Data Management and OpenClinica®  Infrastructure…..from scratch Elisa L Priest, MPH Manager of Clinical Trials Baylor Health Care System Dallas, Texas May 11, 2011

Background I’ve given advice to small groups trying to develop  g g p y g p an FDA compliant data management  infrastructure.  These groups are often looking at  OpenClinica because it is open source and freely  because it is open source and freely available.  However, OpenClinica is just one small  part of the data management infrastructure that  needs to be developed.  This presentation  d t b d l d Thi t ti describes the process that Baylor Health Care  System has gone through in the past 4 years to  y g g p y develop an FDA compliant data management  infrastructure to support investigator‐initiated  trials. trials

Overview • • • • •

Business Need Business Need Document Requirements Evaluate existing infrastructure l i i i f Phase 1: Basic Infrastructure Phase 2: More Infrastructure

• Lessons Learned

BUSINESS NEED BUSINESS NEED

Business Need Business Need • One small FDA regulated Phase II trial: – Original plan – In‐house DBMS – Would require extensive programming and validation to comply with  FDA regulations

Then… • One larger FDA regulated, multi‐site Phase II Trial – Larger trial with similar startup timeframe – Multi‐site trial introduces complexity into data management

• One larger non‐FDA regulated, multi‐site Trial

Business Need Business Need

to capture and manage clinical data in  to capture and manage clinical data in electronic format in a manner that  meets FDA requirements meets FDA requirements

DOCUMENT REQUIREMENTS DOCUMENT REQUIREMENTS

Document Requirements Document Requirements • Regulatory (FDA) requirements Regulatory (FDA) requirements – Investigator initiated (Investigator = Sponsor)

• Data Management requirements Data Management requirements – Paper based  Eventually EDC – Multi‐site – Multiple Trials – Different phases and therapeutic areas

• Additional business requirements

Regulatory Requirements Regulatory Requirements • 21 21 CFR part 11 CFR part 11 • Guidance for Industry: Computerized Systems Used  in Clinical Investigations, May 2007 g , y • Guidance for Industry: Part 11, Electronic Records’  Electronic Signatures‐ g Scope and Application, August  p pp g 2003 • General Principles of Software Validation; Final  Guidance for Industry and FDA Staff, January 2002

Data Management Process  Requirements • • • • • • • • • •

Creation of Entry Screens that mimic paper CRF y p p Edit checks Generate queries Manage discrepancies (queries/SECs) Study lock Extract data Coding dictionaries (MedDRA) E External data loading l d l di Paper tracking/process tracking Reporting Reporting 

EDC versus Paper Requirements EDC versus Paper Requirements  •

EDC‐ Electronic Data Capture Trial

•

Paper‐based Trial

– Generally longer study start‐up – Sites may begin enrollment  before db is fully developed – All database building completed  around the time of first patient in – Extensive edit checks built into  forms (out of range, invalid, logic  checks)

– Need data entry screens as the  first sets of data come in (after first sets of data come in (after  CRFs are monitored against  source) – Edit checks on entry (usually too  late for verifying source) late for verifying source)

– Study Coordinators enter data at  site – Specialized data‐entry staff – Monitors compare source with  electronic forms l f – Need 10‐12 weeks from Final  Protocol

– Monitors compare paper CRF  with source – Need 8 Need 8‐12 12 weeks from final CRF weeks from final CRF

Additional Business Requirements Additional Business Requirements • Training/Customer Support Available Training/Customer Support Available • IT support available • External Hosting l i

EVALUATE EXISTING  EVALUATE EXISTING INFRASTRUCTURE

Existing Infrastructure Existing Infrastructure • Human Resources Human Resources – One Access database programmer – One data manager/ SAS programmer One data manager/ SAS programmer

• Health Care System IT network • SAS • Microsoft Access

Existing Infrastructure Existing Infrastructure Basically No data management infrastructure Basically…..No data management infrastructure

PHASE 1: BASIC INFRASTRUCTURE PHASE 1: BASIC INFRASTRUCTURE

Phase 1: Basic Infrastructure Phase 1: Basic Infrastructure • • • •

Data management/Data Capture Software Data management/Data Capture Software Software Training Define Data Management Processes fi Develop SOPs

You should also have a Documented Plan! You should also have a Documented Plan!

Phase 1: Software Phase 1: Software

Software Training

Validation

DM  Processes Training

SOPS

Software Review Process Software Review Process • Business needs/ Requirements Business needs/ Requirements • Determine priorities for software system: Determine priorities for software system: – Time: When must you have the solution in place? y p – Budget: What dollar figure must you not exceed? – Value: To what extent does the product have to meet your  needs? – Scalability: how long do you expect to use this product?

Software Infrastructure Software Infrastructure Review

Training

Develop SAS  abstraction b t ti

Choose

Purchase

*Validation

Installation of  Installation of multiple  environments

Validation Plan Validation Plan • Review Review Validation Documentation Provided by  Validation Documentation Provided by Akaza Research • Develop User Requirements Develop User Requirements • Develop Performance Qualification Test Plan • Develop or modify additional documentation • Perform validation according to plan and  document

Timeline

July 2007

• Review Regulations

Fall 2007

• Software Software Requirements and  Requirements and Evaluation

Spring 2008

• BIS software recommendations • PO for OpenClinica for BIIR

Fall 2008

• Installation of OpenClinica at BIIR

Fall 2009

• IHCRI requirements for OpenClinica

Spring 2010

• RFP for OpenClinica and PO started • Data Management SOPs

Sept 2010 October 2010

• PO Approved • OpenClinica Installed • Validation Project Began

November 2010

• Site Audit of Akaza

December 2010

• Finalization of Validation Plan

May 2011

• Continue “moving towards  compliance”

Phase 1: DM Processes Phase 1: DM Processes

Software Training

Validation

DM  Processes Training

SOPS

Data Management Processes and  Phases of Research h f h Concept  & &  Design

Statistical  Statistical Analysis

Reporting 

Preliminary tests of data collection tools/process SAS or other  Exposure, outcomes,  Data Management Plan (DMP)‐ Id data  programs confounders identified confounders identified requirements and standards and processes requirements and standards and processes Finalize data collection tools edit checks DMP Finalize data collection tools, edit checks, DMP

Reports and  publications

Pl Planning i

E Execution i

Edit checks/cleaning Estimates of  Determine measurement Id all variables used in protocol and event  Build electronic capture (database, spreadsheet) effect table  Analysis  Validate e‐capture Statistical requirements for  documentation Id all data from statistical analysis plan variables Collect data

Id all data for required reports: DSMB

Data receipt/tracking

Id all external data sources (charts, lab  Data entry/ verification measures, ect)  Document variable formats  Id edit checks for variables

Queries and corrections Edit checks/cleaning Coding

Data collection tools: paper Case Report  Data transfers from/to  Forms (CRFs) Annotated CRF

Data integration Data integration Data process reports SAS or other programs

Estimates of effect E ti t f ff t

Termination

Archive data Archive  A hi documents  Metadata/docs Make data  available for re‐ il bl f use

Developing Data Management  Processes Draft SOP  creation??

OC Training OC Training

Develop OC  Expertise  + +  Use previous  Experience

DM Process  Training

Create DM  Processes

Refine  Processes

SOP creation

Standardize Standardize  Processes  across studies

Developing Data Management  Processes • May be different for EDC vs. Paper May be different for EDC vs Paper – May need SOPs for both

• Start Start as simple as possible and then add  as simple as possible and then add processes as need • Does not need to be perfect at first Does not need to be perfect at first • Depending on software capabilities, may need  additional infrastructure for tracking forms additional infrastructure for tracking forms,  queries… – Access Access based study database for tracking data  based study database for tracking data management processes including forms and queries

Developing Data Management  Processes • Our approach: Our approach: – Develop processes and familiarity with  OpenClinica with a paper with a paper‐based based process process – Test EDC capabilities in a subset of patients to  ensure clinical site comfort ensure clinical site comfort – Move to full EDC in future trials after  infrastructure development p Paper

Paper +  EDC

EDC

EDC first‐‐Why EDC first Why not? not? • Infrastructure for EDC Infrastructure for EDC – Knowledge/Expertise of OpenClinica – Provide Training for Clinical Site Provide Training for Clinical Site • General data entry training Study‐Specific Specific data entry training data entry training • Study

– Provide Support for Clinical Site • Answer questions when needed Answer questions when needed

– IT infrastructure for site – Maintenance of security/permissions  Maintenance of security/permissions documentation

EDC first—Why EDC first Why not? not? • Infrastructure for EDC Infrastructure for EDC – Edit Check programming – Additional validation Additional validation – User acceptance testing completed by clinical site – eCRFs CRF ready for production (and validated) by  d f d ti ( d lid t d) b time of study start

Why SOPs? Quality: GCP 5.1 Why SOPs? Quality: GCP 5.1 • The sponsor is responsible for implementing and  gq y q y maintaining quality assurance and quality control  systems with written SOPs to ensure that trials  are conducted and data are generated ,  d documented (recorded), and reported in  t d( d d) d t di compliance with the protocol, GCP, and the  applicable regulatory requirements. applicable regulatory requirements. ICH Guideline for Good Clinical Practice E6(R1) 1996

Data Management SOPs Data Management SOPs • SOPs on DM Processes: all phases of research SOPs on DM Processes: all phases of research – Study Planning/Start up – Execution • • • • •

Filing/Storage Paper work flow Paper work flow Paper tracking Discrepancies/Queries Data Extract

– Study closure/ Archiving

Data Management SOPs Data Management SOPs • Create Listing of all SOPs needed Create Listing of all SOPs needed – caBIG – Good Clinical Data Management Practices from  g the Society for Clinical Data Management – Practical Guide to Clinical Data Management  (Prokscha)

• Prioritize Listing • Create SOPs • Responsibilities for personnel

Data Management SOPs Data Management SOPs • 25 Priority 1 SOPs 25 Priority 1 SOPs • Around 400 hours of creation time SOPs – SOPs (versioned) – Procedure descriptions/work instructions (not  versioned) – Template Forms

SOPs

Scheduling

Scheduling

• Estimated Effort vs. duration – – – –

Draft Committee Review Committee Review Editing/Final Committee Review and sign off Executive Review and sign off

Example Weekly Schedule and Tasks Example Weekly Schedule and Tasks • This week 05/17/2010 • Writing:  Due to Executive on 06/04/2010 – Study Database Data editing – Data Cleaning and Review

• Reviewing: Due to Review this week.  Due to Executive on 5/28/10 – Database design and Creation (Change name to Study Database  Design and Creation) Design and Creation) – Study Database Validation – Study Database Edit Check Programming

• Final Final Editing Monday/Tuesday/Wednesday: Due to Executive  Editing Monday/Tuesday/Wednesday: Due to Executive 05/21/2010 – CRF Forms and Flow Management – Data Management Roles and Responsibilities g p

SOP on Programming SOP on Programming

Phase 1: Basic Infrastructure Phase 1: Basic Infrastructure • Maturity of Training  Maturity of Training 1 Person Trained d

“Informal Informal Training Training” of  of others as needed

• Training Topics – Data entry a ae y – Study Setup – Data Principles Data Principles – CRF Creation

Baylor‐specific Baylor specific  standardizedtraining

Train the Trainer  h

Develop ‘customized’ Develop  customized   training for each study

PHASE 2: MORE INFRASTRUCTURE PHASE 2: MORE INFRASTRUCTURE

Phase 2: More Infrastructure Phase 2: More Infrastructure  • • • • •

Infrastructure Strategic Planning Infrastructure Strategic Planning EDC Support Data Management SOPs: Priority 2/3 SO i i 2/3 Data Management Training Data Management Competencies – Job Descriptions Job esc p o s – Training Matrix – Personnel Development Plans Personnel Development Plans

Lessons Learned: Implementing  Clinical Trials in OpenClinica®   Elisa L Priest, MPH Manager of Clinical Trials Baylor Health Care System Dallas, Texas May 11, 2011

Lessons Learned: Implementing  Clinical Trials in OpenClinica l l l l II will review my experience building  will review my experience building infrastructure using OpenClinica. Specifically,  lessons learned while implementing multiple lessons learned while implementing multiple  investigator initiated trials in OpenClinica. This  will include organizing and tracking eCRF will include organizing and tracking eCRF development, versioning eCRFs, and creating  paper CRFs compatible within the OpenClinica paper CRFs compatible within the OpenClinica framework

Lessons Learned: Overall Lessons Learned: Overall  • Organization • Documentation • Change is certain!! g

Paper CRFs Paper CRFs • Read Read the protocol the protocol • Create your own study event chart • Compare the provided study event chart (if  C h id d d h (if there is one) with yours to identify  di discrepancies i • Identify/Organize data into repeating modules Identify the eCRF sections and headers • Identify the eCRF, sections, and headers

Study Event Chart/ CRF Matrix Study Event Chart/ CRF Matrix

Paper CRFs Paper CRFs • Recreate paper CRFs to be more compatible with OC  interface – Listen to your gut: reformat if necessary!

• Be consistent in formatting to represent – eCRF title – Section S i – Headers

• Limit the number of columns • Limit unnecessary repeating groups (for data extract) • Be consistent in answer choices (1‐yes, 2‐no) B i t ti h i (1 2 )

Formatting of eCRF and Sections Formatting of eCRF and Sections

Paper CRFs Paper CRFs • Require Require testing of pCRFs testing of pCRFs with hardcopy proof  with hardcopy proof that they have been tested • Send only PDFs to review Send only PDFs to review • Use a draft watermark and only remove it  once CRFs are approved CRF d • Versioning and tracking changes is critical

Organizing and Tracking eCRFs Organizing and Tracking eCRFs • Use Use study event chart format study event chart format • Track initial design, testing, versioning • Track response options and response options  k i d i text • Be consistent on response options and text  across eCRFs in a study

Tracking development Tracking development

R Response Options O ti

eCRF development • Be organized and consistent Be organized and consistent – Naming folders and files – Naming variables across eCRFs Naming variables across eCRFs – Response Options and text

eCRF development • Develop in stages – – – – –

Local environment: Development and 1st testing Review and approval by second person Upload into testing environment Upload into testing environment Testing with real data by at least 2 different staff Versioned and uploaded into production

• Cannot Cannot change variable labels, types, response options  change variable labels types response options  rename variable • Versioning/change control is time consuming 

pCRFs and eCRFs and eCRFs versioning • Link paper CRF versions with eCRF Link paper CRF versions with eCRF versions. versions

Lessons Learned Lessons Learned • You will make mistakes ou a e sta es • Learn from your mistakes Learn from your mistakes • Use the community Use the community – Prevent mistakes – Help others prevent mistakes p p – Contribute to Mantis when you find a potential  problem

Thanks! • Dr. Sunni Barnes

• Consultants – Suzanne Prokscha – Dr. David Hardison

• IHCRI Clinical Trials Team IHCRI Clinical Trials Team – – – – – –

Monica Anand Tyson Bain Candice Berryman y Kristen Rose Deepa Putuvakkat Alicia Turoff

• Akaza Research

Thank you! Thank you! Elisa L Priest, MPH Manager of Clinical Trials Baylor Health Care System Dallas, Texas [email protected]