LOINC code and Category code problem

186 views
Skip to first unread message

Daya

unread,
Nov 21, 2018, 11:13:51 AM11/21/18
to Cerner FHIR Developers
Hi

The patient ID in context is 5354007 

If I pull all the Observation data, which is about 105 Observation resources (following the link url for next set), there are about 63 Observations with Category unknown ("valueCode" : "unknown") where LOINC code may or may not be present. My questions with context are as follows.

 

  • Context: both LOINC code and Category code are absent
    1. If we consider one such Observation Id "M9185513", has no LOINC code (i.e. "valueCode" : "unknown"), with the valueCodeableConcept has "text" : "Dialysis catheter". 
      1. Does this mean Observation will never have LOINC code or has it not been coded properly? Another example of the same type is Observation Id M9185335 "text" : "Oxygen Therapy".
      2. If no observation matching the LOINC codes are found then should we assume the uncoded observations are not relevant since they are not coded?
    2. Related to #1, on our end if we search Observation (by code in API call or later) search for a specific LOINC code related to say 43558-6 (Guided Dialysis Catheter) change then we won't find it. Yet the information related to Dialysis Catheter is present but not coded. Is the recommendation then to not search Observation by code? If yes then what's the recommendation?
  • Context: LOINC code is present but Category code is not
    1. Observation Id M9185333 which has LOINC code 59408-5 for "SpO2" (oxygen saturation), should this not be of Category code "vital-signs"?

We need to finalize our strategy of searching Observation either by code or by category. What strategy would you recommended?


Best Regards,

~Daya


Benjamin Eichhorn (Cerner)

unread,
Nov 21, 2018, 12:39:03 PM11/21/18
to Cerner FHIR Developers
Hi Daya,

There's multiple parts to digest here so hopefully I can communicate this clearly.

For Observations that have no Observation.code and no Observation.category that's a result of our service unable to identify both the category and the code that would pertain to that Observation. This is usually due to missing mappings. You're seeing a value for Observation.code.text because that is the underlying Millennium display for that Observation, it does not necessarily represent a specific LOINC code. 

You shouldn't assume that a Observation without a code is irrelevant. Again, that's a result of our service unable to identify the specific code for the Observation. This can be a result of missing mappings. We return the information we can and leave it up to the developer/consumer to determine what to do with it in this situation.

In terms of searching by either a category or a code, that's up to you as a developer to decide. If you are only looking for specific LOINC (such as only diastolic blood pressure) you probably would want to perform a search by code. If you wanted to get the BMI, Blood Pressure, and Weight of a patient you would likely want to perform a search by category.

Overall, I should clarify that this is specifically more common in our sandbox domain rather than a clients domain, we've logged an internal investigation to close some of the gaps on the missing categories and codes you are seeing. When onboarding an app into a client's domain, whether you search by code or by category, mappings will be completed for that domain for the specific LOINC/Category that your app consumes. In a clients domain, when a specific result either is missing a category or code, that would be entirely dependent on the build which will potentially require more configuration to correct.

Let me know if you need further clarification,
Ben (Cerner)

Daya Sharma

unread,
Nov 21, 2018, 1:19:23 PM11/21/18
to cerner-fhir...@googlegroups.com
Hi Benjamin,

Thank you for your quick response.

Most of your reply is clear except this "In a clients domain, when a specific result either is missing a category or code, that would be entirely dependent on the build which will potentially require more configuration to correct.". Specifically the underlined part and the meaning of build here, are you talking about Cerner SMARonFHIR deployment in test/prod environments and that it will be client's responsibility to fix the configurations so that correct mapping can happen? What are these configurations called exactly? we have a direct line of communication with our client and we can use this information to ensure correct mappings.

Best Regards,
~Daya

--
You received this message because you are subscribed to a topic in the Google Groups "Cerner FHIR Developers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/cerner-fhir-developers/e3PC0UJx6Ew/unsubscribe.
To unsubscribe from this group and all its topics, send an email to cerner-fhir-devel...@googlegroups.com.
To post to this group, send email to cerner-fhir...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cerner-fhir-developers/5cfe9f31-ec88-4922-aed7-1cf882bf9220%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Jenni Syed (Cerner)

unread,
Nov 21, 2018, 1:28:16 PM11/21/18
to Cerner FHIR Developers
Hi Daya,

The build is done when we deploy your application into a client's domain. All mappings for FHIR are done in the cloud.

When your application goes through the code program validation, we take note of the codes and fields your application needs to have mapped, especially those that aren't likely to be mapped by default when we turn FHIR on for a specific domain.

~ Jenni

Daya Sharma

unread,
Nov 21, 2018, 1:40:47 PM11/21/18
to cerner-fhir...@googlegroups.com
Hi Jenni,

What do you mean "when we deploy your application into a client's domain"? Our's is a SaaS app and my understanding is that just like sandbox we will be issued a fhir_app_id, fhir_client_id and fhir_url (fhir spec) while we give you a redirect url, is that what you mean by deployment?

While we are on this topic could you please confirm those are the only configuration parameter we will need to enable our app in client's domain (test or prod).

Regards,
~Daya

--
You received this message because you are subscribed to a topic in the Google Groups "Cerner FHIR Developers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/cerner-fhir-developers/e3PC0UJx6Ew/unsubscribe.
To unsubscribe from this group and all its topics, send an email to cerner-fhir-devel...@googlegroups.com.
To post to this group, send email to cerner-fhir...@googlegroups.com.

Jenni Syed (Cerner)

unread,
Nov 21, 2018, 3:44:05 PM11/21/18
to Cerner FHIR Developers
Hi Daya,

For Provider and System applications, they must become part of the code program and go through validation prior to going to production. There are many "how to go to prod" questions that get answered as part of that process. While I can't speak to *everything* you might have to configure for your application from site to site, since the app is standalone, it would need to configure the FHIR URL. In our system, the *production* id will likely be the same regardless of site (it will vary on the sandbox side, so it's worth allowing the key and secret to be configured). You don't need the app id for a standalone application, nor would you need a redirect if it's B2B/System access.

I also recommend reading through the common issues page in our documentation: https://fhir.cerner.com/millennium/faq/common-issues/

In order to run an application at any given client site, we (really, the client as they have to approve production changes we make) have to whitelist the application to each site.

~ Jenni

Benjamin Eichhorn (Cerner)

unread,
Feb 1, 2019, 2:03:52 PM2/1/19
to Cerner FHIR Developers
Hi Daya,

You should now get back the category of vital-signs for SpO2.

Thank you for your patience,
Ben (Cerner)
Reply all
Reply to author
Forward
0 new messages