May 22
May 22 QRPH Face-to-Face Meeting Minutes
Agenda
- Floyd Eisenberg (Siemens)
- Jason Colquitt (Greenway Medical Systems)
- Landen Bain (CDISC)
- Harry Solomon (GE)
- Daemon Whittenberg (Greenway Medical Systems)
- Jacob Reider (Misys)
- Joann Larson (Kaiser)
- Patty Craig (The Joint Commission)
- Thom Kuhn (ACP)
- Chad Bennett (IFMC)
- Michael Ibara (Pfizer)
- Delane Heldt (AMA)
- Judy Logan (OHSU)
- George Cole (Quintiles)
- Clinical Research Review
- Updated White Paper with content based on discussions of May 19-21 inclusive: White paper for review May 22 AM
Pending Paper updated during the worksession and uploaded to the ftp site (Link above).
Drug Safety Reporting
Michael Ibara (Pfizer) Review: Aster – Project at Partners Data – real world experience Metadata – spontaneous safety falls between different initiatives ADE Spontaneous Triggered Event Reporting (ASTER) All post-market reporting DIA Conference 2 years ago – RFD to test model ASTER to implement automated ADE collection in ambulatory clinic setting using the longitudinal medical record in ambulatory clinics at Partners CRIX – not for profit organization – will help host form and process information
Data from LMR – ICSR2 format and ICH (International Council of Harmonization) E2B Standard format as well
Spontaneous reports are not submitted due to convenience and effort and understanding When reporting, the Pharmaceutical company becomes the “middle man” as FDA is not generally set up to receive. ASTER plans to send the report directly from the LMR to FDA.
a. Physicians don’t see events as adverse events b. Physicians who recognize the events are not at a location to report c. Data information is sparse
Sponsor – Dr. David Bates, Dr. Jeffery Linder; Norman Marks (MedWatch at FDA) strong supporter Live pilot 4th quarter – September hopefully Plan to change timing for reporting from several days (regulatory requirement for manufacturers of 15 day report) to minutes. Workflow: • Physician in ambulatory clinic • Patient seen with severe nausea due to antibiotic and antibiotic discontinued • Discontinuation includes a reason captured as an adverse event (currently within the Brigham LMR) • Trigger of reason triggers standard RFD transaction sends notice to CRIX – LMR fields are mapped to E2B so that the physician sees returned a partially filled form similar to a MedWatch form • CRIX populates the form from the LMR and passes it back to the physician • The physician must interact with the form - Must for “outcome” section, rest can be added • Maximum of 120 sec, average 60 seconds • Submit – information arrives at CRIX • CRIX codes and matches to “outcome” and DME – designated medical events • Attempt to triage serious events • Mapping in ICH E2B format, and ICSR2 format and send to respective gateways.
MedWatch from - approximately 15 fields to map to E2B. The rest (lab, etc) is identifiable more easily. More complex are items such as definition of "adverse events." Requires subject matter expertise on the safety side. Also expertise in the structure and the data model of the EHR.
Should be flexible to incorporate other elements as they are required. The effort is being centralized from current efforts among manufacturers, removing the need for identical efforts in each manufacturer. The effort also offloads the work from the FDA.
Significant relationships and agreements are required as well as Safety Knowledge expertise. Such expertise is not necessarily present in existing HIEs. The design requires safety expertise for the set up and initiation. It creates a flexible model for reuse. The minimal difference between serious and non-serious is:
- outcome - needing hospitalization or not
- coding of the event
With these two items, can determine seriousness In the ambulatory setting outcome is determined before the patient is in the clinic. In hospital adverse events may be identified during the event and the outcome will not be available when first identified. Outcome as a regulatory trigger is generally known in the ambulatory setting. Some follow up may still be required for clinical outcomes.
A bi-directional communication could allow a business rule to allow a request for a 3-day or 5-day update. This process is somewhat aligned with a case management model.
Reporting of Success: Currently have the pre-intervention number and distribution of adverse event identification as part of the routine medication discontinuation process. The pilot will begin with the more robust reporting physicians to be able to monitor post-intervention success.
- Does this improve the quality report - requires agreement as to what represents a "quality report"
- What new issues will be created - there are 3 categories of information:
- That information requested by safety (the ICH fields)
- That information requested that cannot be found in the EHR (e.g., outcome)
- Information contained in the EHR that can add value to a safety report yet has not previously been considered. - Will represent an additional requirement (potentially) for future reports.
Privacy and security - the LMR is encrypting the identifier - local pseudonymization.
Question about supplemental information - Decision support capabilities might request sending a medical summary to add further information for analysis. (Discussion from PCC with respect to immunization May 21).
A Drug Safety Profile will require a trigger. QED uses templates. To use a medication template, re-use the appropriate structure and package. The profile also requires that data enters into the form is expected within the EHR. In the Clinical Research Profile, the data elements may not be desired within the EHR. In the Drug the data elements generally are the same as those within the EHR. The report itself, however, would not be saved within the EHR for risk management reasons.
- The trigger events specifically are based on local implementations (e.g., an order discontinuation, an entry to a problem list, a new entry to an allergy / adverse event list). A trigger is the initiation point for the Drug Safety Profile.
- The data source for the form filler requires discussion.
- Total number of data elements that required analysis = 15, most were one-to-one maps.