USUBJID

From PHUSE Wiki
Jump to: navigation, search


Sub Project Overview

Goal: Develop best practices/guidance for USUBJID

How do we ensure a consistent single USUBJID for a John Doe within an application where the John Doe may be a re-screened subject, and/or participated in multiple studies in an application?

Project Leads

Name Role Organization E-mail
Mark Sullivan Industry Co-Lead PAREXEL Informatics mark.sullivan@parexel.com
Monica Mattson Industry Co-Lead Celgene Corporation mmattson@celgene.com
Lisa Lin FDA Liaison FDA Wei.Lin@fda.hhs.gov
Mina Hohlen FDA Liaison FDA mina.hohlen@fda.hhs.gov

Project Updates

Objectives and Timelines

Objective Timeline

Project Content and Activities

Meeting Minutes March 17,18 2014

Date: March 17-18, 2014

SUBTEAM: USUBJID

MEMBERS: Joanna Koft (joanna.koft@biogenidec.com); Lisa Lin (wei.lin@fda.hhs.gov), Lisa Pacelli (lisa.pacelli@bms.com), Majdoub Haloui (majdoub.haloui@biogenidec.com), Matthias Boerner (matthias.boerner@bayer.com), Michael Rubison (michael.rubison@capishknowledge.com), Sterling Hardy (Sterling.Hardy@bms.com), Mina Hohlen (mina.hohlen@fda.hhs.gov), Paula Baker (paula.baker@bms.com), Ratnakar Gundlapalli (ratnakar.gundlapalli@iconplc.com), Shannon Labout (slabout@cdisc.org), Snehal Patel (snehal.patel@iconplc.com), Fred Wood (f.wood@accenture.com), Gail Stoner (gstoner@its.jnj.com), Douglas Warfield (douglas.warfield@fda.hhs.gov), Mark Sullivan (Mark.Sullivan@parexel.com), Monica Mattson (mmattson@celgene.com), Christine Fenaughty (Christine.c.fenaughty@pfizer.com)


LOGISITICS:

Leaders: Monica Mattson, Mark Sullivan Meeting time: Fridays, every 2 wks, 9-10 am EST Conferencing: WEBEX Minutes: Rotate Meeting Minutes Google doc for minutes (Majdoub Haloui) Best practice document due date: 4-5 first draft (writer TBD)

DISCUSSION:

• USUBJID = Study id/Site id/Subject id

• USUBJID issue – discussion with FDA • Goal: Consistent single id for John/Jane Doe across studies in an application - Want to know when a subject enters more than once in an “application” • Alot of variability in USUBJID implementation across companies • USUBJID is only required within an application, not across applications • Duplicates found within an application - Open CDISC will fail DM if duplicates in datasets e.g. when you go into an extension study, you lose uniqueness of subject. Some companies have taken ID in study 1 and used in study 2 - the dups will fail validation. • How do we fix the problem? Can we have traceability with the solution? • Is this a CDISC problem or can we decide this solution can be outside of CDISC?

• Application vs. Submission • Pharma tends to use these terminologies interchangeably • FDA confirmed what “application” means – it is not equivalent to submission. Application refers to the whole product e.g. NDA, INDA, not deliverables or subcomponents comprising the applications

• In Scope scenarios for USUBJID:

1. “In Study” Duplicates: • Screen Failures/rescreens • Randomized but not dosed and then re-enrolled • Run-In Failure: includes long run-in studies where randomization happens later in study (e.g. 6 wks out) • Switching sites (“snow birds”) • Shop around for ‘active” drug • Relocated

2. “Across Studies” duplicates: • Sponsor/Partner/CRO conducting multiple studies • Studies integrated for ISS • Extension studies with new database and new protocol • Post marketing commitment • Cross study/cross sponsor • Virtual clinical trial • Registry study (do these get submitted as part of application?)

• Out of scope for USUBJID: • Databases not controlled by sponsors – data is used but outside control of sponsor • Subjects shopping around – may not be truthful across study sites • Scientific Data Repository (need to address if this is in scope) - “across applications”

• Current USUBJID practices across companies: • CRO perspective: • can handle rescreens in some cases but not in EDC • Needs to be an EDC solution • Capture prior SUBJID in same site scenario • Celgene: • In study: Use numbering of subject to identify rescreens • John Doe : 1001 but fails screening • John Doe: 2001is rescreened and entered into study • Across study: Standard CRF collects “previous study” information • Study no and SUBJID collected • BMS: • Captures only BMS trial info • Requires investigator follow-up • Non-patient database • TA level database with demographics • Assigned a USUBID • Other : • Non-patient (AP domain) data domain (found in IG) used to track screen failures. Use AP as a link back to actual data – SVAL can be used • Some companies use “SCR” for screen failures after USUBJID (drop “SCR” for analysis) • Use language on CRF that states “first” study you were in • Programmatically look at all subject id’s to create USUBJID • Assign a USIBJID upfront when expect a subject might participate in multiple studies

• SUBJID – why is this variable not on all Domains? • General consensus that SUBJID should be • Unique id within study • When multiple CROs are used using own systems and processes, may lose uniqueness of a subject

ACTION: ALL • Cut and paste your company solutions for USUBJID in a document for discussion for next meeting on APRIL 4, 104 9-10, EST • All communications to members should include in subject line: PHUSE USUBJID WORKING GROUP:

File:USUBJID Project.pdf

Meeting Minutes April 4 2014

USUBJID Working Group 2014-04-04 Meeting Minutes

Agenda

Writer for white paper: [missed name]

Wiki assignment: Mike Rubison, Majdoub Haloui

Google doc setup: Majdoub Haloui

Roll Call

AI: New members: New members are asked to send email to Monica and Mark today, next week at the latest.

Review of main goal: Develop best practices/guidance for USUBJID

Discussion of requirements from FDA Technical Conformance Guide and 3.2


Team noted that the 3.2 SDTM IG (Section 4.1.2.3) includes verbiage that might indicate the scope is not across a submission, but a product: To identify a subject uniquely across all studies for all applications or submissions involving the product, a unique identifier (USUBJID) should be assigned and included in all datasets.

Monica - Do we interpret (or try to handle) that unique should be across submission, or product, or all trials?

Lisa - Doug defined scope in previous meeting as NDA. But not same across multiple applications for a product

Mina – Not clear FDA internally agrees on what they need, or industry perspectives.

Mark – What is the frequency of occurrence? Perception is that this does not happen that often, but would like to know how common this is.

Lisa – BMS has been tracking and this does seem to occur often. Rollover to long term is common and tracked.

Mina – What is the definition of roll-over?

Lisa indicated that there is not much difference between different types of studies. The same subject is the same subject, don’t get stuck on the differences.

AI - Mark and Monica come up with definition of roll-over vs extension

Review of “re-screening within the same protocol” vs “cross-study scenarios”

Review of Glossary

AI: Mark & Monica – enhance glossary

Review of methods to maintain unique identifiers for subjects

Discussion of method proposed by Fred Wood, previously agreed to by the FDA in two instances

AI: All - Submit comments on the method

AI: All - Request from others how they are handling this

AI: Mina – Internally looking to clarify SUBJID and USUBJID

Example of implementing plan discussed so people can visualize:

USUBJID SUBJID RFICDTC RFPENDTC ARMCD
DM 001-001 001 2013-06-30 2013-06-30 SCRNFAIL
DM 001-001 101 2013-12-15 2013-12-16 SCRNFAIL
DM 001-001 201 2014-01-01 2014-04-10 ARM-1
USUBJID VSSEQ VSTESTCD VSDTC
VS 001-001 1 SYSBP 2013-06-30
VS 001-001 2 DIABP 2013-06-30
VS 001-001 20 SYSBP 2013-12-16
VS 001-001 21 DIABP 2013-12-16
VS 001-001 50 SYSBP 2014-01-01
VS 001-001 51 DIABP 2014-01-01
IDVAR IDVARVAL QNAM QVAL
SUPPVS VSSEQ 1 SUBJID 001
SUPPVS VSSEQ 2 SUBJID 001
SUPPVS VSSEQ 20 SUBJID 101
SUPPVS VSSEQ 21 SUBJID 101
SUPPVS VSSEQ 50 SUBJID 201
SUPPVS VSSEQ 51 SUBJID 201

File:2014-04-04 USUBJID meeting slide deck.pdf

Meeting Minutes June 13 2014

USUBJID Working Group 2014-06-13 Meeting Minutes

  • Should we have a single “Best Practice” recommendation? General consensus/agreement is Yes.
  • Other examples can be included in an Appendix or addendum as “considered but rejected” (or similar)
  • Confirmed that focus of Best Practice is a unique USUBJID within an application (not cross applications)
  • NOTE: This was addressed in an early draft by of the White Paper by Richard Lewis, revisit including this in the current version.
  • Comment: this effort should be driven by the FDA perspective, and we don’t have that (because the FDA does not have it yet). Possible that the FDA’s perspective could be available by early July.
  • Discussed Scenario 1 (single record in DM for each USUBJID) versus Scenario 2 (allow multiple records in DM for each USUBJID, and key off SUBJID).
  • Advantages of Scenario 1
  • Fits best with current expectations (based on language in documentation, available tools, etc.)
  • Helps with development of ADSL
  • Adheres to expectation that USUBJID is key to linking to all other domains
  • Advantages of Scenario 2
  • Used successfully in at least two submissions
  • Easier to track data for an individual in an unambiguous manner, even for unusual cases such as a protocol allowing/expecting multiple valid enrollments/subject
  • Discussed putting SUBJID as a required Identifier variable in all General Observation Class domains (to facilitate tracking back to correct information in the DM domain). Noted that SUBJID would have to be defined as SITEID || SUBJID to ensure uniqueness within a study.

Action Items

  • Action: Monica -- Include confirmation that Best Practice is focused on unique USUBJID within an application (revisit verbiage from Richard’s early draft of the White Paper).
  • Action: Mina -- Seek and communicate FDA perspective on USUBJID, target July 3.
  • Action: ALL – review the current draft of the White Paper and provide feedback to Monica by COB on Thursday, June 19. Thanks to those of you who have already responded.

Timelines Mark and Monica spoke following the meeting and came up with the following planned timelines:

  • June 19: Comments on current draft of white paper due to Monica
  • June 27 currently-scheduled meeting of the team will be cancelled
  • June 30: Monica to incorporate comments received, send to Writers
  • July 3: Update from Mina on whether FDA perspective on USUBJID is available
  • July 9: Writers to meet and discuss draft v2
  • July 11: Full-team meeting to discuss status, next steps

Meeting Minutes July 25 2014

Date: July 25, 2014
Near-term schedule

High-level plan is to have each PhUSE Working Group review the draft White Papers of other Working Groups.
August 6: next comments due to Monica and Mark on USUBJID White Paper draft v2.
POST-MEETING NOTE: In a communication with the heads of the Best Practices teams, the plan is to submit the next drafts to other Working Groups on August 8.
We will need a couple of days to incorporate comments, so the request is to have comments submitted by COB on Wednesday 06-Aug.

Based on feedback to come up with a single “Best Practice”, the White Paper draft v2 has been structured to describe the “Best Practice” as a solution that adheres to current tools/interpretations (identified as solution 1B in the first draft). An alternative solution (solution 2B in the first draft) is presented as a solution that is more descriptive but would require changes to documentation/software to implement in a widespread fashion.

Rationale for including second option:
Keeps data in the correct variables
Reduces the number of SUPPDM records
Better defines relatedness of data values (as there is no linking within SUPPQUAL of related records)

Acknowledged that there has been a general shift within the industry to collect data on every enrollment (which historically has not always been the case) Current validation tools at FDA check for 1 record per USUBJID in DM. Second option may be the better solution for ISS datasets

Current validation tools check that ADSL dataset has only 1 record per USUBJID
For any subjects enrolled in multiple studies, if those studies are integrated, then there will be multiple USUBJID values.
If validation tools could be changed to look at both USUBJID and SUBJID (concatenated) instead of just USUBJID, and there would be better assurance that the values would be unique.

Near-term schedule

High-level plan is to have each PhUSE Working Group review the draft White Papers of other Working Groups.
August 6: next comments due to Monica and Mark on USUBJID White Paper draft v2.
POST-MEETING NOTE: In a communication with the heads of the Best Practices teams, the plan is to submit the next drafts to other Working Groups on August 8.
We will need a couple of days to incorporate comments, so the request is to have comments submitted by COB on Wednesday 06-Aug.

Based on feedback to come up with a single “Best Practice”, the White Paper draft v2 has been structured to describe the “Best Practice” as a solution that adheres to current tools/interpretations (identified as solution 1B in the first draft). An alternative solution (solution 2B in the first draft) is presented as a solution that is more descriptive but would require changes to documentation/software to implement in a widespread fashion.

Rationale for including second option:
Keeps data in the correct variables
Reduces the number of SUPPDM records
Better defines relatedness of data values (as there is no linking within SUPPQUAL of related records)

Acknowledged that there has been a general shift within the industry to collect data on every enrollment (which historically has not always been the case) Current validation tools at FDA check for 1 record per USUBJID in DM. Second option may be the better solution for ISS datasets

Current validation tools check that ADSL dataset has only 1 record per USUBJID
For any subjects enrolled in multiple studies, if those studies are integrated, then there will be multiple USUBJID values.
If validation tools could be changed to look at both USUBJID and SUBJID (concatenated) instead of just USUBJID, and there would be better assurance that the values would be unique.

Meeting Minutes August 8 2014

USUBJID Working Group 2014-08-08 Meeting Minutes

Current Status as of February 8 2015

(1) This working group received the comments on the USUBJID White Paper from the PhUSE Steering Committee (see below). Comments from Wayne Kubick are included as well. Monica will schedule a meeting of the team within the next couple of weeks to discuss edits to the paper required as a result of the comments.

(2) After the comments have been incorporated, the working group will post the White Paper to the PhUSE Wiki for a public comment period.

Also we have been informed that the CDISC SDS Leadership Team will be reviewing the White Paper to provide their comments/thoughts. This may not be compiled anytime soon; as most of them will be at the PhUSE CSS, those comments will likely be discussed during the best practices session.

As an aside, there is a CDISC Members-Only webinar on the USUBJID topic scheduled for Thursday, 12-Feb. Monica Mattson will be co-presenting with Mark Sullivan and Gary Walker. The information presented will follow the recommendations of the paper, emphasizing that this is the recommendation of a PhUSE Working Group, and not an approved CDISC approach. Wayne Kubick and Fred Wood have been asked to join the Q&A Panel at the end of the webinar to be able to respond to any questions directed to the SDS Leadership team.

Archived Content