25-Apr-2014 FDA/DCDISC Workshop
Dec 15, 2015
› 9:00-9:15 – Gathering & Welcome
› 9:15-9:30 - FDA/CDER State of Data Standards – Mary Ann Slack, Ron Fitzmartin - CDER
› 9:30-9:45 - FDA/CBER State of Data Standards – Lise Stevens, CBER
› 9:45-10:45 –Challenges with Incoming Data – Doug Warfield, Lisa Lin – CDER, eData Team
› 10:45-11:00 – Break
› 11:00-11:40 – “The Importance of CDASH – A Statistical Reviewer’s Perspective” – Ben Vali, CDER
› 11:40-12:15 –Challenges with Submission Data – DCDISC Team
› 12:15-1:00 – Lunch/Networking
Agenda
› The ideal conditions for this are:1. The problem is clearly defined 2. The input sources are known and have stable
formats3. The desired functionality and output are
distinctly described
Technology Provider’s Task: Apply Knowledge in Order to Solve a Problem
• Problem: High incidence of spilt milk
• Input: Milk • Functionality: (1) Allow
drinking of milk from the cup (2) Minimize spilling when the cup isn’t oriented properly
1. The problem: Can it be determined from the submission data if the drug under review is safe and effective?
2. The input: Submission Data
3. The functionality: The analyses that need to be performed to determine the quality and completeness of the data and the safety and efficacy of the drug.
Technology Provider’s Task and Submission Data
› Data standards (CDISC SDTM and ADaM) – Organize submission data in a systematic way– Control terminology – Reduce the work needed to locate and understand
the data and create functionality– Increase the potential to reuse functionality
The Role of Data Standards is to Regulate the Input
› Success using data standards has expanded the scope of the problem to be solved
The Problem Definition has Changed
Is this drug safe and
effective?
How can the question “Is
this drug safe and effective”
be answered in the more
efficiently and effectively?
How can data collected for
answering the question “Is this drug
safe and effective” be reused to inform
drug safety and public health in
other ways?
How can submission data be
better modeled for its intended use, effective
reuse in drug safety and public health AND
assist drug development the pre-
competitive space?
scope
time
› Data standards change and grow as the problem changes and grows– Adherence to principles of the standard– Early communication
› Correct implementation of the standard– Requires
› Sponsors understand which standards are accepted› Data validation requirements are unambiguous
– Single Set of Validation Rules › OpenCDISC rules, FDA rules, rules from other organizations› Rules based on IG versus common sense rules
– Checks between standards (SDTM, ADaM)› The implementation guides are as specific as possible and cover
expected scenarios– Pharma SUG 2012 paper (Octagon)
› EX domain often implemented incorrectly› Time and practice: implementation is improving
Challenges to Standardized Submission Data
› Analyses for safety are well defined– Therapeutic area domains may require special analyses for
safety
› Efficacy analyses, particularly those using therapeutic area domains, may be complex and present new challenges
› Guidance needed for derived variables– How much support should be provided for missing values
that can be derived (eg imputing BLFL, STDY)– Should derived SDTM variables be validated?
Well Defined Functionality and Output
• Adherence of submission to data standards is improving
• Many of the new challenges result from an expanded problem space
• There are areas that need attention and work
Technology Provider Perspective on Submission Data
› Good Technology Choices– Designing tools to handle expected change
› Terminology changes, variable additions/deletions, validation rule additions/changes
› Anticipating therapeutic area domains, interoperability needs
– What is the role of the Pre-competitive Community (FDA, CDISC and other Standard Development Organizations, Third Party Data Providers, Technology Providers, Sponsors,…) › Should information technology and data standards usage
decisions involve all stakeholders and open discussion about risk/benefit issues– Metadata Repository– Changes to Standards
Technology Provider’s Responsibility
Adventures in CDISC Consulting and eCTD Preparation
Jane Diefenbach, PharmaStat LLCDCDISC Implementation Network at FDA
April 25, 2014
Preparing for a data submission
• 18 to 24-month lead time• Legacy studies and Phase III may just be
starting• Everything is changing: people, processes,
knowledge of the product, often even the company structure
• When standards change too it can be hard to accommodate them in this environment
What needs to happen
• Planning– What studies? What data formats? What standards?
• Production – legacy studies– Align the studies with a unified modeling plan– Perform the conversions, verify with reanalysis
• Production – current studies– Real time conversion and linear data flow is difficult
• Documentation, maybe most important
Planning the Submission: Company XYZ Data Flow
ADaM
eCTD
SDTM
Studies:
CDMS
101, 102, 103
Analysis
CSR
Item 11
eCTD
201, 202
CDMS
Analysis
CSR
eCTD
SDTM
203
CDMS
ADaM
CSR
eCTD
SDTM
301
Integrated Data for ISS
104
CDMS
ADaM
CSR
eCTD
SDTM
CDMS
ADaM
CSR
eCTD
SDTM
ADaM
eCTD
ISE
Credit John Brega 2014
Experience to Date
• Work volume is astronomical, easy to underestimate
• Clarity of purpose is sometimes absent– Standards are overwhelming; sponsors don’t know
what is most important– We imagine perfection at kick off; eventually just
need to get done– Downstream changes to standards, modeling are
costly; sometimes we need to say Next time
What makes it easier?
• Use of CDASH in data collection• Predictability of new standards, fair warning• OpenCDISC – has done more to standardize
implementation than anything• Electronic distribution of linked standards
(SHARe) and a data standards warehouse• Widespread adoption
What makes it easier?
• Communication!– The more we know about how you interact with
the data the better we’ll do– Guesswork leads to busy work and busy work does
not advance Science or Health
› Implementation/Acceptance of Standards by Agency– Set time period for accepting submissions in new
version› Allows sponsor time to decide when to upgrade
– Work with CDISC to implement changes rather than publishing guidance› Saves time searching through documents
– Set policy on use of provisional / sponsor-defined domains› Are sponsor-defined non-X_, Y_, Z_ domains acceptable› Will provisional domain templates be accepted
Sponsor Company
› ADaM vs Sponsor-Formatted Analysis Data Sets– Will agency be requiring ADaM
› Versioning of CDISC Standards– Some products take 10+ years to develop– Acceptance of submission data in multiple versions
Sponsor Company