Stream: Vulcan/Medication Definition
Topic: Connectathon 2021-01
Hugh Glover (Dec 08 2020 at 10:40):
This is for the January 2021 connectathon discussions
Andrei Idu (Jan 14 2021 at 15:36):
Hello and welcome to the Medication Definition track!
Rik Smithies (Jan 14 2021 at 16:21):
2021_01-Connectathon-Medication-Definition-Track.pdf
Rik Smithies (Jan 14 2021 at 16:22):
We will kick off the session in 9 minutes on the zoom, the slides are above, and we will have meet and greet after that
Andrei Idu (Jan 14 2021 at 17:19):
FHIR Server URL https://fhir-connectathon-server-ema.developer.azure-api.net/ please message me on here for credentials
Rik Smithies (Jan 14 2021 at 18:41):
Here is a google sheet for participants to log what that have done with which parts of the resources https://drive.google.com/file/d/1oDKQxomxc-789B_j4zDYE33V_1_NDo-D/view?usp=sharing
Jose Costa Teixeira (Jan 14 2021 at 20:03):
I'm trying to upload a PackagedProductDefinition, but I don't have one ready, so I started with an example from the build, but I have a few errors.
Andrei Idu (Jan 14 2021 at 20:04):
Which FHIR version are you getting the example from? The server supports 4.4.0 (R5 preview 2)
Jose Costa Teixeira (Jan 14 2021 at 20:04):
perhaps I shouldn't be looking at the current build - is there a URL is hould look for, to get the resource type definitions that are on the server?
Andrei Idu (Jan 14 2021 at 20:05):
Yes, please use https://hl7.org/fhir/2020May/
Jose Costa Teixeira (Jan 14 2021 at 20:06):
that is it, thanks
Jose Costa Teixeira (Jan 14 2021 at 20:15):
you mentioned the Everything operation - is it supposed to get a product and all related products?
Andrei Idu (Jan 14 2021 at 20:21):
It's supposed to get the medicinal product together with all other resources related to that medicinal product, not the cross referenced products
Andrei Idu (Jan 14 2021 at 20:21):
Try id 752b4bd5-8614-469a-8b8d-d5243a462a73
Jose Costa Teixeira (Jan 14 2021 at 20:33):
so the resources that point to X but not the resources that X points to, right?
Andrei Idu (Jan 14 2021 at 20:35):
I think that by the spec it should be both, however I've only implemented resources that point to Medicinal Product for now, as in it won't traverse the references within a Medicinal Product (yet)
Jose Costa Teixeira (Jan 14 2021 at 20:35):
btw, it contains duplicates, and I don't know why
Jose Costa Teixeira (Jan 14 2021 at 20:36):
Andrei Idu said:
I think that by the spec it should be both, however I've only implemented resources that point to Medicinal Product for now, as in it won't traverse the references within a Medicinal Product (yet)
ok. I am not worried about testing the server :) I just asked to see how the specification is intended - cross-references between products is a big question
Andrei Idu (Jan 14 2021 at 20:41):
ok, as far as I understand it, the spec does not mention including cross referenced products, $everything is the actual medicinal product and all other types of resources related either directly or via N-level references. I guess in the end it depends on your business context if you want to include other medicinal products
Andrei Idu (Jan 14 2021 at 20:42):
the duplicates are just the history of me testing out unfinished implementation :)
Jose Costa Teixeira (Jan 14 2021 at 21:58):
thanks
Andrei Idu (Jan 14 2021 at 22:29):
I just managed to submit a new product with all related resources, all seem to be reflected in the $everything response, ID is cdf1aed6-081a-4886-abaa-546192b997dc
Andrei Idu (Jan 14 2021 at 23:04):
I'm calling it a night for today. See you all tomorrow!
Jose Costa Teixeira (Jan 15 2021 at 10:21):
I'm trying to see what is the substance of behind a product. Basically I want to query "all products that contain amlodipine"
Jose Costa Teixeira (Jan 15 2021 at 10:22):
(or "all products that contain a dose of 5mg of amlodipine")
Jose Costa Teixeira (Jan 15 2021 at 10:23):
(minor question - European Union as a language?)
Andrei Idu (Jan 15 2021 at 10:38):
Jose Costa Teixeira said:
I'm trying to see what is the substance of behind a product. Basically I want to query "all products that contain amlodipine"
I think that is not supported by 4.4.0 spec, but it is a use case we also need for SPOR. I think the relationships should be there in the latest FHIR spec to be able to follow the path up from SubstanceDefinition/Ingredient to MedicinalProductDefinition
Andrei Idu (Jan 15 2021 at 10:40):
Jose Costa Teixeira said:
(minor question - European Union as a language?)
EU is the country code in the EU-en country-language combination for centrally authorized products, EU is a special case in the Countries list from RMS (referentials) in SPOR
Jose Costa Teixeira (Jan 15 2021 at 11:01):
yes, that was my understanding as well - so I was expecting the language display to be "EU English"
Andrei Idu (Jan 15 2021 at 11:18):
Which product and field are you looking at specifically?
Jose Costa Teixeira (Jan 15 2021 at 11:22):
the one you created , cdf1aed6-081a-4886-abaa-546192b997dc
Jose Costa Teixeira (Jan 15 2021 at 11:22):
aaaaaaaaaaahh
sorry
Jose Costa Teixeira (Jan 15 2021 at 11:22):
it is country and then language
Jose Costa Teixeira (Jan 15 2021 at 11:22):
So it is correct
Andrei Idu (Jan 15 2021 at 11:23):
indeed :)
Jose Costa Teixeira (Jan 15 2021 at 11:25):
this is a model question then. Normally we use a codeableconcept for language
Jose Costa Teixeira (Jan 15 2021 at 11:25):
but the MPD resource does it differently, it has a special construct
(a normal example of language is here: https://www.hl7.org/fhir/practitioner.html -->.communication)
Jose Costa Teixeira (Jan 15 2021 at 11:26):
so the example is according to the spec.
I'll keep this as feedback on the spec itself.
Andrei Idu (Jan 15 2021 at 11:30):
language is Codeableconcept in MedProdDef.countryLanguage as well, it is referencing an RMS list in the example I made due to EMA requirements, I think the spec allows for the same behaviour as Practitioner.communication as far as I can tell
Andrei Idu (Jan 15 2021 at 11:31):
off to lunch for 1h
Jose Costa Teixeira (Jan 15 2021 at 12:19):
yes, you are correct. I just notice a complex element, and I have questions like
"countryLanguage.country - is it the product country or the language country? " (same with jurisdiction)
In other words, isn't https://tools.ietf.org/html/bcp47 sufficient?
Jose Costa Teixeira (Jan 15 2021 at 12:20):
another question, @Andrei Idu - you use "ibesartan" as a classification?
Andrei Idu (Jan 15 2021 at 12:49):
Jose Costa Teixeira said:
another question, Andrei Idu - you use "ibesartan" as a classification?
yes, as per ATC classification system
Andrei Idu (Jan 15 2021 at 12:54):
I think bcp47 allows for language code only, while specifying country and language as separate fields ensures accuracy
the countryLanguage field is related to the Name of the product, so the country language combination which applies to the text in the parent name entry... you could have multiple translations of a name. So, the country is definitely not of the product
Jose Costa Teixeira (Jan 15 2021 at 12:56):
ok so 100000093533 is the EMA code system that reflects ATC somehow, and
https://spor.ema.europa.eu/v1/lists/100000093533#100000095144 corresponds to ATC#C09CA04
Jose Costa Teixeira (Jan 15 2021 at 12:56):
right?
Jose Costa Teixeira (Jan 15 2021 at 12:59):
our language codes allow things like nl-BE, fr-BE and de-BE. We even have a standard extension for all strings, so that does not seem the only requirement
Jose Costa Teixeira (Jan 15 2021 at 13:01):
we could easily have a product.name.language, instead of 3 separate elements - I'm keeping that as possible feedback.
(this not critical, just trying to see if there are some wrinkles that can be ironed out)
Andrei Idu (Jan 15 2021 at 13:04):
yes that's right about the ATC codes
Andrei Idu (Jan 15 2021 at 13:11):
about languages, bcp47 allows just "de" German, i think a separate country field ensures a country is specified. Also an Internet standard might not be sufficient for wider domain (real world) use cases
Jose Costa Teixeira (Jan 15 2021 at 16:00):
Andrei Idu said:
about languages, bcp47 allows just "de" German, i think a separate country field ensures a country is specified.
I believe these are valid:
sl-IT-nedis (the specific variant of the Nadiza dialect of Slovenian that is spoken in Italy)
de-CH-1901 (the variant of German orthography dating from the 1901 reforms, as seen in Switzerland)
Also an Internet standard might not be sufficient for wider domain (real world) use cases
I don't understand this - we have other real world use cases in FHIR that use this.
Andrei Idu (Jan 15 2021 at 17:09):
I hope there are no documented product names in those specific languages :)
Andrei Idu (Jan 15 2021 at 23:44):
Great to have met everyone! Server is up until tomorrow afternoon. Have a nice weekend!
Last updated: Apr 12 2022 at 19:14 UTC
