Orders & Observations

San Antonio Working Group Meeting

May 2006

Meeting Minutes

Table of Contents

Attendees

Monday Q1/Q2, 2.7 Proposals

Tuesday Q1 – OO/Rx/Lab

Tuesday Q2 – Lab, IHE, O&O, Struc Doc

Tuesday Q3 – OO/Pt Safety/Rx/BTO/PHER

Tuesday Q4 – OO/Pathology

Wednesday, Q1 – OO/Dev

Wednesday – Q2, OO/Lab/BTO

Wednesday - Q3 – Genomics/Lab

Wednesday – Q4 – II

Thursday Q1, Order Set Overview

Thursday, Q2 – OO/Rx – Dynamic Model

Thursday, Q3/4 – Clinical Statement

Attendees

Attendee / Company/E-Mail / Mon AM / Mon PM / Tue AM / Tue PM / Wed AM / Wed PM / Thu AM / Thu PM / Fri
Liora Alshuler / / √
Rita Altamore / / √
Kay Avant / / √
Calvin Beebe / / √ / √
Fred Behlen / / √
Deb Belcher / / √
Charlie Bishop / / √
Keith Boone / / √
Bill Braithwaite / / √
Louise Brown / / √
Nicholas Brown /
Hans Buitendijk / / √ / √ / √ / √ / √ / √ / √ / √
Michael van Campen / / √ / √ / √
Jim Case / / √
Todd Cooper / / √
Garry Cruickshank / / √
Richard Dixon-Hughes / / √
Bob Dolin / / √
Dick Donker / / √
Kristi Eckerson / / √
Dav Eide / / √ / √ / √ / √
Brett Esler / / √
Peter Elkin / √
Mollie / …@harvard.edu / √
Clive Flashman / / √
Chris Foye / / √
Ken Fuchs / / √
Marguerite Galloway /
Rick Geimer / / √
Ken Gerlach / / √ / √
Gay Giannone / / √
John Gilbertson / / √ / √
Sarah Glamm / / √
William Goossen / / √
Freida Hall / / √
Rob Hallowell / / √ / √ / √
Dick Harding / / √
John Hatem / / √ / √ / √
Peter Haug / / √ / √
Rob Hausam / / √ / √
Masaaki Hinai / / √
Jim Horton / / √
Mary Jarquin / / √ / √ / √
Gaby Jewell / / √
Yaria Jaroch / / √
Dave Johnson / / √
Jeri Jones / / √
Tom de Jong / / √ / √
Kyunghee kang / / √
Bertil Keppin / / √
Gert Koelewijn / / √ / √
Helmut König / / √
Alexander Krans / / √
Austin Kreisler / / √ / √ / √
Thom Kuhn / / √
Yun Sik Kwak / / √
Sungkee Lee / / √
Andrew Lindy / √
Patrick Loyd / / √ / √ / √ / √ / √
Chris Lynton-Mell / / √
François Macary / / √
Joginder Madra / / √
David Markwell / / √
Michael Martin / / √ / √
Brett Marquard / / √
Jim McCain / / √
Ken McCaslin / / √ / √ / √
Doris McGinnes / / √
Larry McKnight / / √
Gary H. Meyer / / √ / √
Patrick Mitchell-Jones / / √ / √ / √
Jun Nakaya / / √
Karen Nocera / / √
Thomas Norgall / / √
Wendell Ocasio / / √
Mike Ostler / / √ / √ / √ / √
Charles Parisot /
Craig Parker / / √
Jaqui Parker / / √ / √
Fola Parrish / / √
Chris Peck / / √
Diana Perez-Lopez / / √ / √ / √ / √ / √
Andrew Perry / / √
Backy Reed / / √ / √
Melvin Reynolds / / √
Scott Robertson / / √ / √ / √
Craig Robinson / / √ / √
Dan Russler / / √
Tod Ryal / / √
Gunther Schadow / / √ / √ / √ / √
Barry Schell / / √
Dave Shute / / √
Karen Sieber / / √ / √ / √ / √
Ioana Singuraehu / / √
Corey Smith / / √ / √
Rik Smithies / / √
Harry Solomon / / √ / √
Rene Spronk / / √
Helen Stevens Love / / √ / √
Nick Steblay / / √
Lise Stevens / / √
Jeff Sutherland / / √ / √ / √ / √ / √ / √
Sadamu Takasaka / / √ / √
Jack Taylor / / √
Anita Walden / / √
Mead Walker / / √
Jan Wittenber / / √
Bob Yeneha / / √

Communication with declared O&O participants can be done through . You can sign up through the HL7 website, List servers for focused aspects of the O&O domain are: , , , , and .

Monday Q1/Q2, 2.7 Proposals

See attached document for proposals discussed to date with discussion and disposition.

Tuesday Q1 – OO/Rx/Lab

Common Order

  • Context Conduction should be mandatory with default values so it can be used based on exception.
  • ObservationEventCriterion
  • Effective Date/Time in ObservationEvent Criterion: Example is where a headache lasts more then an hour. Agreed to add.
  • No recursion in the general model. Rx will work on getting it out of Rx model. Only if use case arises will this go back in.
  • ActDefinition
  • Changed to 0..*
  • ObservationRequest2
  • Rx to put in Status code.
  • Entry Point in Rx
  • Rx considering choice box so it can come into Header, Supply, other.
  • Agreed to have Common Order have a choice box. Rx would follow suit. Lab will constrain to a specific entry point.
  • InformationRecipient
  • Do we need InformationRecipient for notification to clinical trials?
  • In V2 we only reference clinical trial identifiers, not the actual recipient.
  • We should consider to put the identifiers in Common Order. Rx will likely pick it up immediately, while Lab is contemplating whether it is needed.
  • Hans to check whether the V2 use case for Orders (CTI) is still relevant.
  • Consent
  • In ControlAct or focal act. Should
  • Patrick and Michael to take to I&M.
  • CallbackContRequestChoice
  • Rx agreed to take this on.
  • DataEntryLocation
  • Should it be in the main model or just in the Control Act?
  • Due to reports and queries we agreed to keep DataEntry in the model.
  • Annotation
  • Agreed to change R_AssignedPerson to R_AssignedEntity.
  • Performer
  • Need to be able to point to a person or organization.
  • We all agree that we need to recognize performing organizations, not a scoper of an implicit person.
  • Specimen
  • Problem that the entity is CE. Consequently cannot support SNOMED.
  • Propose EntityCode from CE, CD as harmonization proposal.
  • Against: 0; Abstain: 2; In Favor: 25.
  • 0…*/1…N Participation or Act
  • Add sequence number to Participation and ActRelationships and priority to ActRelationships. Where cardinality is greater then 1.
  • Coverage
  • Use case indicating special authorization. Requires 3+ attributes.
  • A_Coverage is too strict. Need to relax to support Identifier, Effective Date, Insurance Carrier.
  • Rx and Lab will check on how much data they need.
  • Common Order will already include A_CoverageUniversal as a placeholder.
  • Need to check with some PT knowledge how much they need as minimum as they will providelikely use cases for this area.
  • SupportingClinicalInfo CMET.
  • Does not support previous, cannot support nested. In short, not fit for purpose.
  • Considering to eliminate SupportingClinicalInfo altogether.
  • Patrick to check whether SupportingClinicalInfo is indeed a proper subset of Clinical Statement.
  • We will forward to Clinical Statement that we’ll drop SupportingClinicalInfo CMET and get Clinical Statement into CMET form.
  • ReplacementOf
  • Old order becomes obsolete. Requested that an obsolete transaction should come before the replace transaction so that the trigger event only works on one focal act.
  • Everybody agrees that the orders space should allow for three approaches:
  • Obsolete
  • Replace – Obsolete, New
  • New
  • Level of authorization and use case drives which one to use.
  • Move to agree that the ReplacementOf trigger event should be able to cover both an obsolete and replace trigger.
  • Against: 0, Abstain: 0; In Favor: Unanimous: 28

Tuesday Q2 – Lab, IHE, O&O, Struc Doc

Objective: how Lab SIGballot, Clinical Statement, IHE CDA lab report, and CCD all inter-relate.

Francois provides IHE objective overview.

  • Scope excludes Blood Bank and Pathology. Includes CDA R2.
  • The Lab Report is not part of the Order Fulfillment process, rather AFTER it’s completed, including historical data.
  • Three use cases:
  • Post Discharge
  • Ambulatory Lab
  • Physician EHR Lab
  • Bottom-Up Approach starting with existing paper reports from various sources focusing on lab data.
  • For Level 2 using Sections while at Level 3 using Organizers for “promised items”.
  • Slide 16
  • Problem: CDA does not support R_Specimen and therefore the procedure to collect the specimen cannot be attached to the specimen directly, rather then the organizer. Can be resolved with Level 3, but not Level 2.
  • Problem: Unclear which organizer to use for the battery in CDA.
  • Slide 17
  • Problem: Unclear which method is used (MIC vs. Kirby) or all susceptibility forced. Probably a rendering question.
  • Slide 18 – Issues summary
  • Two more issues.
  • Showing previous result tied to current
  • Expressing global comment on a battery
  • Other Issues:
  • When to use Lab result message versus CDA Document?
  • Single document vs. differentiated document type code?
  • How can we have a Cumulative report that crosses multiple orders? Should there be many or if there are many just zero?
  • Flow to resolve issues for IHE.
  • Lab first (Core Domain First)
  • Clinical Statement second (for long-term issue resolution) (Pattern resolution)
  • Structured Document third (for short-term issue resolution and any long-term implications) Extensions back to Clinical Statement. (If used as the communication vehicle).
  • Need formalize short-term issue resolution approach. Currently Liora and Glen facilitate reconciliation. Based on best available knowledge where things may be heading.
  • Need to get the constrained model for Lab (as other domains) for Clinical Statement.
  • Revalidate Clinical Statement.
  • Most burning issue: Are the resolutions short-term acceptable?
  • Could consider to go from procedure to specimen or put procedure on top.
  • Criterion can be on choice-box.
  • Problem is that the specimen is the product of the procedure.
  • What if multiple specimens and multiple procedures? Which belongs to what.
  • Use the CDA participant relationship and specialize.
  • Public Comment to all parties and IHE to package the specific requests.

Tuesday Q3 – OO/Pt Safety/Rx/BTO/PHER

  • Rx Harmonization
  • Need to ensure that Patient Safety is kept up-to-date with changes.
  • Need to schedule harmonization meeting.
  • Rx has list of changes.
  • Circulate ICSR ideas to Rx in between. Exchange documentation on list servers. Rx to include Patient Safety and Patient Safety joining Rx list serve.
  • Adverse Reaction and Allergies standing Wednesday Q3 meeting.
  • BTO
  • Generating story boards.
  • Not enough materials yet for Rx and Patient Safety.
  • Establishing conference calls. will be the vehicle to establish the timing.
  • Immunization
  • PHER took on Immunization as a project. POIZ stays as separate domain. Patient Care and O&O able to absorb meeting time if needed.
  • Working on receiving use cases to create HL7 Immunization Landscape.

Tuesday Q4 – OO/Pathology

John Gilberson - Pathologist - Presented overview

co-Chair Pathology SIG

Identifying a specimen and doing a study on it is key issue.

Need an Anatomic Pathology domain. Lab SIG model can be used.

Want to invert data on multiple reports on the same specimen.

1. Trying to develop an AP domain model in Visio.

2. Will map model to Lab model.

Timeline, resources, anyone want to help?

Have 5-6 pathologists interested who will come to next meeting.

1. Will develop coherent model by next meeting.

2. Will look at existing specimen model and comment. Need to understand it better.

3. Patrick Loyd has a task to determine whether V3 or CDA is best for AP domain.

4. Will begin developing an implementation guide for CDA.

Could do an AP domain model in V3 and then describe how this would be used in CDA.

DICOM Pathology Working Group must use HL7 specification model. They have story boards.

Lab SIG has initial specimen model. Patrick and Gunther have it.

Tissue banking needs more detail in the specimen model than traditional pathology documentation needed for cancer registries, etc. Newer systems support this while older systems do not.

Specimen model in current ballot was reviewed.

Block, level, and ordering of slice of speciment not in model.

Speciment treatment needs a time stamp.

When you get a specimen of prostate, for example, you describe physically where the block came from in a specimen. In a big specimen, some tissue will get thrown out. Every block gets a letter or number and free text description (currently). Should be coded better.

A good system will describe site, block number, slide number from block and describe procedure done on it.

Action items: (John Gilbertson)

1. Look at Lab specimen model carefully.

2. Come back at next meeting with issues.

3. Determine how to represent model in traditional pathology report.

Wednesday, Q1 – OO/Dev

Reviewing various device questions covering their modeling activities.

  • Recursive Act Relationship
  • Use Lab as example for recursive construct of Component on Act. Objective is to insert into a new ballot round in September.
  • The IHE implementation guide will need to put this back into V2. Can either use Parent/Child or OBX Sub-ID using the “.” approach.
  • Null Value
  • Various null flavors. Appear to be insufficient.
  • Examples are in V3 data type. If anything is missing, will work with Gunther/Patrick for a harmonization proposal.
  • We need a V2 proposal for OBX to allow to accept null code values. Gunther and Hans to check what needs to be changed in V2.
  • MedicalDeviceAct.code
  • If IEEE has a list, use that, otherwise further work with Vocab to establish.
  • Filter Mechanism – What should we use?
  • Year 1 has three profiles for Point of Care Devices.
  • Primary profile is the inter-enterprise communication of device data.
  • Third is PCD broker enabling requests for certain data, e.g.,
  • Check with Chapter 5 for V2.
  • Unsolicited Observations against Standing Orders
  • Check R30, R31, R32
  • Personal Health Record Updates
  • See Clin Statement for relationship with “non-clinicians” as authors/informants to deal with patients or persons on their behalf providing results.
  • In V2 check out the Role Segment and utilize the role code for participation/role definitions.
  • Message vs. Document
  • Key differentiation is the readable/formatted aspect of a document versus message. Next is that attestation is more likely on a document then on a message. Beyond that characteristics could be applied to either message or document.
  • Critical that to the extent that data can be structured by the source, the same structure is used. Clinical Statement is a good example of a pattern that conveys the core data as part of a message or a document. The packaging around it varies.
  • Risk Management Paradigms
  • See slides by Jan Wittenber

Wednesday – Q2, OO/Lab/BTO

  • Strange problem occurred with BTO content impacting Lab while no known changes had occurred. Diana to check with Woody what happened and determine what BTO may need to do to ensure it does not happen again.
  • BTO (definitely) and Lab (likely) will skip the next ballot round.
  • Lab needs SupportingClinicalInfo to be extended to meet their needs. We also want to pursue turning ClinicalStatement pattern into a CMET.
  • Prefer to not change SupportingClinicalInfo for backwards compatibility.
  • Agreed to pursue SupportingClinicalStatement CMET that can be serialized (current tooling restriction) as replacement to SupportingClinicalInformation CMET. Patrick Loyd and Patrick Mitchell-Jones will pursue.
  • Need to know what can be added to a Clinical Statement without being non-compliant. Rules need to be documented.

Wednesday - Q3 – Genomics/Lab

  • Update
  • Updated the approved DSTU documentation
  • Genotype Topic now contains two models.
  • GeneticLocus – appears as CMET
  • GeneticLoci
  • Can get to all core classes from the entry point.
  • Associated properties/observation classes consolidated
  • Phenotype choice box replaced
  • Observed Phenotype class to be replaced by SupportingClinicalInformation CMET
  • Haplotype class removed and replaced with place holder.
  • tagSNP and translocation classes removed and re-inserted as codes
  • Family History
  • Added new class – FamilyHistory
  • Family History is entry point
  • Added classes to hold results of risk assessment
  • Various other changes
  • Question about sensitivity and specificity as input parameters. Suggestion to use control variable instead. To be reviewed.
  • Family History Issues
  • Suggest to use observation construct to carry clinically relevant ethnicity/race, rather then attempting to extend the more administrative/government defined ethnic group and/or race code.
  • Harmonization required with CDA and CCD/CCR.
  • Provided guidance that the fact that the family history references a Genotype CMET that is still in DSTU state should not be a limitation on the Family History to go normative, as long as the CMET reference is optional.
  • V2
  • Various ideas how to use OBX.
  • Clinical Statement/Genomics

Wednesday – Q4 – II

  • Common Order Status
  • Motion to put Common Order in Draft for Comments
  • Against: 0; Abstain: 1; In Favor 8
  • Ballot Feedback
  • Resolved negatives and comments and is ready for Membership.
  • Challenge is that a CMET is using the Clinical Statement DMIM as the parent. That may cause a problem when Imaging is normative while Clinical Statement is still on its way there.
  • Clarification required on extension guidelines to
  • Workflow Model – Recursive relationship
  • There are two alternative ways to model workflow process steps. One is to have ProcedureRequest relate directly to the ScheduledProcessStep and PerformedProcessStep, or through a ScheduledProcedure and Performed Procedure respectively.
  • II does not need the middle layer while others do. The choice is therefore to accept two different modeling techniques, leading to similar data being sent in two different ways and requiring clear guidelines when to use one or the other.
  • Infoway Proposal
  • Need registry query (list of studies for individual), order, and detailed observation request involving images

Thursday Q1, Order Set Overview

Craig Parker provided an overview of the approach to Order Sets. Attached are the materials used.

Various questions/discussions were conducted during the review. The primary conclusion we reached is that we need to continue to discuss the approach to ensure we end up with a consistent approach based on O&O order patterns.

Thursday, Q2 – OO/Rx – Dynamic Model

Topics

  • Project Team
  • We seem not to be able to get a project team going to focus on the Dynamic Model that cuts across the different areas.
  • Should we consider an out-of-cycle, e.g., Sunday prior, or Friday, on Dynamic Model?
  • Should we focus on creating documentation rather then just making decisions?
  • Hans to send out request for interest to meet on Sunday prior or Friday for a full day.
  • Objectives
  • Initially focus on common set of rules on how to deal with order management state transitions after the creation of an order.
  • Covers lab, Rx, Rad, PT, Dietary, NRS, Procedures, etc.
  • Subsequently:
  • Create set of rules on how to deal with observation state transitions
  • Synchronize with referrals and appointments.
  • Referral is patient care provision in order mood.
  • Wiki
  • We agreed to use the wiki site to channel discussions.
  • Hans to populate raw data. Notes and initial seed of the documentation
  • Copy-n-paste the general act state machine as well.
  • Documentation
  • We should take current state machine of act and elaborate under the common order model.
  • Alternatively, we can use storyboards as starting point.
  • We need to provide both the storyboards and state machine.
  • Approach
  • Should we use the HDF approach to generate the artifacts necessary?
  • Should we go back to the topics discussed to date and start to go for each one of them and start to produce these artifacts and refine our approach?
  • We agreed to address Storyboard (wiki/pub db), Activity Diagram (visio), Interaction (pub db), Trigger Events (wiki/pub db), Application Roles, State Transition Definition. (those in bold are needed for ballot).
  • Left to Right or Right to Left?
  • Do they all end up in the ballot and therefore, do we need to do them?
  • Consider listing/prioritizing all topics, or take 1-2 storyboards and flesh them out completely. The latter would not touch all topics, but complete all necessary steps.
  • Tom and Rob to provide initial story boards.
  • Michael/Garry C initial “pub drivers”
  • Wendell initial “visio driver”

“Combined State Transitions”