Stream: patient administration WG
Topic: Follow-up on #17519
Christine D (Jul 27 2018 at 15:21):
Checking in on a request for the Encounter resource. For Clinical Research, the Encounter is the 'visit'. These typically have a number (sequence) which we can map to the ID, but they may also have a name (e.g., screening visit) that we capture. We were wondering about an extension to Encounter, that would allow us to capture a name for the Encounter. What are you thoughts on this? How should we proceed with the tracker entry?
https://gforge.hl7.org/gf/project/fhir/tracker/?action=TrackerItemEdit&tracker_item_id=17519
Brian Postlethwaite (Jul 29 2018 at 11:27):
The values that you've suggested in the tracker look more like descriptions for coded values.
There is an appointmentType on the appointment that like is the same type of concept that you're asking for (it's expected to be a user defined valueset).
If this was included on Encounter would it cover your use case also?
Christine D (Aug 01 2018 at 14:21):
Hi @Brian Postlethwaite , it sounds similar, but I am not sure how we would accommodate encounters that are just part of the regular visits (e.g. visit 1, visit 2, visit 3) and need to be given a name. They aren't really coded per se. But I do see how there is some synergy with the appointmentType in that some visits will have a purpose that is specialized.
Brian Postlethwaite (Aug 01 2018 at 21:32):
Using the CodeableConcept type (which is what we would use) you have the ability to just use the text property without a coding, giving you the best of both worlds
Christine D (Aug 02 2018 at 17:52):
That could work. Will the tracker be updated with the new proposal?
Christine D (Aug 08 2018 at 14:50):
@Brian Postlethwaite - this is one we didn't discuss in the PA meeting this week. Will the 'encounterType' proposal move forward? (your suggestion to implement something similar to appointmentType) Is there any chance that would be ready for the Connectathon?
Cooper Thompson (Jan 16 2019 at 16:08):
@Christine D what systems are you communicating this Encounter name to? Most systems I'm aware of track discrete values for encounter type and class, or appointment type for this sort of information. I'm curious who (in the interoperability space) would actually be interested in a free-text Encounter name.
Brian Postlethwaite (Jan 16 2019 at 16:09):
Did you consider using the Encounter.type property in the way described above?
Christine D (Feb 08 2019 at 14:19):
@Cooper Thompson apologies for the delay! this would be in clinical research, as the site is communicating study data (including the encounter information) back to the study sponsor. a number or ID may not be sufficient detail to understand the context of the encounter.
Christine D (Feb 08 2019 at 14:20):
@Brian Postlethwaite based on the definition, encounter.type didn't quite seem to fit.
Cooper Thompson (May 08 2019 at 18:42):
@Christine D Is this name something that you expect an end user (e.g. registrar) to manually type in when they are creating the encounter? Or is the value from a fixed list of allowed values? I'm not aware of any workflow where end users are entering an encounter name when they create an encounter. All implementations I'm aware of have a pre-defined list of valid "names" (we call them "types").
Cooper Thompson (May 08 2019 at 18:43):
You could generate a more "customized" name by combining the patient name and the Encounter.type to produce something like "Well Child visit for Wesley Crusher". Since that is a value composed of existing data, we wouldn't want to add it as a new property.
Christine D (May 08 2019 at 18:46):
Hi @Cooper Thompson , we would expect this to be something typed in. There isn't necessarily a fixed list of values, since the study team may call an encounter (these are "visits" in clinical trials) "Pre-screening" or they may call it "Visit 9". We need a bit of flexibility for the study teams to determine what they would use. We would envision the clinical trial site typing in this data.
Cooper Thompson (May 08 2019 at 18:47):
Are you here at the WGM? Any chance you could join PA in Salon Jarry? We're discussing this right now.
Christine D (May 08 2019 at 18:48):
I am not. :( I had an internal meeting and could not attend. I am hoping to be in Atlanta.
Cooper Thompson (May 08 2019 at 18:49):
Is your clinical trial staff re-entering the same values for every set of patients? I.e. if you have 100 patients enrolled in you study, will all 100 of those patients have both a pre-screening and a visit 9?
Christine D (May 08 2019 at 18:49):
Yes
Cooper Thompson (May 08 2019 at 18:50):
If you have staff entering in the same thing over and over again, wouldn't it be better to build a fixed list of types? For example:
Cooper Thompson (May 08 2019 at 18:50):
- Pre-screening
- Visit 1
- Visit 2
- Follow up 1
- Follow up 2
Christine D (May 08 2019 at 18:50):
With one caveat. There could be a patient who requires and additional visit for something like an additional lab (e.g., something was out of range).
Cooper Thompson (May 08 2019 at 18:51):
Then you'd have an optional visit type of "lab follow up" that would only apply to the subset of patients that needed that.
Christine D (May 08 2019 at 18:51):
It could be possible to build a list. We were just a bit worried about the variance between studies. (sometimes they have novel treatments)
Christine D (May 08 2019 at 18:52):
And we could have hundreds of visits. Two year study with weekly visits = very long pick list. :)
Cooper Thompson (May 08 2019 at 18:52):
Do you have a full example set of visit types you'd have for studies? I'd expect you'd end up with a list of pretty generic types.
Christine D (May 08 2019 at 18:53):
We don't have a full example set because they do vary by therapeutic area, study length, etc.
Christine D (May 08 2019 at 18:53):
oncology may have different visit types than neuroscience
Cooper Thompson (May 08 2019 at 18:55):
Is the visit type really just "research visit" in an oncology department, where the actual differentiated is the treatment or protocol that is performed in that encounter?
Cooper Thompson (May 08 2019 at 18:58):
As a point of reference, we (Epic) have a "Research Encounter" type. That encounter would take place in a specific department, and be associated with a particular study. The encounter would consist of whatever treatments or consultations were defined by the study.
Cooper Thompson (May 08 2019 at 18:59):
If a patient were admitted for reasons unrelated to the research study, you could also administer treatment related to the study as part of that admission (which would have a different encounter type all together).
Christine D (May 08 2019 at 19:06):
I think it may be a similar concept. In the trial, the visit (encounter) is associated with the study and is related to the protocol-specified treatments / procedures / etc. that need to take place during that visit. They typically give that visit a name. Sometimes it is just Visit X, but sometimes it describes the scope of what is being performed during the visit.
Christine D (May 08 2019 at 19:29):
@Cooper Thompson I just saw that the Change Request had been rejected and the suggestion was to add an extension. The CR was for an extension. We were not asking to modify the code resource.
Christine D (May 08 2019 at 19:30):
*core resource
Cooper Thompson (May 08 2019 at 19:37):
You can add your own local extension for now. We were still not sure we understood what the best data model was for this type of information, so we weren't ready to add a standard extension.
Christine D (May 08 2019 at 19:39):
Ah, ok. Thanks!
Last updated: Apr 12 2022 at 19:14 UTC
