Stream: committers
Topic: Consent.status
Grahame Grieve (Nov 29 2018 at 02:46):
@John Moehrke I do not understand why Consent.status is bound to EventStatus. how is a Consent an event? What does 'on-hold' or 'stopped' mean? How is a consent 'currently occurring'?
John Moehrke (Nov 29 2018 at 14:23):
I presume it is bound to the broader valueset as they had little experience with this concept to limit the valueset. There is a strong need for experience with realworld consent. So this STU phase is very important to get engagement... Is this right @David Pyke ?
John Moehrke (Nov 29 2018 at 14:27):
I think the valueset is also large due to the scope of Consent, it would be more controlled if consent was just Privacy.
John Moehrke (Nov 29 2018 at 14:27):
I think there is a vision that some 'profiles' of the various types of consent would be created. within that would be a smaller set with defined use.
David Pyke (Nov 29 2018 at 18:04):
It's bound to the Event workflow, that was given to us when the workflow alignments happened. I"m happy to change it to a better suited workflow and resulting codesets.
John Moehrke (Nov 29 2018 at 20:21):
ah, so the mapping to Event didn't consider creating a subset of that vocabulary. That is an opportunity for R5.
Grahame Grieve (Nov 29 2018 at 20:23):
Consent is not an event, and the vocabulary doesn't make sense
John Moehrke (Nov 29 2018 at 20:24):
ah, why is it not an Event? It records the results of the patient giving consent for some activity. Isn't that what the Event pattern is about?
Grahame Grieve (Nov 29 2018 at 20:24):
it was complicated much further in that the short description still contained the old codes. Caused me some confusion for a while there, but I've corrected it.
Grahame Grieve (Nov 29 2018 at 20:25):
There is an event, the patient creating the resource. But the resource is not a statement about the event, it's a statement about the state of things going forward .
Grahame Grieve (Nov 29 2018 at 20:25):
if you want to track the making of the statement, then it has appropriate codes. But do we really care about the fact that a patient has started working on their statement, and then that they have finished?
Grahame Grieve (Nov 29 2018 at 20:25):
I don't think we do
John Moehrke (Nov 29 2018 at 20:26):
hmmm. I see your point on 'things going forward', but I could say that about anything
Grahame Grieve (Nov 29 2018 at 20:26):
not so. The obvious equivalent resource is AllergyIntolerance, actually
Grahame Grieve (Nov 29 2018 at 20:26):
it's a statement about things going forward.
Grahame Grieve (Nov 29 2018 at 20:27):
where as Observation is a statement about a point in time
John Moehrke (Nov 29 2018 at 20:27):
if you want to track the making of the statement, then it has appropriate codes. But do we really care about the fact that a patient has started working on their statement, and then that they have finished?
I don't think it is trying to track the event process. Just the result.... which my understanding of Event pattern is that this is what Event Pattern is about... Event Pattern is not about tracking the progress of the event, but recording that the Event happened.
John Moehrke (Nov 29 2018 at 20:28):
on the tracking vs documenting... I would agree that the Consent resource is trying to be too-many things.
Grahame Grieve (Nov 29 2018 at 20:28):
right. but Consent is not about recording that the event of consenting happened, but about what the agreement made is going forward
John Moehrke (Nov 29 2018 at 20:28):
okay... so then what is your recommendation?
John Moehrke (Nov 29 2018 at 20:28):
I am not clear on why that matters
John Moehrke (Nov 29 2018 at 20:30):
right. but Consent is not about recording that the event of consenting happened, but about what the agreement made is going forward
I think Consent resource is intending to be both... it is capturing that the consent happened, and it is encoding the rules that will be used by some decision engine. It is NOT the decision engine.
Grahame Grieve (Nov 29 2018 at 20:30):
well, I stumbled into this when doing R3/R4 maps. Here's the status codes for the R3 consent:
Grahame Grieve (Nov 29 2018 at 20:31):
draft | proposed | active | rejected | inactive | entered-in-error
Grahame Grieve (Nov 29 2018 at 20:31):
that list totally makes sense to me
Grahame Grieve (Nov 29 2018 at 20:31):
here's the list of R4 consent status codes:
John Moehrke (Nov 29 2018 at 20:31):
is there a requirement that there must be logical progression from R3 to R4? I thought that under STU we are free to break things
Grahame Grieve (Nov 29 2018 at 20:31):
preparation | in-progress | not-done | on-hold | stopped | completed | entered-in-error | unknown
Grahame Grieve (Nov 29 2018 at 20:32):
I don't know how to map those because they're different lists, and because they're about different things.
Grahame Grieve (Nov 29 2018 at 20:32):
no, you don't have to.
Grahame Grieve (Nov 29 2018 at 20:32):
but in this case: what's the status for saying:
Grahame Grieve (Nov 29 2018 at 20:32):
'I don't agree with this consent any more'?
John Moehrke (Nov 29 2018 at 20:32):
ah. I do like the STU3 vocabulary... as @David Pyke mentioned this might have been an overreaction at the mapping to Event pattern
Grahame Grieve (Nov 29 2018 at 20:33):
it wouldn't be the first. But in this case, Consent is not an event
Grahame Grieve (Nov 29 2018 at 20:33):
The committee can choose to change their minds on this.
John Moehrke (Nov 29 2018 at 20:34):
I think it is Lloyd that dictated it... the committee likely had no idea how to evaluate that dictate or what alternatives there are
Grahame Grieve (Nov 29 2018 at 20:34):
we could revert to the R3 mapping. Or we could add a second status for the agreement, as opposed to the event. Or we could add documentation about what is going on. But I don't think we can do nothing
Grahame Grieve (Nov 29 2018 at 20:35):
well, @Lloyd McKenzie is not here for 10 days, so we can't ask him for an opinion
John Moehrke (Nov 29 2018 at 20:41):
so. lets first figure out this question of if it is an Event or not an Event... I am still struggling with why it is not an Event.
John Moehrke (Nov 29 2018 at 20:42):
Many organizations will use Consent purely to document the consent ceremony. They will not use the .provisions at all
John Moehrke (Nov 29 2018 at 20:42):
much like the discussion ongoing on mapping between BPPC and Consent.
John Moehrke (Nov 29 2018 at 20:43):
that is clearly an Event pattern... right?
Grahame Grieve (Nov 29 2018 at 20:43):
yes that is
John Moehrke (Nov 29 2018 at 20:43):
that is mostly why we are an Event pattern.
Grahame Grieve (Nov 29 2018 at 20:44):
so what if a patient later says 'I don't agree with that statement anymore'. I believe that this is (and should be) something we support.
John Moehrke (Nov 29 2018 at 20:46):
yes, that would be either a revision of the Consent resource... with the new version expressing the new ceremony. Or it would be a new Consent resource and a deprecation of previous... depending on the organizatinns management style
John Moehrke (Nov 29 2018 at 20:46):
hence, why I agree with you on the old vocabulary being more useful.
John Moehrke (Nov 29 2018 at 20:47):
which, Dave seemed to indicate they didn't know they could specalize the .status vocabulary
Grahame Grieve (Nov 29 2018 at 20:48):
seems to me that we have 2 status codes - one for tracking the process of the consent agreement, and one for tracking the validity of the agreement itself
Grahame Grieve (Nov 29 2018 at 20:49):
the record of the event remains valid irrespective of later decisions, but the agreement itself is subject to later decision making.
John Moehrke (Nov 29 2018 at 20:50):
I think validity is all that is needed.
John Moehrke (Nov 29 2018 at 20:51):
Tracking process is not typically needed. At most there is a need for a way for a patient to propose their terms, so that an organization can respond with a consent terms best mattching those terms.
John Moehrke (Nov 29 2018 at 20:51):
There is a timeframe to a consent. It is not valid automatically after the sunset timeframe.
John Moehrke (Nov 29 2018 at 20:51):
It can be terminated early...
John Moehrke (Nov 29 2018 at 20:52):
so there is already different vectors for validity period. There just needs to be a .status method for deprecated.
Grahame Grieve (Nov 29 2018 at 20:54):
ok, well, when does the committee next meet?
John Moehrke (Nov 29 2018 at 20:56):
tuesday... right @David Pyke ?
David Pyke (Nov 29 2018 at 21:02):
Right, and I'd be happy to revert the status. As John said, we were given the Event pattern to follow and got errors/warnings if we didn't make sure all parts were aligned. Tuesday at Noon ET is our weekly meeting. If someone can tell me how to revert/change back without triggering build warnings, I'd appreciate it.
Grahame Grieve (Nov 29 2018 at 21:05):
I'll have a look.
John Moehrke (Nov 29 2018 at 21:09):
dave, see source/suppress-workflow-warnings.xml
Grahame Grieve (Nov 29 2018 at 21:42):
so all the examples had a status of 'in-progress' - which would mean that they were not actually completed. But all of them appeared to be completed. So I do believe that the change is required, to the point where I will commit the change, though I will have to reverse it if the committee does not endorse the change
John Moehrke (Nov 29 2018 at 22:41):
the status of in-progress is not logical. You either have consent, or don't... As in STU3 where the status element was 'active'
Grahame Grieve (Nov 29 2018 at 22:52):
well, reverting to R3 codes should be up on build.fhir.org in about 30min
John Moehrke (Dec 04 2018 at 17:07):
GF#19718 approved
Grahame Grieve (Dec 04 2018 at 18:13):
thx
Lloyd McKenzie (Dec 10 2018 at 22:31):
Consent is definitely an Event. It's something that happened - not something that was asked for or "could" happen.
Lloyd McKenzie (Dec 10 2018 at 22:33):
That said, 'suspended' and 'stopped' make no sense here, so a constrained value set is certainly appropriate. Consent is an "instantanious" event.
David Pyke (Mar 22 2022 at 17:27):
(deleted)
Last updated: Apr 12 2022 at 19:14 UTC
