Using ICD-10 for diagnoses, but really

Tags: #<Tag:0x00007f1f78769398>

Most implementers here have probably had a client request that the full ICD-10 reference list be used as a diagnosis list in OpenMRS.

The usual thing to do is to explain as we always do that the ICD-10 list is not really appropriate as a diagnosis list, that doctors don’t think in codes, that it doesn’t differentiate things that are importantly clinically distinct, etc.

But what do you do when you have a really intransigent client, who insists with absolute authority over the project that the diagnosis list must be all (or most) of the ICD-10?

Do you create a concept set, and automate the creation of tens of thousands of ICD-10 concepts? Do you keep those somewhere different from your main concept server/repository so as not to clutter things up? Do you say “well you can’t use OpenMRS then”?

There is nothing that prevents an OpenMRS from presenting clinical users a list of ICD-10 diagnoses. This could be done either by restricting the concept dictionary or by using a large convenience set. This is not the tact advocated by CIEL. There is no reason for clinicians to be forced to choose from ICD-10 code language to document care if clinical diagnoses are mapped to ICD-10. You can always use the concept_map table to output the ICD-10 codes and descriptions to a report. It is rarely required to use anything beyond the mapped ICD-10 code. A bigger question is scope of coverage of CIEL for ICD-10 codes. We have elected so far, to not provide clinical terms for all ICD-10 codes that are not likely to be used. However, there is a growing realization that we should probably at least have a single concept that would be mapped to each fully-specified ICD-10 codes (not including header terms that are not actually usable codes). I am currently undertaking a project that would be bringing in SAME-AS Vietnamese translations for ICD-10 which would create new concepts previously missing from CIEL.

Of note, there is not a good way for capturing disjunctive terms (Abscess and Cellulitis of Face) as the actual meaning of the code is to capture Abscess of Face OR Cellulitis of Face. Patients usually have one or the other and documenting that they might have one of a list of things in the EMR is not correct. Also, patients do not have “Other disease of blood and blood-forming organs”. Patients have a disease or disorder which should be named, and it just might map to that ICD-10 code this year, but in 2 years when that code is specifically added, then it would have to be remapped to that code. Always best to document what the patient actually has and let the mapping tables assign the ICD-10 and SNOMED codes, for example.


@bistenes , i did some work related to that sometime back, we were creating a custom concept set of Cancer Diagnoses constrained to a given rage of ICD-10

1 Like