200005GovSIG - HL7

Document Sample
200005GovSIG - HL7 Powered By Docstoc
					                        Government Projects Meeting
                            Tuesday, May 23 AM

Meeting participants:

Name                        Affiliation                    E-mail
Denise Koo                  CDC                            Dkoo@cdc.gov
Dan Jernigan                CDC                            Djernigan@cdc.gov
Marvin Lerfald                                             Marvin_lerfald@srm.com
Robin Zimmerman             Kaiser Permanente              Robin.j.zimmerman@kp.org
John Berry                  3M                             Jjberry@mmm.com
Mark Williams               Cimarron Medical Informatics   Sitkacmi@gci.net
Mike Lincoln                VA Salt Lake                   Michael.lincoln@hsc.utah.edu
Wally Fort                  VA San Francisco               Wally.fort@med.va.gov
Jim Demetriades             VA                             Jim.demetriades@med.va.gov
Ken Rubin                   EDS                            Ken.rubin@acm.org
Christian Bremeau           Meta Integration               Bremeau@metaintegration.co
Bill LeClair                Meta Integration               Leclair@metaintegration.com
Mead Walker                 IDX                            Mead_walker@idx.com
Angelo Rossi Mori           CNR                            Angelo@itbm.rm.cnr.it
Nancy Orvis                 DoD                            Nancy.orvis@tma.osd.mil
Karen Van Hentenryck        HL7 Headquarters               Karenvan@HL7.org

I.     Introductions and Review of Minutes from Atlanta Meeting –

       Denise Koo opened the meeting by inviting participants to introduce
       themselves. The minutes from the San Diego meeting were reviewed and
       accepted without change (anyone finding required changes should send
       them directly to Karen Van Hentenryck).

II.    NCVHS Mapping –

       Mead Walker provided a summary of the mapping of the NCVHS core
       data elements to HL7, both Version 2.3.1 and the RIM. Mead indicated
       that the HL7 Standard is generally well suited to transmit the contents of
       the NCVHS core data elements as a minimum data set. Some of the
       NCVHS fields are not precise (e.g., Medication Prescribed) and will
       require clarification before mapping. The questions now are how should
       this be implemented (with the point made that the NCVHS
       recommendations for core data elements were not necessarily intended to
       be an actual data set), what would it mean to ask people to support a
       minimum health record, and what should be the next step. Other
       organizations may also have minimum data sets. A question was raised
       regarding the Type of Facility/Place of Encounter field. Both the DoD and
       X12 both have tables for this field, which is mandated by HIPAA legislation
       as an X12 standard. It was noted that the X12 codes should be part of the
       HL7 standard under the US Realm within the HL7 vocabulary tables. PRA

       table suggestions should be forwarded to Liora Alschuler for the PRA.
       Karen Van Hentenryck will post Mead’s document to the Government
       Projects page, and the mapping will be shared with NCVHS

       Click here to view the mapping document

       Click here to view the Powerpoint presentation

III.   RIM Harmonization –

       Denise Koo reported on the progress of harmonizing the Public Health
       Conceptual Data Model (PHCDM) to the HL7 RIM. Ideally, the PHCDM
       would overlap substantially with the HL7 RIM, although not all elements of
       the PHCDM can or will be harmonized to the RIM.

       The PHCDM class Party is similar to the RIM Stakeholder class, but the
       PHCDM class is broader and doesn’t have the legal requirements of the
       RIM Stakeholder class. Consequently, the PHCDM model has two
       classes: Party and Organization. Under the Individual class, the PHCDM
       model has two specializations – person and non-person living organism.

       The PHCDM class Health-Related Activity is similar to the RIM Service
       class. Within the PHCDM model, Intervention is a subgroup. The CDC
       feels that there are interventions such as educational campaigns (e.g.,
       billboards to help stop smoking)that don’t just involve a medical
       procedure. The PHCDM model also includes a Case class (a type of
       Observation for determining cases of diseases or conditions of importance
       to public health) and a class called Outbreak (a specialization of Case).
       Notification is also new class in the CDC model. There was a suggestion
       to put a more specific description on the Case class. Maybe it should be
       called Public Health Case.

       The PHCDM class Location is essentially used to pinpoint how to track
       down a person, whether by physical location, telecommunications
       location, and potentially, text directions (go to the bar and ask for him, in
       case of homeless persons, for example).

       There was a question regarding how much the scope of the RIM should
       expand. Some items that are specific to public health may not need to be
       included in the RIM. The RIM cannot expand indefinitely. All interested
       parties are invited to sit in with PAFM from 1:30 - 3:00 today to discuss
       changes to the RIM, to harmonize with the PHCDM model. After cookies
       all interested parties are invited to sit in with Orders/Observations, where
       we will also be discussing needed changes to the RIM. On Wednesday,
       Orders/Observations will have additional discussions about Living

      Click here to view the entire presentation.

      Click here to view the comparison document

IV.   Lessons Learned from HCFA’s Implementation of HL7 –

      Christian Bremeau and Bill LeClair provided an overview of how HCFA is
      using HL7 technology in their ESRD (End Stage Renal Disease) project.
      The current ESRD system uses lots of paper forms and they want to shift
      from collecting information on paper to collecting it electronically using

      The first step involved mapping the data to the RIM, not to integrate the
      two, but to discover the common elements and thereby determine the
      appropriate HL7 messages to use. Currently, ESRD facilities are
      reporting to HCFA via paper forms. The ESRD project utilizes a Vision
      Data Collection Tool, HL7 XML and a SIMS network (which is an SQL
      database) to transmit the information. HCFA would eventually like to have
      a live transfer from the hospitals to their database but cannot support this
      feature now because of security issues.

      The main goal of this project was to build HL7 XML message for HCFA
      forms. To achieve this goal, they used

         The HL7 MDF V2.3

         Rose Tree

         RIM095 with no modifications

         Standard coding schemes (LOINC and ICD-9)

      They felt fortunate that USAM (Unified Service Action Model) existed.

      The XML will be shipped via Hypership, which is used by the US post
      office to encrypt the XML fields and the address (www.wareonearth.com).
      The XML file is sent to the US Postal service, which encrypts the data and
      verifies that it has not been tampered with.

      Phase 1 will be implemented in August.


         PRA

            Does not contain enough RIM information

            Not all the information seems to be in the RIM (i.e.,
             Clinical_Document_Header with patient relationship)

            Closest to what we need if timeliness is an issue

        Could not find other messages dealing with documents.

     Lessons learned

        The RIM is fairly complete – found no needs to create new
         relationships or to create new attributes

        It is sometimes hard to choose a path when several paths are possible
         and valid (service for example)

        Yes/No questions require interpretation

        Members of HL7 are very cooperative

     Waiting on

        Static – not yet defined

        Used Xeena tool from IBM (free validating XML editing tool)

     Click here to view the Powerpoint presentation

V.   Lessons Learned: DTDs and XML –

     Nancy Orvis and Bill LeClair reported on the lessons learned from using
     HL7 PRA. Bill reported that one of the biggest problems is that an invalid
     tag can pass though the messages. Another major concern is the size of
     the messages. XML is noisy – like 20% data and 80% noise. Readability
     is also an issue – should it be human readable or machine readable.
     Many of the rules are within the comments, and thus the message is hard
     to parse via tools. They would like the DTDs to be readable enough that
     non-HL7 experts can use off-the shelf tools. It would be ideal to be able to
     incorporate coded schemes from different people. Other agencies wishing
     to stand behind Bill’s recommendation should respond. Mead suggests
     that we get some time on the XML committee’s agenda to discuss these

It was suggested that Bill LeClair contact Paul Biron to convey their
suggestions. Everyone is encouraged to join the PRA ballot pool.

Angelo Rossi Mori reported that he met with the clinical templates group
on Tuesday and that it is his understanding that templates will not require
a ballot. Instead, the group needs to use the HMD and RIM to produce
non-balloted messages. There would be two different types – one would
be a balloted set of messages and the other an unballoted set of
messages that simply reply on the HMD and the RIM. In the second
instance the SIG or particular group (such as HCFA) would assume
responsibility. The proposal was to form a new clinical templates SIG.

                        Government Projects Meeting
                           Thursday, May 25 AM

Meeting participants:
Name                         Affiliation                  E-mail
Denise Koo                   CDC                          Dkoo@cdc.gov
Dan Jernigan                 CDC                          Djernigan@cdc.gov
Glenn Sperle                 HCFA                         Gsperle@hcfa.gov
Sheila Frank                 HCFA                         Sfrank1@hcfa.gov
John Berry                   3M                           Jjberry@mmm.com
Bernd Blobel                 HL7 Germany                  Bernd.blobel@miz.uni-
Hyejung Chang                KHIDI (HL7 Korea)            Hjchang@khidi.or.kr
Wally Fort                   VA San Francisco             Wally.ford@med.va.gov
Ann Hueber                   PAML                         Ahueber@paml.com
Ken Rubin                    EDS                          Ken.rubin@acm.org
Christian Bremeau            Meta Integration             Bremeau@metaintegration.co
                             Meta Integration             Leclair@metaintegration.com
Mead Walker                  IDX                          Nead_Walker@idx.com
Angelo Rossi Mori            CNR                          Angelo@itbm.rm.cnr.it
Nancy Orvis                  DoD                          Nancy.orvis@tma.osd.mil
Karen Van Hentenryck         HL7 Headquarters             Karenvan@HL7.org
Roger Metcalf                VA                           Rober.metcalf2@med.va.gov
Peter Rontey                 VA                           Peter.rontey@med.va.gov

I.     Introductions

       Denise Koo opened the meeting by inviting participants to introduce

I.     Review of Standards-level Taxonomy Document –

       Ken Rubin provided an overview of his standards-level taxonomy
       document. The Enterprise viewpoint shows how a system is to be use in
       the organization; the Information viewpoint describes the what – what info
       is being carried by the system (this is like the RIM); the Computation
       viewpoint describes how (the design), what are you doing with this; the
       engineering viewpoint provides the details of how to produce it; and the
       technology viewpoint describes the components from which the system is
       made. There was some discussion regarding whether this document
       should be presented for use and acceptance by HL7 and if so, the process
       for accomplishing that. Mead Walker indicated that the document might
       have value to HL7 but that it should be presented to the MnM group for
       consideration. Mead felt that it would difficult at this particular time in HL7

       to get enough attention from the MnM and other groups to get buy in for
       this document.

       Action Item: Ken will present his document to the MnM group for their

       Click here to view the entire presentation

II.    Public Health Data Standards Consortium Update –

       Suzie Burke_Beebe reported that the consortium currently supports six
       active workgroups. One of their current projects is to get data elements
       into the X12 standard. For example, the consortium would like to use the
       Mother’s Medical Record Number to link to the baby’s number and has
       thus requested that it become part of the X12 standard. Denise Koo
       reported that the CDC asked PAFM to allow for repeating fields for
       race/ethnicity. Suzie indicated that they are looking specifically at the X12
       837 transaction. For public health surveillance issues, they would like to
       collect the granular and clinical data that is captured in HL7. Suzie will
       forward the NAHDO report to the list server.

       Click here to view the x12n Elements Important Discharge Data

       Click here to view the High Priority & Non ANSI X12N 837 Elements for
       Institutional Transactions Summary Matrix

III.   Clinical Templates –

       Angelo Rossi Mori reported that a group interested in clinical templates
       met on Monday and decided to approach the Technical Steering
       Committee at the St. Louis meeting with a petition to form a clinical
       templates SIG.

       Angelo provided a presentation that outlined:

          Key requisites and new roles for e-terminologies

          Four kinds of e-terminologies, including their usages, topics and action

          Three pillars for semantic interoperability (repositories, templates and
           value domains)

       key requisites for e-terminologies include:

   support multiple usages of clinical data within information systems

   build coherent "families”, each made of many specialised coding

   restore continuum between content (values) and containers (data

   achieve complete interoperability among applications

New roles of e-terminology

   NO MORE "just labels"

operational perspective of (clinical) data

   key for retrieval of operative data
    from master tables and knowledge bases

   internal tables, within applications

   commercial drug databases

   triggers, e.g. warnings and alarms

   Arden syntax, MLM and decision support

Four kinds of e-terminologies usages, topics, actors

   triggers for services within applications

     life-cycle states (messages, workflows, activities)

     names of messages and segments

    system developers

    HL7 maintains closed tables - CNE

   organisational & administrative parameters

     sex, race, religion, dietary preferences

    facility managers

          HL7 maintains reference tables – CWE

         clinical / communication & scheduling

           lab tests, document names, activity names

             master tables builders & users

             HL7 endorses adequate domains

         clinical / recording

           findings, History & Physical, assessments, diagnoses, health issues

             recording professionals

             HL7 registers compatible domains

Three pillars for semantic interoperability:

         data dictionaries and metadata registries, including names of
          observations (LOINC) – this should consist of a simple list of data
          elements and an integrated repository

         clinical templates – to aggregate data elements from the repository (or
          from the RIM) into meaningful fragments (i.e., batteries of lab tests,
          blood pressure (systolic + diastolic) + circumstances (patient’s position,
          device), pathways in USAM-RIM

         tables with enumerated value domains – there are hundreds of explicit
          tables with enumerated value domains


         Draft data dictionaries and metadata registries including names of
          observations (LOINC)

         Clinical templates

         Tables with enumerated value domains’

      Combining the three bullets above will help achieve interoperability at the
      semantic level. He sees this as an extension of the RIM – each data
      element should be explicitly registered as a child/refinement of a RIM
      class, under control of the respective HL7 technical committee. The
      registry would be a means of mapping data elements that are outside of
      the RIM to a particular class within the RIM. For example, blood pressure
      might be mapped to the class Observation.

      Angelo will provide a written report at the next meeting. We would also like
      a report on Metadata registry and how it relates to clinical data templates.

      Click here to view the entire presentation

IV.   Status reports

         GCPR framework project – Nancy Orvis reported that the project is
          currently in phase II. The prototype in a lab environment was
          successfully demonstrated in early April. Phase II, scheduled for roll
          out in Alaska, will involve a test or mirror database at some sites and
          information to communicate lab results, patient demographics and
          outpatient medications.

         United States Health Information Knowledgebase (USHIK) – Glen
          Sperle reported that this project’s goal was to build and populate a
          database across multiple organizations. It originated as a response to
          HIPAA legislation, is based on the ISO/IEC 11170 metadata standard
          and utilizes the Australian high-level model. HCFA has offered the
          registry to the HISB and SDOs to utilize for a year to determine its
          benefits. Seven SDOs have accepted: X12, ADA, ASTM E31, HIBCC,
          HL7, IEEE 1073 and NCPDP. Five additional organizations have
          agreed to submit their data elements, including the CDC, DoD,
          NCVHS, GCPR and HCFA (for ESRD). The Initial project startup
          involves loading metadata from dictionaries. Next, HCFA will be
          looking at how to incorporate the registry into the SDOs’ operations.
          HISB will assume management of the registry.

Click here to view the entire presentation

   Denise Koo reported that the US government is negotiating with
    SNOMED for a license. NLM is taking the lead on this project. They
    don’t want SNOMED to be a requirement for which there is a
    requirement of payment for each report to the government. This is an
    HHS wide initiative.

   Denise Koo also reported that the NVCHS will be celebrating their 50 th
    anniversary on June 20 in Washington, DC.

Agenda for St. Louis Meeting:

   Follow-up issues (next steps) from the Plenary programs (possible
    participation in plenary on version 3)

   Status reports

   Update from clinical templates

   Update from Suze Burke-Beebe (consortium)

   Update from Public Health Conceptual model/RIM Harmonization

Meeting preferences for St. Louis

Wednesday AM (or some other time that does not conflict with clinical


Shared By: