Everyone interested in concept management: We have discussed the need to bring together those people most involved with maintaining concept dictionaries and maps. To help implementers and designers better understand concept management and curated dictionaries like CIEL in particular, I am proposing an open “Office Hour” to occur on alternating Wednesdays and Thursdays at 8:30pm IST | 6pm Nairobi | 5pm Cape Town | 1pm UTC | 10am Boston | 7am Seattle. This will start on Wednesday January 20th, then Thursday January 28th, etc.
If you would like to get on a calendar invite/mailing list, please let me know by providing me your email address. Also, if you cannot make these dates/times and are still interested, also reach out to me.
We will provide zoom information to invitees. Everyone who is interested is invited.
Fantastic, @akanter! Is this something we can put on the OpenMRS Calendar and include in our What’s Happening This Week posts? Let me know how I can help.
Only caveat: the FHIR Squad call happens every week at this time, at 10am EDT on Wednesdays. So I’m selfishly wondering: is there any chance we could move the office hours to 8am EDT, or 12 EDT?
Whoops, clearly I need to read twice before posting a reply. That totally makes sense. Thank you Andy!!
And agreed with @jennifer above; it would be great if @jennifer could add this to the OMRS Community Calendar. Would that be okay with you @akanter? (We could ensure you’d have the ability to change the event if needed through using your openmrs.org email account)
Today’s discussion focused on mappings to implementation dictionaries (as opposed to SNOMED, etc.). Thank you Ellen. We identified a couple of issues with OCL and OpenMRS which we will document. Next week, the office hours will be on Thursday. Everyone is welcome.
Today’s discussion was about the definition of “grand parité” as a risk factor for pregnancy. Is the concept interested in whether the mother is older (having had >5 deliveries) or is it actually looking for Grand multiparity which only cares about the 5 deliveries? If the latter, why is this important if you record the number of deliveries using a numeric concept Parity? CIEL has concepts for grand multiparity to be used for visit coding (but these imply relationship to a current pregnancy, either during pregnancy, with an antenatal problem, or after delivery).Hard to know which to use as a pre-pregnancy risk assessment. We decided that the risk is Grand multiparity with current pregnancy and this is the CIEL concept mapped to the generic SNOMED concept for “grand multiparity”
Thanks for these updates Andy! I’m sorry I’m normally double booked during these office hours but I’ll try to come from time to time; in the meantime these notes are quite helpful.
CIEL has some legacy concepts mapped SAME-AS to the same SNOMED CT code. The example was HIV positive and HIV Infected. This should not happen. HIV Infected had a synonym of HIV POS. These two concepts are the same and one should be retired. The question about which and how needs to be communicated to the CIEL users.
Also, it appears that the challenge of information model comes up here as well. SNOMED has different concept IDs for observable entity and for findings/diagnoses, etc. Observable entities are designed to capture a question. So for example, Patient test status is an observable and could potentially be tested or not tested. The finding that a patient has been tested would be another concept. CIEL occasionally uses a finding as a coded concept with Y/N/U to support certain UIs. But this creates a potential conflict where there is an OBS with an affirmative concept, but with an answer that negates it. In the above example, HIV INFECTED had as an answer Negative. When doing queries of the database it is very important to know whether you are searching the question concepts or the answer concepts for your particular implementation.
The second case was around the use of plurals. The particular example was for an anatomical site using the name “peripheral nerves of lower extremity”. In general, for findings/diagnoses, we use singular forms of the concept, since the finding is usually at a particular site, and if it is meant to include more than one site, then the term explicitly says multiple sites. This comes up since many of the target codes (like in ICD) are groupers intended to categorize any site to that particular code. That is not clinical, but a reporting function. However, given anatomical site might not need to be as specific (as compared to disease sites) I agreed to allow “grouper” plurals. Diagnosis/findings will continue to be singular unless specifically required.
Please be aware that the US changes to Daylight Savings Time next week. These office hours are staying at the same time in the US which means that it will be 1 hour later for those joining internationally
Yesterday’s office hours focused on some PIH terminology and laboratory terminology. There are challenges in capturing data with the proper units/coding that can be used to enforce interoperability. One area we identified as a real mess in CIEL is the Stool Exam for OCP (ova, cysts and parasites). This is going to take some cleanup work so those folks who have been using the very old concepts in CIEL (304 and 161451) may need to consider how future concepts should be modeled. Stay tuned.
@akanter I’m noticing a few strange things in some CIEL concepts especially when it comes to Reference Ranges. Is this the best place for me to post them for consideration during your office hours, esp. since I am usually double-booked during the office hours?
Here are 2 examples:
Creatine: Reference range looked off to a physician; can you double-check?
Respiratory Rate: This looks very off to me, e.g. CIEL in production currently says a “Normal High” can be “70 breaths/minute”
@grace@akanter Ranges are another area that we should discuss about OCL and “non-breakable” changes to the concepts. PIH occasionally will change concept numerical ranges to match lab equipment. Is this considered a “non-breakable” change when we start with a CIEL concepts?
(In addition, OpenMRS concepts don’t handle ranges based on age/gender which we’ve discussed earlier.)
