Group 2 – Assess and Improve Existing Validation Rules (Amy)

From PHUSE Wiki
Jump to: navigation, search

Group 2 – Assess and Improve Existing Validation Rules (Amy) When: Monday, March 19th
Place: Silver Spring Civic Center | Sliver Spring, Maryland, USA
Facilitator: Amy Malla

Name Organization
Amy Malla FDA/CDER (FDA Co-Lead)
Hany Aboutaleb Biogen Idec (Industry Co-lead)
Max Kanevsky OpenCDISC (Industry Co-lead)
Akira Soma Mitsubishi Tanabe Pharma Development America, Inc.
Aldo Schepers Novartis Pharma
Anne Russotto Hoffman-La Roche
Carina Soldner Boehringer Ingelheim
Carol Vaughn Sanofi Pharma
Dhanajay Chatre FDA
Doug Robinson SAS/JMP
Justin Steele Clinical Trials & Surveys Corp
Michael Rippin Clinical Trials & Surveys Corp
Michael Whitworth Quanticate
Muntaha Walp FDA-DTS/CSC
Samuel Ogunbo FDA
Sharon Powell Eli Lilly and Company
Tatiana Sotingco Janssen Pharma
Todd Miller Novartis
Sergey Sirichenko Pinnacle 21

General discussion

Group 2 – Assess and Improve Existing Validation Rules (Amy)

  • Discuss general issues with existing validation rules
o What are common and recurring issues?
o What does it mean to pass or fail (not black and white)?
o What are the main contributing factors?
1. Bad or incorrectly implemented rule
2. Unclear or conflicting underlying standards
3. Lack of understanding
  • Review and evaluate Top 20 SDTM and Define.xml issues from CBER/CDER
o Review metrics from CBER and CDER
o Evaluate each rule to determine contributing factors
o Propose changes to rule definition or underlying standards

Summarize group findings

  • Topic: Additional Areas of Focus for the Working Group
o FDA collaboration with Health Canada, EU, etc.
o Validation checks for earlier in the process EDC/CDASH
o Implementation/business rules such as timing variables that impact quality or ability to stack (WG3)
o Definition of concepts change based on study or study design, how to address this issue
o Integration with CAB project
o Integration with CDER Validation project
o Common ways to respond to rule issues
  • Action Items
o Group will prioritize when existing projects are completed
o Based on prioritization new projects will be initiated.
  • Topic: Top 20 Rule Failures
o SKIP_SD0006:
o This is extraneous information and should be hidden from the user
o SD0006: No baseline result
1. SDTM requires baseline for EG, LB, QS, VS, not all TEST have a baseline
2. Validation rule and standard needs to align with use cases
3. Baseline results is defined per protocol
o DD0010-ODM attribute should not be included in define.xml. ODM attributes that are optional and are not explicitly specified in define.xml should not be included.
1. Investigate potential user errors
2. Investigate CDISC documentation
o CT0037: Value for –BODSYS not found in SOC controlled terminology codelist
1. Case sensitivity mis-match between EVS/MedDRA/OpenCDISC
2. Case sensitivity an issue for analysis of mixed with study or across studies.
3. Check to see if MedDRA is available either in mixed or only upper case.
o SD0029: Missing units on value
1. Not all tests have units
2. Identify all tests with no units
3. Validation rule and standard need to align with use cases
o SD0031:Start date expected when end date provided
1. Imputation of dates for ISO8601 results in errors
2. Information relating to rules for imputing date for ISO8601 is difficult to locate and implement
3. User errors, new year and year is not increased
4. Investigate the domains that this error is occurring in with greatest frequency (AM, CM, EX)
5. Obtain understanding of how this check fires
o SD0081: Observation date must be less than or equal to latest disposition date (EG, LB, VS)
1. Protocol design: no last contact date. Can you create a record in DS for last subject visit?
2. Disposition incomplete – bad data. Happens in AE with deaths. Last record in DS may need to be death.
3. Review codelists to determine if what options are available and if additional choices should be added
4. May need business rule or additional information in IG
o DD0024: Invalid Codelist in define
1. OpenCDISC codelists are obsolete
2. Extra terms/codelist extensions are not added to the define.xml
3. User error can be corrected, need additional directions for Sponsors.
o SD0026: Missing Units (Original)
1. Same as SD0029
o SD0080: AE start date must be less than or equal to latest disposition
1. May be a partial date issue
2. No disposition date or date is missing
3. May need business rule for example: Followed 30 days post exposure
4. What DSSTDTC is this rule looked at?
o SD0070: No exposure records found for subject
1. May be screen fail or not assigned to an arm
2. Follow on studies, exposure occurred in first study
3. Randomized not treated (high risk trials – subjects die before treatment)
4. Not identified as screen failure
5. Phase 0 study looking at population, no exposure (gene/biomarker), used to select population for exposure
6. Need clarity on how rule is executed
o SD0088: Reference end data cannot be null for randomized subject. ARMCD≠SCRNFAIL
1. Need clarity on how rule fires
2. Expected not required. Is this appropriate to meet review needs?
  • Topic : Business Process to Review and Identify Resolution Pathway
o Discussion
o General Steps
1. Investigate intent of validation rule and what causes it to trigger
2. Investigate what is in the IG that informs industry practices
3. Assign to a bucket (could be multiple)
• IG Clarification/update
• Rule change or language clarification
• Guideline from Agency
4. Interface group with CDISC, OpenCDISC, FDA, CAB Validation project for resolution
• Create Communication Template to document issues, investigation and next steps (Anne)
• Create tracking spreadsheet (for example Item/Status/Category/Priority) (Sharon)

Group Agenda

Group Discussion

Group Findings

Group Follow up

Back to Working Group 1:Data Validation and Quality Assessment Homepage

Last revision by Hany.aboutaleb,04/11/2012