SNOMED is included in the SNOMED codes that are in the pick list.

 

--

Nancy Anthracite

> >> Nancy Anthracite

> >> ---

> >> You received this message because you are subscribed to the Google Groups

> >> "Hardhats" group.

> >> To unsubscribe from this group and stop receiving emails from it, send an

> >> email to hardhats+u...@googlegroups.com <javascript:>.

[Nancy Anthracite]

https://www.nlm.nih.gov/research/umls/mapping_projects/snomedct_to_icd10cm.html

SNOMED to ICD10 mappings

--

Nancy Anthracite

 

On Sunday, October 11, 2015 09:31:14 AM Nancy Anthracite wrote:

> I would think that UMLS might have it as David said. One problem with that

> for open source software is the SNOMED license, which will not be a problem

> only in the US.

>

> I would think narrowing the choice, not one to one mapping, would be

> helpful. ICD9 was mapped to SNOMED before the transition on October 1. I

> did not check, but I suspect that was one to one so taking advantage of

> that code will not be a likely solution, but taking advantage of the

> Lexicon utilities to look up codes might be something that could be

> leveraged. It may be that it is too much to ask in a huge hospital with

> innumerable queries to do this, but probably not in a small practice.

>

> I am almost positive that the US extension to SNOMED is included in the

> SNOMED codes that are in the pick list.

>

> > David, Nancy (and others),

[Kevin Toppenberg]

Thanks Nancy.  I had this link in my initial post.

Yesterday, I spent all day going through the problem list code, figuring out how it works.  I have a partial solution that asks the user for a linking ICD right after they choose the SNOMED problem diagnosis.

For example, I will show below the process of adding anew problem.

I do a search for "bursitis" because the patient has bursitis of their left knee, and the dialog search function gives me options for bursitis of various body parts.

After selecting the SNOMED code that I want, the user will now immediately be asked to select a linked ICD code

After clicking OK, I have the problem ready to accept.

And yes, this little bit took me all day.  As a side note, there are other changes to the above form that have to do with problem-oriented progress notes, comprised of multiple subcomponents, one for each problem.  But I am not done with that part yet.

Kevin

[Syed]

It is kind of similar in I h s world of course they get patches from VA too. Focus is to keep the problem list up to date at every visit. In hospital environment because of good support from coders and billers it is lot better for providers.

I still do not understand why snomed terminology was introduced when we have icd coding system. Looks like to me every time with new icd patches we also need to keep snomed terminology uptodate or vice versa.

Add to the problem list check box in the encounter was lot easier during icd9 as compared to now ( snomed + icd10).  This will open new challenges and ofcourse with the time we will see resolutions

[Kevin Toppenberg]

From what I read SNOMED is more clinically oriented and works better in EMR's etc.  But I agree, it seems redundant. 

Kevin

[stevenli]

Perhaps for Pathologist and Radiologists SNOWMED works, but for the remainder ICD9 ran the world in billing/documentation/forms. 

When VA went SNOWMED for MU2 compliance on the problem list, it forced clinicians to update the PL to snowmed even though the rest of the world remained on ICD9 at the GUI layer and chose to get MU2 compliance by doing this under the staff facing PL which remained ICD9.

Epic did this too and mapped out to ICD10 with only minor hiccups on Oct 1 where mapping from 9 to 10 wasn't available.  I don't know what SNOWMED is happening underneath, but it certainly is working compared to what's happening at the VA right now.

My colleagues essentially have had to abandon their entire PL and rebuild their ICD10 PL (with an inefficient ICD10 search engine to boot).  If they inadvertantly choose an old SNOWMED code, it gets kicked out and worse, ties up the pairing for the next checkout.

The end result-- many clinicians now is use a single ICD10 code at checkout to avoid causing even more downstream rework of the PL which they had to rebuild just 2 years ago from ICD9 to snowmed.  

That's going to mess up VERA, HCC and 3rd party collections big time.  

[David Carter]

For clarification.  SNOMED and ICD10 both were decided upon much higher than VA or any vendor.  Government Departments have to follow Government laws, which also includes updating code sets quarterly. All of the code sets are handled by the Lexicon and unless something has changed that is what the search tool uses.  

I am vendor neutral because I have to be for what I do since the VA accommodates to several vendors even proprietary ones.  If a vendor did well with their transition then hats off to them. 

As for the comparison and opinion it would be interesting to see the findings/results of the comparison other than hearsay or opinions.  Please feel free to email me directly or have your colleagues look me up if you like.  I'd be glad to help them or reach out to other SMEs that might be able to help them even more if I'm not able to address it.  However, please let them know I do not not douse the fire of frustration with gasoline.  It never helps the user and is a disservice that makes matters worse.

It is a entirely different environment to collaborate in and work jointly with several vendors (with software tied in within numerous places) than what it is with just one.  My statement in a previous post was not to put in a plug for one organization or to pit one vs. another, but to give credit where credit was due.  It wasn't a pop shot at best or even an opinion.  Whether the glass is half empty or half full or one pours smoother than another its still a man made glass and will not be perfect.  There isn't a man made system that will ever be perfect and/or never contribute to hiccups.

Nancy, 

The gnashing of teeth has already begun as stated before.  Clinically updating the chart was part of what I was referencing earlier.  It gives the feeling of double the work at first.  From your and Kevin's feedback I think you are already seeing it too and mitigating the issues.

  

--

--
http://groups.google.com/group/Hardhats
To unsubscribe, send email to Hardhats+u...@googlegroups.com

---
You received this message because you are subscribed to the Google Groups "Hardhats" group.

To unsubscribe from this group and stop receiving emails from it, send an email to hardhats+u...@googlegroups.com.

[W.G...@t-online.de]

Within this context the German "ICD-10-Dignosen-Thesaurus" might be of interest: It was originally proposed and constructed by me, using the BAIK Thesaurus, maintained since the late 60s using wordlists from dictated physician reports. To link these wordlists to ICD-9 and ICD-10 has been successful. DIMDI, the German counterpart of NLM, decided in the late 90s to replace the German translation of the alphabetical ICD-10-index with this "ICD-10 Diagnosen Thesaurus". They maintain it now officially with yearly updates. It is used in Austria, Switzerland and Germany (as well in hospitals and ambulatory care).

 

Best regards

 

Wolfgang Giere

 

PS: Regarding the early medical text analysis efforts I would like to mention and thank Scotty Pratt from NLM . Perhaps some oldies of the readers still know him or of him.

[Kevin Toppenberg]

I love it when grass-roots efforts by users take root and become useful for others.

Kevin

[Nancy Anthracite]

So is this written in German? Also, is it open source (and I think the latter
answer is no as I recall).

--
Nancy Anthracite

[W.G...@t-online.de]

Yes, it is written in Geman. No, it is not open source if bought as book, but the files are available from DIMDI without payment, open source:

 

Quote from www.dimdi.de:

 

>>Seit Version 2005 wird der ICD-10-Diagnosenthesaurus als Alphabetisches Verzeichnis zur ICD-10-GM weitergeführt. Die Dateien finden Sie im Downloadcenter Klassifikationen im jeweiligen Jahresversionsordner zur ICD-10-GM im Unterordner Alphabet.<<

 

Best regards

 

Wolfgang Giere

 

 

 

Am Dienstag, 13. Oktober 2015, 09:23:21 schrieb Nancy Anthracite:

> So is this written in German? Also, is it open source (and I think the

> latter answer is no as I recall).

>

> > Within this context the German "ICD-10-Dignosen-Thesaurus" might be of

[Nancy Anthracite]

How would Google translate do translating them? I suspect it would be quite
amusing! :-)

--
Nancy Anthracite

[Kevin Toppenberg]

On Sunday, October 11, 2015 at 8:48:17 AM UTC-4, Kevin Toppenberg wrote:

David, Nancy (and others),

Thanks for your replies.  See below. 

On Friday, October 9, 2015 at 10:31:38 PM UTC-4, davidc wrote:

SNOMED was ushered in with Meaningful Use.  It is terminology based problem system just as Nancy and Kevin have pointed out whereas ICD10 is a more granular and robust Dx coding system with more clinical value that 9 had.  10 is used as a billing code, but so are other code sets also like CPT.  It just depends on who they are doing the billing to and what company accepts what.  

As for the linking of the ICD10 and SNOMED there isn't a 1 to 1 matching and as stated above there is a clinical decision involved since there are so many to pick from that can be used in various conditions.  

Mapping from one code set to another is always a problem.  As to 1:1 mapping, a Dr. Kin Wah Fung with UMLS has tried to make this as close as possible with an I-MAGIC tool (http://imagic.nlm.nih.gov/imagic/code/map) that did a good job with the few sample entries that I tried.

 

Side note: It really gets exciting when it comes to mapping the taxonomies in the reminders since all code sets have to be mapped in.

There are several new tools in VistA with more coming. There also several online that are helpful for providers.  The Vehu campus has some good material to help get through the curve too.  I will try to get a few of those online outside source links and reply back to this thread again. There is a software tool I have seen in house, but I don't have it.  I have seen it in conference calls though.  Without knowing a name of the one mentioned above I can't say whether its the same one.  My brain isn't wired to Dx and my brain was already spinning, so I didn't get it.  I have gotten a even higher respect for the knowledge and skills of our providers just listening in and helping present.  All that Latin just makes my head spin when they start jumping into the weeds with all the different cases.  I really do appreciate what they do.  

I started looking at this presentation about mapping: https://www.nlm.nih.gov/research/umls/mapping_projects/mapping_2012himss.pptx any my head was spinning as much as yours, and I am a provider...

 

The Project Managers for both have done absolutely awesome too.  The enormity of it all was amazing and they were all over it and still continue to ensure a pretty smooth transition.  

There's going to be some moaning and groaning along with some gnashing of teeth for a little while, but the users will adapt.    

My concern is not so much about adapting and more about having to do double work: first looking up a problem with SNOMED, and then later having to look up the problem again with ICD-10.

I have put in a request to download the mapping file.  They will have to approve my request and get back with me in 3 days.

One BIG factor is that it has a very complicated license, including restrictions that I will ensure that it is not distributed outside the United States.  Since I am making open source software, I think I will probably not be able to use that.

So I got approval for my license for the UMLS SNOMED->ICD10 mapping file yesterday, and I downloaded it.  There is a CSV (actually TSV) format file that could be used to narrow down the possible ICD's for a given SNOMED code to just a few possible ICD's.

I could write open source software that takes this file and import it for use in VistA.  Then let every individual user get their own license and import the file....

I'll have to think about this.

Kevin

[Nancy Anthracite]

I assume this was the UMLS license and it is related to SNOMED, so those in
the US should have no difficulty with having it passed on to them or obtaining
a UMLS license and feeling no compunction about using it.

--
Nancy Anthracite

[Kevin Toppenberg]

This was the UMLS.  I can't comment on the legality of giving this to another user in the US.  I suspect that each user is supposed to have their own license, but I'm not sure.  The license was many pages in length.  https://uts.nlm.nih.gov/license.html

Kevin