An Agile Approach to Data - SCDM 2019 Annual Conference … · agile approach to data management...
Transcript of An Agile Approach to Data - SCDM 2019 Annual Conference … · agile approach to data management...
An Agile Approach to Data Management: Improving the
End-to-End Clinical Data Strategy
Speakers:
Julie Savoia, eClinical Solutions
Cecilia Calcagno, PRA Health Sciences
Elizabeth Thompson, Sarepta Therapeutics
Chair:
Diane Lacroix, eClinical Solutions
Speakers
Elizabeth ThompsonCecilia CalcagnoJulie Savoia Diane Lacroix
Associate Director,
Clinical Data
Management
Sarepta
Therapeutics
Manager – Data
Management and
Training, Professional
Services
eClinical Solutions
Lead Data Manager
PRA Health
Sciences
Director – Data
Management
eClinical Solutions
Agile Data Management:Key Considerations for a
Strategic Approach
Presented to: Society for Clinical Data Management
Date: 01OCT2019
Julie Savoia – Manager, Data Management & Training
eClinical Solutions
Agenda
• The new reality of clinical trials is they are always in "startup mode“. To support adaptive designs requires a mindset change and a more agile approach to data management for successful trial execution. We will discuss lessons learned and implementation best practices used to achieve agile data management including design considerations for data acquisition, intelligent analytics, workflows and process changes.
AGENDA ITEMS• Clinical Trial Landscape
• What is “Agile Data Management”?
• Ways to be Agile
• Key Takeaways
Clinical Trial Landscape
• Master protocols as complex novel trial designs
• Accelerated drug development and decreased timelines
• Higher number of amendments
• Higher number of data sources
• Increased complexity creates new challenges
impacting cross functional study teams
What is “Agile” Data Management?
• Embrace change
• Up front planning and preparation for changes
• Identify Data integration platforms
• Tools that support real time eCRF and non-eCRF data integration.
• Deliver completed work frequently
• Implement continuous cross functional data cleaning strategies so that high quality data is always available.
• Collaborate cross functionally
• Stakeholder participation in data review and communication of requirements throughout the trial lifecycle.
Ways to be Agile
Data Acquisition Strategy
• Cross functional stakeholder investment in design components early on in the process
• Configuration planning to minimize downstream impacts when changes occur
• Flexibility vs Rigidity and Over-engineering
Centralized Data Integration, Aggregation and Data Review
• Ability to integrate and aggregate eCRF and non-eCRF data from multiple different sources
• Data review workflows supporting incremental review through identification of new or changed data, resulting in reduced cycle times for data entry to ‘clean’ data
• Data issue management with ability to triage data issues for facilitation of data issue resolution
• Aggregate analytics with drill down, swim up, filter and sort capabilities to support data review and data exploration
• Subject level and study level reporting including reporting on data review progress
Traditional vs Agile Data Management
Technology: Subject Status
Technology: Exception Listings (Incremental)
Technology: Graphical Patient Profiles
Clinical Data Analytics
Data Review and Cleaning Strategy
• Define cross functional data review objectives in a Data Review or Integrated Data Review Plan
• Organize objectives by data domain, role or type of review
• Prioritize or identify the criticality of the data point/domain
• Link objectives to a specific analytic, listing or domain
• Define the frequency for the review
• Define the data review deliverables including the data subsets or ‘cuts’ to be included
• Include documentation review was performed and when
• Conduct data review meetings prior to study deliverables
Metrics - Study and Operational Oversight
• Aggregate approach to integrated operational metrics.
• Identification of risks earlier in the process.
• Define and share key metrics to quantify and evaluate success.
• Publish metrics across the organization to keep data quality top of mind.
Key Takeaways
Key Takeaways
• Resist the urge to over-engineer even though data acquisition configuration options may seem attractive
• Consider downstream impacts when making data acquisition decisions
• Leverage data integration, aggregation, analytics and review tools to drive data review of eCRF and non-eCRF data
• Define cross functional data review objectives and timing to keep data cleaning current, decrease data review cycle times, deliver high quality data at any point in the trial lifecycle, and decrease LPLV to DB Lock
• Define, share and publish key metrics across to keep data quality top of mind
Questions?
IKEA for Data Management
Presented to: Society for Clinical Data Management
Date: 01OCT2019
Cecilia Calcagno – Lead Data Manager
PRA Health Sciences
How to face your role as Data Manager?
• We receive training
• We have a mentor
• We achieve certifications in our role
• We face the real world of DM
Is that enough? NO
We need the extra that each person brings to this role.
Start-Up Stage
Protocol Understanding
Collector & DR tools design
Testing Phase
System Integrations
(Vendors Database and ours)
FSR – First Site Ready
Start-Up Stage: Lessons Learned
Case scenario: Integration challenge – eDiary & EDC
• Early involvement (even informally) in design & integration tools ensures transparency
• Have a PLAN B for scenarios where new components will not be ready by FSR to mitigate downstream impacts
In-Life Stage
Two key players at this point in the life of a study:
Raw data coming from the subjects
More subject data coming from VENDORS
This means…
In-Life Stage: Lessons Learned
• Case scenario: Data Flow Bottleneck
• Identification of possible critical safety issues and timely receipt by adding to concomitant medication form
• Developed report to ID duplicate information between concomitant medication form and the eDiary form
• Proactive escalation to avoid repeating the same scenarios in trials using the same tools
DBL Stage
Several factors have to be aligned in order to close the study successfully:
All data must be present in our database(whether it comes from the site and/or the VENDORS)
All data must be clean
All deviations and not resolved discrepancies must be identified and documented
DBL – DataBase Lock
DBL Stage: Lessons Learned
Case scenario: Preparing for timelines not in DM’s control –impacts to DB lock
• Examine data earlier in process – at startup, request from Clinical at least 2 data loads in life of study, if not more
• Acknowledge risks of loading data only once and/or too-near DB lock
• In-Life meetings need to proactively raise how key turnaround times are tracking (ex: sample analysis by vendor), and back into how those turnaround times relate to key milestones such as DB lock
In Conclusion
• We are part of a whole
• We need to not only develop our capability as a DM but also aim for the best performance of the team
• We always have to have a PLAN B or C or D or E…
• ALWAYS be audit ready, at any phase of a study
• We have to be able to combine the structure that the SOP provides with the flexibility that real life demands
• And ALWAYS remember…
In the life of a Data Manager,
tomorrow won’t ever be the same as today!
Questions?
Agile DM Strategies: Achieving & Maintaining 95% Clean Data
Presented to: Society for Clinical Data Management
Date: 01OCT2019
Elizabeth Thompson – Associate Director, Clinical Data Management
Sarepta Therapeutics
DMD Overview
• DMD: Most common muscular dystrophy, and most common fatal genetic disorder diagnosed in children
• Affects ~1 in 3500-5000 males worldwide
• Caused by mutations in DMD, the dystrophin gene, on X chromosome
• Results in absence of dystrophin, a key protein required for muscle health and function
• Male who inherits a single mutated X chromosome is affected
DMD PrimerDuchenne Muscular Dystrophy
• Symptoms usually appear between 1 and 3
• Delayed motor milestones (mean walking age 18 mos)
• Progressive muscle weakness (age 6-11)
• Waddling gait, difficulty climbing stairs (starting around age 8)
• Loss of ambulation (age 9-13)
• Pulmonary decline (starting around age 10)
• Difficultly with upper limb mobility (starting around age 14)
• Dilated cardiomyopathy (teen years)
• Death in mid to late 20s on average
Sarepta’s Data Strategy: OverviewSarepta is currently working on approximately 20 studies in DMD
• Includes studies in start up, enrolling, open but closed to enrollment, and recently final locked
• Includes both interventional and observational trials
• Phase I-IV, ranging from 12- 200+ patients
• Most interventional studies have weekly dosing with associated visits
An avalanche of data
• Weekly visits range from 4-24 data pages generated in EDC
• One large study contributes 9000 pages a month in new data
Multiple data sources
• EDC
• IVRS
• Video data
• ePRO
• External vendors (central lab, cardiac assessments, etc)
Sarepta’s Data Strategy: Speed and Precision
Use of eCRF from global library with limited changes from study-to-study
• Speeds database build, edit check programming and help text drafting
• Sites in multiple trials become experienced in data entry
• CRAs working on Sarepta suite of studies know what data to SDV and where it is in the casebook
Ongoing data review: multifaceted approach
• Target Data Management review ~95% of entered data at any point
• Standing cross-functional meetings for Medical and safety data review
• Monthly external data reconciliation
All with the goal of having data as clean as possible for the length of the study
Technology: Integrated Data Management
DataProviders
Collection
Organization
Review & Analysis
importer
Data Hub
SAS Server
web services
Clinical Cloud
API orweb services
exporter
sFTP: configuredbatch job
InvestigationalClinical Sites
SponsorsOther External Data Providers
(e.g., Labs, IXRS, other clinical assessments)
Sarepta’s Data Strategy: Rolling Softlocks
With the amount of data involved and the frequency of deliverables, Sarepta DM initiated rolling softlocks for all clinical studies
• ~every 6 months perform a complete soft lock through a milestone date including:
• External data reconciliation
• Medical and safety data review
• DM review
• SDV
• Coding review
• PI signature
• These can be tied to a specific deliverable (i.e. regulatory submission) or a cut date of the team’s choosing.
We want to have the data as quickly as possible for regulatory locks-moving the timelines from LPLV to topline results
Sarepta’s Review Strategy: Dynamic DM Review
Data Management is now all about dynamic and robust review
• DM: Project managers of the data
• DMs need to have ownership of what data is important, what data should look like and where more intensive review is needed
• Cross-functional review teams spearheaded by DM
• Internal data review plans and robust, specialized listings
• Moved away from logic check and programming to meaningful manual review
• Provide direction for where more review is needed from the larger team
DM’s focus is mindful data collection to produce to clean, analysis ready data
Sarepta’s Review Strategy: Technology Driven
Data Management uses next-generation listings for team level review
• Dynamic analytics allow us to drill into an individual datapoint to see other related data and get a more complete picture of a subject or event
• Listings can be built to reflect data across forms, sites and subjects to create a holistic view of an asset
ePRO brings data collection into the home
• Ability data collected on smart phone or tablet
• Video data recorded and sent direct from the device• Tutorials in application• Facial blurring
DM leverages technology in data collection and data synthesis to spearhead targeted reviews by Clinical Operations, Medical, Pharmacovigilence and other teams
Takeaways
• Reuse tools and process where possible to increase speed and accuracy
• Take a multi-faceted approach to achieve ongoing data review
• Softlocks enable readiness for regulatory locks and keep everyone proactive
• Dynamic DM Review makes DMs the strategic project managers of the data and facilitates meaningful manual review that’s focused where it is most needed
• Leveraging technology in data collection and data synthesis spearheads targeted reviews by Clinical Operations, Medical, Pharmacovigilence and other teams
Questions?