Stream: committers
Topic: Extensions in multiple contexts
Bryn Rhodes (Jul 14 2016 at 22:10):
So if I have an extension that I want to be able to use in both the Resource and DataType contexts, I need to define different extensions, right? Is there a convention for how to name those extensions?
Grahame Grieve (Jul 14 2016 at 22:11):
what context do you want to use it in?
Bryn Rhodes (Jul 14 2016 at 22:11):
For example, I want to define a qualityOfEvidence that can be used in both Resource and DataType contexts.
Bryn Rhodes (Jul 14 2016 at 22:11):
qualityOfEvidence
Grahame Grieve (Jul 14 2016 at 22:11):
which specific paths?
Bryn Rhodes (Jul 14 2016 at 22:12):
DocumentReference, root and Attachment, root.
Grahame Grieve (Jul 14 2016 at 22:24):
it feels like there should be a way to do that
Grahame Grieve (Jul 14 2016 at 22:24):
would you want it on extensions that are attachments?
Bryn Rhodes (Jul 14 2016 at 22:25):
Yes, I could definitely see that. The idea being that an attachment can be made into "supporting evidence" by adding this indicator to describe the quality of the evidence.
Grahame Grieve (Jul 14 2016 at 22:27):
given that DocumentReference must refer to things by Attachment, why not define it just for Attachment/
Bryn Rhodes (Jul 14 2016 at 22:28):
That would work for this case, yes.
Bryn Rhodes (Jul 14 2016 at 22:29):
I have another one that raises the same question though.
Bryn Rhodes (Jul 14 2016 at 22:29):
I'll do that for the qualityOfEvidence though.
Grahame Grieve (Jul 14 2016 at 22:30):
I'll look forward to the next one...
Bryn Rhodes (Jul 14 2016 at 22:31):
:)
Bryn Rhodes (Jul 14 2016 at 22:31):
I want to be able to define an expression extension that can be applied anywhere.
Bryn Rhodes (Jul 14 2016 at 22:31):
So I created three different extensions (actually six, three for fluentPath, three for CQL).
Bryn Rhodes (Jul 14 2016 at 22:32):
And named them based on the context in which they could be used.
Bryn Rhodes (Jul 14 2016 at 22:32):
fpExpression, fpDataTypeExpression, fpExtensionExpression, and similarly for CQL
Grahame Grieve (Jul 14 2016 at 22:32):
Anywhere? so, lik,e just on Element?
Grahame Grieve (Jul 14 2016 at 22:33):
and Resource?
Bryn Rhodes (Jul 14 2016 at 22:34):
On the root of a resource is a little weird, but on any element within a resource, yes.
Bryn Rhodes (Jul 14 2016 at 22:35):
And even at the root of a resource, actually. The interpretation being the expression defines the entire resource. (I actually do that in some of the order set and decision support actions).
Grahame Grieve (Jul 14 2016 at 22:35):
well, then ContextType= DataType, Context = Element,Resource
Bryn Rhodes (Jul 14 2016 at 22:36):
Clever.
Bryn Rhodes (Jul 14 2016 at 22:36):
And that will work for extensions too?
Grahame Grieve (Jul 14 2016 at 22:36):
sure. they are element
Bryn Rhodes (Jul 14 2016 at 22:37):
Perfect, thanks!
John Moehrke (Jul 15 2016 at 14:20):
@Bryn Rhodes to your specific qualityOfEvidence... this seems like an element that is duplicating the "Integrity" meta tags that you can use through the meta element using the HCS securiity_tags https://www.hl7.org/fhir/v3/SecurityIntegrityObservationValue/index.html
Bryn Rhodes (Jul 15 2016 at 17:47):
The quality of evidence extension in CQF is specifically about describing the relative quality of the supporting evidence for a particular knowledge asset. Looking through that list, I don't see anything that applies in that space, but maybe I'm missing something?
John Moehrke (Jul 15 2016 at 18:05):
Hi @Bryn Rhodes so these codes don't apply? Then what kind of codes are you suggesting? HRELIABLE highly reliable
RELIABLE reliable
UNCERTREL uncertain reliability
UNRELIABLE unreliable
CLINAST clinician asserted
DEVAST device asserted
HCPAST healthcare professional asserted
PACQAST patient acquaintance asserted
PATAST patient asserted
PAYAST payer asserted
PROAST professional asserted
SDMAST substitute decision maker asserted
CLINRPT clinician reported
DEVRPT device reported
HCPRPT healthcare professional reported
PACQRPT patient acquaintance reported
PATRPT patient reported
PAYRPT payer reported
PRORPT professional reported
SDMRPT substitute decision maker reported
Bryn Rhodes (Jul 15 2016 at 18:24):
Those codes seem to be about specific data such as observations or medical record content. What would it mean, for example, to say that the patient asserted that the content of this decision support module is accurate?
John Moehrke (Jul 15 2016 at 19:04):
I don't understand... they are codes... they convey meaning... all codes can be used wrongly. You can't build into a code system a forbiddance from being abused.
Grahame Grieve (Jul 15 2016 at 20:55):
I think you're talking past each other. Bryn, waht codes do you have for quality? And John, most people would never think of looking in 'security' for this...
Bryn Rhodes (Jul 15 2016 at 20:57):
I'm tracking those down as part of building the example bindings for the extensions, but they are typically a rating system used to grade the evidence (and another for strength of recommendation).
Last updated: Apr 12 2022 at 19:14 UTC
