Attendees:
Harold Solbrig, Ali Rashidee, Anita Benson, Hiroyuki Furukawa, Mead Walker, Phil Taggart, Smita Hastak, Lise Stevens, Bob Hizer, Susan Bounds, Elizabeth Smith, Paul Frosdick, Gene Ginther, Ted Kline, David Markwell, Clive Flashman, Chris Foye, Julie James, Stan Huff, Tom Oniki, Kent Spackman, Chris Chute, Carmela Couderc, Marguerite Raaen, David Eide, Dan Jernigan.
Discussion:
The meting began with a brief presentation on terminology challenges and a discussion of HL7 internal terminology challenges. There are 70 coded attributes that need vocabulary or code sets:
· Discussed options for binding external vocabularies to RMIM attributes. The UK wants to specify Realm-specific vocabularies. Patient Safety sees two vocabularies, depending on the context (regulated or research). Ted Klein and Chris Chute indicated that it was reasonable to allow a late binding in both cases. There are issues of when and how this binding will be specified in the tools.
· Discussed the options of using LOINC or internal HL7 code sets for observations in the ICSR that record summaries or judgments (e.g., InterventionCharacteristic = suspect, interacting or concomitant). Stan Huff pointed out that with LOINC observations if the value in an observation is a code, then that has to have a value set in some vocabulary.
· Harold Solbrig indicated that he would help Patient Safety prepare its submissions
The terminology practices for HL7 purposes will probably evolve to some international and some specialized terminologies.
FDA stated that this version of the message has a specific use case for regulated reporting, which has certain requirements that are not intended to capture a very broad a patient safety message, such as acts of omission. ICSR is harmonized with ICH E2BM standards for global adverse event reports to regulatory authorities in US, Europe and Japan. Future releases of the model were discussed, then the group agreed that additional releases of the message will expand its use cases for broader patient safety requirements. ICSR may not be able to accommodate all patient safety requirements, and that other messages may need to be developed for other specific use cases. PSSIG asked Vocabulary TC for their recommendations to move the ballot forward.
The NIH National Cancer Institute (NCI) intends to be an early adopter of the message. They would like to implement messaging in five months and would like group to define the value set used for reporting to the FDA. Group agreed to consider Vocabulary TC’s recommendations and the meeting was adjourned.
Conclusions:
Vocabulary TC Co-Chair suggested that PSSIG:
· Publish normative standard that addresses the specific use case, domain and realm attributes
· Publish an HL7 implementation guide that should include examples of message using realm specific/pre-defined code sets. If required, code sets should be registered with HL7. Clarify code variable names
· Separate our use cases: regulated reporting, clinical trial, patient safety, other and provide instance examples for each message
The members generally agreed that an ICSR Implementation Guide is needed. Ideally, it will be flexibile, the code set late binding, and register them. The Vocabulary TC would like to see the list of act codes being recommended.
5/3/04, MONDAY Q2
Attendees: Ali Rashidee, Anita Benson, Hiroyuki Furukawa, Mead Walker, Phil Taggart, Smita Hastak, Lise Stevens, Susan Bounds, Marguerite Raaen, Clive Flashman, Chris Foye, Julie James.
Discussion:
National Patient Safety Agency, UK (NPSA):
Chris Foye presented an overview of NPSA’s Patient Safety Focus. The presentation is being posted on the PSSIG site.
· Emphasis on learning from patient safety incidence, and talked about the National Reporting and Learning System
· Showed us some reporting capability using SAS and Autonomy
· Described patient de-identification process, which allows duplicates because of anonymous reporting. Discussed the possibility of a “black box” where patient identification is known and duplicates are discovered.
Conclusions:
Basic scope: Reporting to national entity in the interest of patient safety
Action items:
Chris Foye will define an HL7 Scope, set priorities, and model the priority areas. This work will be presented in the next meeting at Atlanta.
5/4/04, Tuesday Q1
Attendees: JOINT with RCRIM
Discussion:
Group agreed to classify ICSR ballot comments into 4 categories:
§ Message Scope
§ Vocabulary
§ Structural
§ Miscellaneous Questions
Comments from Eli Lilly & CDISC
§ Agreed to add “and the manufacturer” to text describing where the HCP should report incidentAgreed to add more descriptive comments to the model
§ Agreed to collaborate with OO to capture challenge and de-challenge
§ Agreed to rename pertinentRelatedAE Investigation to pertinentRelated ICSR Investigation
Comments from Ali Rashidee AHRQ
§ Agreed to reword ballot narrative and remove all references to specific government agencies and entities that may be interested in using the message§ Accepted comments concerning harmonization requirements and continue collaborations with Medications and OO. Agreed that realm-specific guides may be necessary because more time is required to create a global vocabulary that everyone would approve in time for implementation of the message.
Comments from Rob Borotkanics AHRQ
§ Agreed to remove reference to MedDRA for coding the adverse event term. However, FDA implementation guide will require the use of MedDRA for regulated industry and reporting between regulatory authorities to harmonize with existing ICH guidelines§ Requested that Rob submit an additional story board to accommodate use of the standard for non-regulatory reporting.
Comments from UK NPSA
§ NPSA Comments 1 & 2: Agreed that the scope of release 1 of ICSR is geared toward regulated reporting. Agreed to have more detailed discussion within PSSIG to define the scope of the group and development of messages to support the scope.§ NPSA Comment 3: Agreed that model currently support anonymous reporting b/c the fields are optional. However, 1 CMET is requiring a RAssignedEntity code. Agreed to find another CMET that does not require code
§ NPSA Comment 4: Agreed that realm-specific implementation guides could be developed to address coding and vocabulary required for NPSA, therefore no new codes would need to be added to the HL7 Act class codes.
§ NPSA Comment 5: Determined comment concerning “condition” as not related. Concept of near miss will be addressed in release 2 of the message
§ NPSA Comments 6-11: Agreed that these comments relate to scope discussion and would be accommodated in release 2 of the message
§ NPSA Comment 12 – 14: Agreed that these issues are covered in the existing model, but may need to clarify/make more obvious
§ NPSA 15-17: Related to scope and version 2 release
Comments from Virginia Lorenzi
§ Agreed to update narrative to include examples of each use case for all interactions reflected in the model§ Agreed to add class of the transition to State Transition field on Trigger Event
The members agreed to schedule conference calls to resolve the remaining comments.
5/4/04, Tuesday Q2
Attendees: JOINT with O & O
Discussion:
Objective is to see how well the ICSR model interdigitates with O&O models.
BLOOD BANKING MODEL:
Diana Lopez presented the blood model that looks generally good, but needs more detail and review by Subject matter Experts.
The Blood AE/error model was presented. Transportation issues for blood, medication, etc. that may lead to error needs to be considered. OO chair indicated that they are working on harmonizing clinical statements and wants to make sure we would incorporate any changes into our model.
Issues that surfaced during the discussion:
· Blood, tissue, organs- can these be merged?- “Transfusion/Transplant Model”
· How to accommodate for pre-donation-transplantation test, and the compatibility testing on a donor-recipient set
· Specimens for the blood tests are taken at the same time as the blood donation, but they are separate specimens. The results of the tests don’t come back for a couple of days.
· Transportation issues for blood, medication, etc. may lead to errors, so needs to be considered. Also to think about Suspect, interacting and concomitant medications, which may, or may not, include challenge-rechallange.
Meet again in January to address remaining areas of overlap (Infections and Pathogens).
General notes from model discussion:
§ Sibling information is captured in supporting clinical information CMET
§ Need to add performer to the CMET
§ Is blood storage and transportation (distribution) info covered
§ Add act for compatibility testing (pre-transfusion)
§ OO wants clarification on what types of tests are performed: specimen vs. donor/recipient matching tests – capture all relevant tests (pre-donation transplantation tests, compatibility matching, HIV, etc)
§ Remove constraint for volume and add ID
§ Add relationships in model
The Current Clinical information model is undergoing changes – see if these can be derived from that model.
5/3/04, Tuesday Q3: Cancelled
5/4/04, Tuesday Q4
Attendees: Medication SIG hosts PSSIG
Discussion:
Medications Co-Chair informed PSSIG group that the ICSR modeler should be at the discussion. Mead Walker was not able to attend this session, but Anita Benson was able to answer some questions
§ ICSR should refer to ballot 7 version of the medications model. Suggested that PSSIG review Pharmacy domain discussion about public health domain and use of the CMET.
§ InterventionCharacterization is used for suspect, concomitant, and interacting medications. New requirement to capture Re-challenge is needed
§ Consider registering NDC codes or other FDA specific drug or vaccine codes with HL7
§ Question about “Labeled strength” of drug. Should it be explicitly modeled or is it part of the name?
Medication model is evolving to be a separate ballotable model, like general purpose CMET for ‘Supporting Clinical Information’
Action items:
Patient Safety will evaluate model in ballot and change as necessary. Med Info/Pharmacy SIG will maintain resulting CMET.
5/5/04, Wednesday Q1
Attendees: Susan Bounds, Marguerite Raaen, Douglas Fridsma, Ali Rashidee, Clive Flashman, Chris Foye, Helen Stevens, Dan Jernigan, Mead Walker, Hiroyuki Furukawa, Phil Taggart, Anita Benson, Joan Miller.
Discussion:
The University of Victoria present a student project, “Noscomial ‘Hospital Acquired’ Infections, surgical Site Infections, an International Approach”
This diagram presented by Helen will have to validate with the RIM
Primary act classes
Procedure
Readmission
Pathogen observation
Class code
Needs Domain identifier
The project is not a message RMIM at this point, but a model can be a basis for a RMIM.
Mead Walker suggested that the model needs to be abstracted to a higher level.
Change the story board names to HL7 published guidelines.
In the future ,early collaboration with the Subject Matter Experts and other stakeholders should take place.
Conclusions:
Future: Nosocomial in PS Domain, How finely grained should it be depends on input from the stakeholders..
Send Letter to the Helen Stevens and the students who worked on the model
Include CDC in future discussions.
5/5/04, Wednesday Q2
Attendees: Ali Rashidee, Anita Benson, Marguerite Raaen, Douglas Fridsma, Clive Flashman, Chris Foye, Smita Hastak, Hiroyuki Furukawa, Phil Taggart, Anita Benson, Bob Hizer, Joan Miller, Susan Bounds.
Discussion: PSSIG Scope
Current ICSR message supports regulated reporting of drugs, devices and vaccines. Designed to support PhRMA requirements for ICH E2B M2 ICSR message specifications, and reporting of adverse events to FDA using the MedWatch 3500A (drugs, devices and therapeutic biologics), and VAERS-1 for vaccines. FDA would prepare future releases of the message to support:
· eRADHealth
· Foods and Cosmetics
· Blood/Tissues
· Animal drugs (individual vs. herd reports)
Challenges and collaborations:
· Integrate with OO’s modeling of BTO (Blood, Tissues, Organs)
· Integrate changes to Medications model
NPSA, UK: Review use case requirements for ICSR
§ Includes all care settings
§ Gap between what was reported and what happened
§ No harm, near misses (includes reports with or without a patient), and group reports that may not
have a lot of details included in the report
§ Incident, alarms, events including death (includes staff involved in incident)
§ High level contributory factors
§ Need to accommodate free texts: actions taken, descriptions
§ Anonymous reporting: reporters, patients, staff: Change one CMET to enable this.
§ Expanded current model or develop a new model for notification reports of patient safety incidents
Future NPSA reporting requirements also include:
§ Reports from patients, or their families, and the general public
§ Provide guidance to local jurisdiction
§ RCA Reports by Health Care Practitioners: Inbound and outbound messaging
§ Health Care Practitioner Reporting Practices (Best Practices Messaging)
NCI:
Reviewed the use case requirements for research and clinical trials: NCI lead discussion about reporting requirements between NCI cancer centers and the FDA. Indicated that there needs to be a way to capture patient roles that differentiate between “patient status” and “research participants”. Other areas for new model development include:
§ Capture long term studies (+20 yrs). NCI wants to report from prevention-quality of life studies.
§ Track types of foods eaten during study, or capture diet as an intervention or substance administration
§ Combined analysis of different types of product areas: classes of drugs, devices (including battery of tests) – How will these be reported?
Other future requirements:
§ Clinical trial modeling of a medication
§ Severity coding (vocabulary): MedDRA vs. CTC
§ Development of use cases for cancer trials
Group brainstormed on the specific areas that should be covered under the PSSIG, and the need to develop standard definitions for:
§ Events, Alerts, Incidents
§ Decision Support: How will our messages integrate with these types of systems?
§ Clinical trial reporting: regulated vs. public health reporting from Health care Professionals (HCP)
§ Modeling of products vs. therapies
§ Need to develop a list or reporting requirements for HCP
Group was in agreement that we needed a work plan. Topics agreed to be discuss at the next meeting:
§ Define the scope of events, alerts, incidents and other sentinel signals (clinical vs. non clinical environment) – What should we cover and how do we support with messages: Chris Foye, NPSA UK, will prepare a draft for discussion
§ Define who our stakeholders are and what’s the best way to support them
§ Vocabulary harmonization
§ Ongoing ICSR development: Address challenges of the current model: expanded scope and size of model and how to work with existing tools for implementation