Minutes for the CIC Workgroup from the October 2015 HL7 WGM in Atlanta, Georgia

Minutes for the CIC Workgroup from the October 2015 HL7 WGM in Atlanta, Georgia

Tuesday 10/6/2015

Q1: Joint meeting with BRIDG and RCRIM to review Bipolar DAM and discuss modeling approaches in general, based on BRIDG experiences.

Question for this group: Should we have a generic Mental Health model developed instead of creating more specific models?

Current DAM model development – plan is to ballot in January 2016; publish in March 2016
Current work is based on the abstraction of data elements from 24 trials
169 written comments received and 65 assessment scales
Use cases and activity diagrams were also circulated

Total of 311 Data Elements abstracted from 24 trials, 29 sources included the trials and 5 Standards

54 new data elements considered

Now incorporating feedback from review division into the content and updating the Model in order to prepare for a January ballot. Ballot will be informative.

Once the DAM passes informative ballot, the content is handed over to CFAST to harmonize with their content. The definitions passed over to CFAST are most visible as benefit will be seen in the CDASH content.

FDA comments: When new CDISC standards are applied to trial, they are different from existing version of the standards and that creates the need to do a great deal of reconciliation across the data. Data collection prospectively should use the new standards, but aggregating across all data is a big challenge. FDA needs better ability to turn on and off business rules. FDA has worked with the vendor to make them aware of the problems when only the newest version of the standards is applied to the data through the validation tool.

Issue – how to reconcile legacy data already collected along with new data (which Bipolar will have).

Anita said that a great deal of information about mapping was included in the User Guide, but the FDA said that there is too much information in the User guide. So the team is now trying to figure out how to take out some of that information.

James Topping developed the model – he started with the Major Depressive Disorder and then made changes for Bipolar.

Common model – a person would need to avoid a large part of the model that does not apply. Would a Data Manager be able to do that? Yet the commonality would be a benefit.

Behavioral health model exists – is that in line with this work?

Anita said that the team reviewed the Behavioral Health Model, but it was too high level for use for this purpose, and finding content was too difficult. Behavioral Health model was also undergoing changes.

BRIDG is looking at the ability to put tags on content that would allow only a subset of the model to display. There would be a common foundation model with other content appropriate to certain domains. BRIDG has become so large that it has lost its ‘domain friendliness’.

How do you handle content of one domain from another?
How do you handle different levels of granularity required?

If another layer were put on BRIDG, a ‘conceptual layer’, related concepts to what a user needs could be surfaced. Then a ‘design layer’ is also needed for implementation views. BRIDG is challenged because one foot is in the conceptual layer and the other foot is in the implementation layer.

How does this content fit into BRIDG?

Domain experts in different mental health areas – schizophrenia, depression, bipolar – think about the same elements differently. It does not always mean the same thing across these groups. It is more work to force harmonization than it makes sense to do.

Is the value of using BRIDG covered in training?

How do organizations now use BRIDG?

Are there vendors using the BRIDG model?

Is there BRIDG training being created or used?

Not quite there yet with compliance but no criteria for compliance yet.

A group cannot state ‘BRIDG compliance’ at this time.

Next WGM – we will meet again in January on Tuesday Q1 at this time. The Bipolar ballot will be going through reconciliation at that meeting.

Q2: Joint meeting with EHR usability WG

Discussion of reviews of literature, Interrater reliability issues; John went through a large set of comments.

Conformance Criteria created –

Ballot reviewed – and the HL7 functional model was reviewed.

Submitted a journal article to JAMIA

The overview chapter of the Usability HL7 User Guide will contain all the definitions of concepts like ‘usability’ and ‘function’. Three categories of Heuristics included in the ballot responses.

Important to include self-describability of the content.
Suitability to the task

Reference to the NIST standard made - ?

Usability testing – nurses and other clinicians are often late or not even involved in Usability and acceptance testing of an application.

Review of the ballot for EHR usability – many comments received.

Many of the comments focused on usability vs. function of content

Discussion of trust and traceability of entries in the electronic record very interesting. While the paper record is always able to be reviewed to validate a series of events, this does not occur in the EHR system, so there is an underlying problem with trust in the system.

Further discussion about vendors implementing EHR system(s).

Q3: Trauma Meeting

New Trauma PSS: Updates being made to the current content, US realm project

Changes will be made to the value sets
Creation of an IG – the new IG will update the version of the underlying template that is being referenced
Balloting to normative
Expanded the scope to include performance measures for a provider

Implementer community is the people with Trauma Registry software

Ed Hammond: Who is the implementer community?

AMS: The American College of Surgeons will know the vendors who are using the software. These are:

… a list of about 5 implementers that can be identified
If there is an identified set of people that will be included, they should be identified in the PSS
Risk associated with divergent harmonization requirements – DEEDS, C-CDA, LOINC, DEEDs

There will be a case-by-case review of standards needed, and they may be different in some cases. The need for interoperability will drive decisions.

American College of Surgeons is the driver of this standard. The ACS is providing all the requirements to have a Trauma registry which is required for an organization to be ACS-certified as a Trauma Center. The ACS has up to a level 4 setting, the state has up to 6 levels.

Project will go to the US Realm; ballot by Tuesday the 13th. Then the PSS must go to the DESD for review and voting.

NTDS - National Trauma Data Standard, created 2002

Voting on this PSS: Jay moved that we vote to approve the ballot; AMS seconded the motion.

Yea – 6

No - 0

Abstain – 0

The PSS was approved to send forward to the US Domain Workgroup, and then to send it on to DESD

Next meeting: Trauma is TQ4 and WQ3

EMS is TQ3

Q4: MAX Tool update

Michael van der Zel and AMs present to answer issues associated with MAX as found by Duke Team (James Topping present).

Probably will use HL7 datatypes

Tag values to hold CDE Identifiers is something that may add value.

Issue: the model and xls were independently created, so not enough connections between the two. That can be fixed by putting in connectors. But also concerned that xls cannot hold more than two levels xml; but MAX can handle that level of detail.

Issue: Create a name for each value set – like the name of a Value Doman – and refer to it by that name in this xls. Then in the referred to spreadsheet delineate the individual value set.

MAX objectives:

Make MAX more accessible throughout HL7
Use MAX on a project that has new requirements

Make those changes in MAX to make them part of the general product

For the Bipolar project, do a transform of the xls and then decide how to represent the content in xml; the format needs to be all the items needed to transport to EA.

Need a change to MAX: Expose the ID of the value set name. Also expose the ID for the class name

James and MAX team have been meeting on Thursdays.

Meeting in January at HL7 WQ4

Wednesday, October 7, 2015

Q1: Hospital Ballot Report - Ballot Reconciliation - Jay Lyle

Ballot comments inventoried. Jay Lyle invited people to attend from Emergency Services and Structured Documents. Emergency Services said someone would attend, but no one has shown up – email sent to remind them. No response from Structured Documents. Both cases were people who had problems with the content included in the ballot.

121 comments – classified in 46 groups of similar content; discussion included the following:

Jo-Anne Johnson from Kaiser – found it very hard to read, had a lot of gaps. In general, ‘it just wasn’t ready’. A lot of discussion and review came from Kaiser. It needs more description and explanation about the goal. It could be improved to make it more helpful to the readers. Need more explanation and context
This is noted as an inpatient encounter with a different encounter for the ED visit. There are lots of niche systems for ED visits. EDs have a completely different EHR system; for major vendors like Cerner the ED system is usually an add-on. So there is a need for a niche system which the physicians may buy themselves. Do we need to cease to make the distinction – the hospital sends the record, the ED sends it, or a combination of the two send the information to the medical record. We can remove the distinction and have the definitions ‘tweaked’.
First ED systolic BP refers specifically to the first BP after arrival. Temporal qualifiers on the data elements need to be avoided.
ED Disposition or Discharge diagnosis – If a person is admitted, that would be Admission Diagnosis. If the patient is sent home it would be discharge diagnosis. It is not more than one thing, but dependent upon the patient’s disposition from the ED.
Hospital diagnosis is less valuable to gather than the ED information.
Ballot will abandon the design that attempts to represent two encounters in one document (based on the comments and discussion)
Procedures are included in an Outcome document, and so the Title of the document is at odds with the inclusion of procedures. The term ‘Outcomes’ will be dropped from this section. There is already a Procedures section, so how is this one different? Do we constrain the section for ICD-10 content only? In reality, few groups may be using ICD-10, so we need to take whatever we can get.
After much discussion, decided to use primary Procedures section for all procedures.
Assessments done of the functional status at discharge is an Outcome. From the Emergency Services perspective all the content is not described in a way the hospital folks will understand. This content is purely to be able to evaluate if the EMS did a good job. This is patient outcomes post EMS care. Major areas of feedback need to come back to the EMS group.
Important – the document needs good explanation that this is Patient Outcomes post EMS care. Use ‘EMS Outcomes Report’.
Rankin Scale – do not limit it to only this scale, but will take other performance status evaluations.

Vote on the items #65, #63, and #58, #46

Move to vote to approve the disposition – Jay Lyle

Second – Dianne Reeves

Yes - 5

No - 0

Abstain – 0

Q2: Patient Outcomes EMS Report – Ballot reconciliation

This Quarter is a continuation of discussion from Q1 to reconcile Ballot comments. We will evaluate industry acceptance of the idea (User meeting next week) that EMS quality feedback is an appropriate use of a Discharge Summary. There is a caveat that some of the content may be ‘masked’. Our intent will be to specialize the Discharge Summary in a few key places for EMS specialized data to make it into an EMS Outcome Report.

Is this a change that will require reballot?
This approach will deal with the bulk of the confusion over this report.
How about a comprehensive statement for the ballot disposition? We agreed that the best way to proceed will be a reballot. This will remove the confusion and will probably make a repeat ballot easier.
If we are going to use the Discharge Summary ‘as is’ and withdraw the Ballot, why can we not move forward? We need to evaluate the Discharge summary to see if a gap exists.
Reuse of the Discharge summary – no ballot needed
If you want to extend the DC summary to add elements as a Standard, we need a ballot
We need to evaluate the Discharge summary now to define any gaps.
It is not the EMS system’s duty to supply a discharge summary report to the Hospital/HIE. Rich suggested that if extra items are needed beyond the Discharge Summary, a special interest group be setup on the next NEMSIS call to hash out what is needed.
Rich would take the Discharge Summary back to the users to identify the elements that are missing. In the meantime the ballot should be withdrawn. How long can the ballot remain open? Probably need to resolve items quickly.
Decision: It is not possible to avoid doing another ballot if we proceed with this plan.
1. An analysis of the gap between EMS needs vs. Discharge Summary needs to be done first; based on the analysis DC may be sufficient or another standard may be needed.
2. Continue with this ballot and reconcile comments - No
3. Identify what the larger community needs – larger requirements gathering activity – No
4. Identify a small set of elements and identify a standard set that could be used with the Discharge summary – No
If we want to approve a Standard Template, there is no alternative to a reballot.
Decision: Take what we have, put them into another project template with notes about how to use it and why. Meanwhile, if the list of elements needed beyond the DC summary is identified it can be added into the next version of the Template
Ballot a document and IG – this is an EMS Standard for Discharge Summary Guide, and here are some things EMS is particularly interested in. The Discharge Summary Template has been balloted, and the constraint is that it contains a handful of things EMS is interested in. The Discharge summary is already Normative. Feedback on the Discharge Summary will not be the goal – only looking for feedback on the additional elements.
Jay wants to reballot in January – tight schedule. Do we aim for January or aim for something faster?
Jay needs to respond to all the comments on this ballot before it can be reballoted. This will be a reballot, not a new ballot or new PSS.
If funding does not permit a new ballot, we can say that we’re going to use the Discharge Summary.
Can use the standard DC Summary template ‘as is’ with extensions in data elements needed for EMS, and avoid a reballot.
If we use the DC summary and add EMS elements, should we ballot straight to Normative? This could be faster than going to DSTU first.

Thursday October 8, 2015

Q1: Report out of CIC Projects

Ed: We need to pick a problem, and then create a plan for how to handle that problem working collaboratively with other groups. In the past, all the work of HL7 was done at the WGM; now almost all the work is done outside the WGMs. We need to identify a common clinical problem that will need standards to solve, potentially in a clinical area like oncology or cardiovascular.

Anita: How do other workgroups do their work? In CIC all of our work has been funded through grants or FDA.

Ed: HL7 is starting to get grants, so can we do a project with HL7 funding? Chuck Jaffe is starting to teach Mark Hamill and Karen to write grants, to get funding for the organization.

Russ Leftwich leading LHS, trying to accelerate learning from publication to use in practice. He has formed a User Group with about 35 people, broader than physicians.

Original mission of CIC was to find a way to keep clinicians involved in the standards work. Once you add technical people into the process, the clinical people don’t come back. There still needs to be an Outreach into the clinical community, and there still needs to be a Translation point between technical content and the clinical group.

Ed: I really want to form a team to solve these problems. It has to be a partnership. I would not bring the clinical group to HL7 to do the work. The plenary session may be of interest to the clinical groups, but HL7 people also need to attend clinical meetings.

Ed: There are many groups that could offer funding to groups that have informatics problems. We could put together a list of funding sources if that is what is needed. At one time the source of these projects was from vendors. Now that is not the primary source. It’s worth reexamining our mission and how we do our work – are we spending too much time in administrative tasks.

Ed: CIMI has voted to become an HL7 workgroup, and are up now for HL7 voting and inclusion.