CGIT WG Minutes Baltimore, Maryland (09/2016)

1

HL7 Work Group Meeting Madrid Spain– May 2017

Attendees: The list is provided in the Appendix.

Monday Q2 Chair: Nathan Bunker Scribe: Rob Snelick Attendees: 6 Quorum: Yes

Reviewed and approved agenda. Rob, Alex 2-0-0

Reviewed and edited project list. Intend to create PSS for standalone conformance chapter for HL7 v2. Question is, do we coordinate this with the publishing of HL7 v2+?

Monday Q3 Chair: Rob Snelick Scribe: Nathan Bunker Attendees: 4 Quorum: Yes

Prepare for FHIR meeting on Wednesday Discussion in FHIR-I about whether to represent version at the resource level or to indicate this in the profile. Background: they want to have information about compatibility. Compatibility can be opposite for both sides of the same connection.

Basic challenges. The version of FHIR resources is needed to parse the resource correctly but if it’s in the resource then you don’t know it.

One idea was posed that normative will not allow changes. But this is impossible because there will always need to have changes.

Tooling in Java (and other languages) is locked into a version. The services for a base URL will always be specific to a FHIR version. But in practice you will get more than one version on different URLs. Even in the best of scenarios there are at least two versions in production, but in practice a service can have many versions. Once an interface is put in place it is often not upgraded for a long time.

The question is, what happens to the underlying resources. The answer depends on how they are handled. If FHIR is a façade, then adaptions can be expected. But it FHIR is used to store the resources then no changes will apply.

No simple solution, Restful services have no magic solution to this.

Notes continue here on this wiki page: http://wiki.hl7.org/index.php?title=Conformance_Position_on_FHIR_Versions

CGIT WG Minutes Baltimore, Maryland (09/2016)

2

Monday Q4 Chair: Rob Snelick Scribe: Nathan Bunker Attendees: 4 Quorum: Yes

Continuing Q3.

Implementation Guide:

Task for us to evaluate each relevant FHIR construct to see if it is sufficient or not. This has been our role to look at this but we have not had the time to do this so far. Put each aspect on the bi-weekly calls so we can eventually cover all the areas. Can we get a volunteer with FHIR?

Take slides from tutorial and work on these and review those. Create a power point with recommendations. Use Slide 65 of the NISTConformanceTutorial as the template for discussions. http://hl7v2tools.nist.gov/portal/#/publications

Conformance Resource and Implementation Guide Resource in STU 2 appears to be some overlap. Haven’t looked at STU 3. Conformance Resource is now called a Capability Statement Resource in STU 3.

Prepare for meeting with HL7 & OO & AID for implementation package and validation of HL7v2 Closely related to the future of HL7 v2. Rob has slides prepared for the meeting. Talk about the use of IGAMT for validating HL7 v2.

Conformance length in HL7 v2 History of this for V2 would be good. Need a place for V2 frequently asked questions on conformance.

With min and max length you can only shrink. You can make your min more and your max less and still be conformant. Otherwise not conformant. Add to the new Conformance chapter that is being pulled out.

TODO: Make sure to add this to the list of things we need for PSS.

Cochairs Dinner/Meeting TODO: By August 1st we need to review our mission/charter.

Wednesday Q3 Chair: Frank Oemig Scribe: Nathan Bunker Joining: FHIR, Templates Attendees: 15 Quorum: Yes

Versioning FHIR Discussion The discussion began by talking about version can be handled with FHIR. There are different aspects depending on how the specification is used. If FHIR is used as a façade adaptations of the data is necessary. If resources are stored directly, this will be done according to a specific version.

CGIT WG Minutes Baltimore, Maryland (09/2016)

3

From a REST perspective, a single endpoint must be the same version, ie. different versions are handled by different base URLs. (On Saturday the discussion was about data at REST.)

But the issue is still important for communication channels. How do you tell which version the channel is in? Also who does the translation?

The issue is what is behind the endpoint? Usually get data to use it. Need to know what the version is whether it’s transmitted or whether it’s added to the data as it comes in.

Looking back in time, how do you know which version a resource was encoded under?

The version number has been requested to be explicit because the profile is not required to be indicated so it may not be possible to understand if the resource is valid or not.

The decision is not going to be decided here or even this week.

Looking at this document: https://docs.google.com/document/d/1mpbLwGIhoWMfSH_TI368mgQrQOyMwkxPJZdtyKpDCiA/edit

Recommendation that we can only have one version for the whole resource and not allow the content not to be another version.

Normally we have Grahame Grieve write a blog post on it to recommend how to continue the discussion.

More in favor or abstain from adding version: Nearly all in favor, nobody opposed. Most people think there should be a version someplace.

Straw Polls, which options are acceptable:

1. Mime Type: 1 2. Extensions: 1 3. Profile declaration with non-canonical URL: 2 4. Profile declaration with canonical URL + version: 4 5. Tag: 3 6. New element for meta: Nearly all

FHIR QA Talking about FHIR maturity model. A measure to the outside world about how stable our models are. 0 is draft. 5 means people have been using this for a while and have wide participation, not expecting changes. Each of these levels requires you to fill out a QA Tracker sheet. Looking at this web page: http://wiki.hl7.org/index.php?title=DSTU_QA_Guidelines

Conformance needs to evaluate the FHIR Resource QA Tracker: https://docs.google.com/spreadsheets/d/18HfXF7mUCUV7jACCG0oejFp6D-ibtvbmcgywNhn76lw/edit#gid=0 Conformance needs to look to see if anything needs to be added.

TODO: Add this project to our list of items to cover in our weekly meetings.

The goal today is just to see what this is about so we can address it.

CGIT WG Minutes Baltimore, Maryland (09/2016)

4

Templates: Want to talk about formal method of converting profiles to templates and back There are gaps in the process of converting CDA templates. Need to add some things to the structure definition or the template definition to do this. What is the benefit of converting a profile definition to a template definition. Maybe it doesn’t have to be perfect, maybe it just needs to get the author close to the right solution. Rick (?) is doing a project to convert CCDA to FHIR implementation guide. The thought is that we will have a manual mapping process forever.

HL7 v2 has a similar issue and it also takes insight to be able to convert between. You have to know what you are doing.

Not talking about the specific use cases being translated. Talking about expression of the template language. There was a question about the profile definition language whether it has been balloted. It has balloted in every version, but the latest version has not yet been balloted. Can we have a single form to express our constraints. Now we have three. Templates ITS, HL7 v2 and FHIR. The short answer is no.

Might be able to make a more user friendly syntax that we could use in both FHIR and HL7 v3. That might be possible.

Managing profiles is going to be a challenge in the future.

Structure Definition in FHIR has two responsibilities to describe the structure and then to validate.

More discussion about how to solve the problem between Templates and Structure Definitions. Templates chose XPath.

In HL7 v2 we wrote our own language for constraints. Can do co-constraints and jump from one segment to another. Can also handle OBX dynamic data type. NIST has written software in SCALA. We can do tooling on the front end and then automatically generate those constraints for the validation engine. About 2-4% needs to jump out to the executable which we call plug-ins. It can be found here: http://hl7v2tools.nist.gov/portal/#/

FHIR Conformance Review Propose to move the meeting to 10am ET.

Wednesday Q4 Chair: Nathan Bunker Scribe: Rob Snelick Attendees: 3 Quorum: Yes

Wrap-up and Admin Agenda is blocked out for next WGM.

Question from PHER: Conformance statement versus small value set What is conformance’s official recommendation about documenting conformance for a small number of values.

Answer: Neither is wrong.

Looking for best practice.

CGIT WG Minutes Baltimore, Maryland (09/2016)

5

Single code: use patterned constraint conformance statement

Hard to argue one or the other. Both are precise. What is the preferred? Perhaps have an explicit value set for two or more. Not too worried about this, it’s great that the author is getting this precise. If you can use constructs that are generic then you want to use them, so using the value set makes more sense.

Thursday Meeting with Vocabulary. See Vocabulary notes.

Attendance Sheet

TBD

z> CD C) CD CD C CD 000 £ '< j3CD D)

CD 3Cfl .Q).—..< 5 CD CD Q 00 D 0 CD CD

.-,.'<..Ø..0.CD.(D n 3 CD

0 CD CD o

CD CDCNCD_CD - Hn

aCD g ,3 =

CD 2D'

rf<= 0'< Q< FD' : :: cTTh:': :: :

a.

CD ) Q) -' =-

I Cl- :

.ni: I:cn:<:z: :z: m:o:si ::n Q: m :o: :z: : r : :m: : : 0: z: :- :ci z:> o: 0 DO_= ::; G)0c,)CDm3

•i2• > (1)0 .cZ.0.fl). c) CCDCD ° oc

'

2.

. - ,=, =- 0. —.

.: :o: : : :: : =: :: : O.< cn: : : : :E: : •

:': :': : ::

: :8: : :

0 0

• • • •, 3. •(,, . • CD..-..

I :cn - 3:CF 0 CD M. - CD -UPWt w 3 =r T 0* CD

fl u 3 P fi w 0. , 50

(D 0~

(D3 3 3 00 0 5

1PI 5D

qD0H t& 8 itt 8 8 8

:o 2 a :3 0

.U) o

o0 . Y) 3:

:o 0 ., 3

o 0

: 0 0

:2 : :0 :2,

0 0

• • . • • CD

cn CD

: : > • —I _s

• : • : -1 F-..)

• -1 ()

JLL -

i

—I • : : : : : =

F'..) —I I

: ' H =

• .

0 —h 0

n) C) CD

0 C a-

0 CD

0 1

3 CD B CD

n) 1-

0

CD (I)

(0

0

CD

a- n)

C) CD

(I)

CD CD

Ca

:uF - u flfl u1u cu

L!i:: I 1 1 _

11111111111 ; 1111 I.

III :11,1111 11

. 1

is ~c

'/s ue

• W ~i qy ~-' <:0Io'C' 01 3:

O3

a .3 111,11.1 111111. I I. I I 1.11.111111 - 0

I

0

I I I I I I I I I I I I I I I I I I I . . ~ I . . I • I

L..J __ J __ .•...1. _ J _.J......L..L..J..J.._ •___ 1 .. 1 ..J..J...L._L_.J..J._J...V..L__J..J...L__L..J.. J__J...~ L__J__J...•___L__a__J I I I I I I I I I I I I I I I I I I I I I . ' I • I I I I I I I I I I I ~

I I I I • I I I I I I I I I I I I I I I I I I I I I I I I I I I I

I I I I • I I I I I I I I I I . I I I I I I I . .•__K...y __w__•__

i

X

III_r _•r••+__+.. .; ..r__r..+..}_..; ..r__•_. +.. .,. __r __r .. +__+__•; •_r__r .. .. .__+__.; .. r ._ r __ + _. } ... r .+..+

I I I I I I I I I I I I I I I I I I I I

1 • I I I I I I 1 I 1 . I 1 I I 1 I ' 1 I 1 I I 1 I I I I I I I I I I I

I I I I I I

1 I 1 I I I I I 1 I I I I

I I .

• I I I I I I

I I I 1 I I I 1 I I 1 I I

I I I 1 I

I I I I I I I I I I I I I I

I I I I

I I I I I I I r .. +.. ~ .._r ..r..+__+.. .; ..r__r_•+•_}__•r ••r__r.. + . __,___r _•r••+__ti.. .; __ r _. r .. + ._ } _..r .. r ._ ~ _. i.. ~ . __ ~ .. r .. +.. }___;. ..r .. •__+

111111111 .::::::::::::1.1.111111, 111111

: 1 1• I

111111111111111.111111'111111111111.1 1

1.111111111,111111111 111.111111111111 1111111,1111111111,

L--J--J---L--L--1--J--J---L--L--J--J---•---L--J--J--J---L--L--1--J--=---L--L--J--J---L--L--•--J--=---L--L--J--J---'---L--J--J

,N • -- • -- . --- t -- • -- • -- ' -- -'---t--•--•------^- '--•--•- K • - t I I I I I I I ~ I I I I I I I I " ~ `---`---•--•--" 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 IX

I~ I 1 1 1 1 1 1 I I I I

r••+_'}___r--r-_+_'+.. .; __r __r .. +__ r ..; ..r__+.. +.. .,__' r .. r .. + _' + .. .; r__r.. i.. r._r .. ~ .. ~ -- ~ _. ~. .. ~ _' ~ __i...~ .._r __ ~ ..r.. ~ W

I I I I I I • I I ~ I I I I I I I I I I I I I : : ' I I I I I . I I I I I

I I I I I I ~

I I 1 I I I I I

I I I 1

I I I I I

I J

r. -4...r -- • - •4•• - r••+_ .... ._ .. 1 .. y._ -+. . r . --,... r . ..... . r .. + .. 4 ._ ........4 W2 1 I t 1 I I I I I I . 1 I • 1 I I 1 I I I . '

I 1 I

• _ _ _ _ _ _ _ I I I I I I 1 I I I I I 1 I I I 1 I I 1 I I I I I I I : w

0

w 0

C) CD

c Q

C) CD

0

3 CD

CD 1̂••,f•

t—f

O

CD Cl) rf

0 D rm•t

` ✓

Q

n)

Cm) CD r

C')'

CD CCD