IRBshare Overview Gordon R. Bernard, MD Vanderbilt University Medical Center Associate Vice Chancellor for Clinical and Translational Research Senior Associate Dean for Clinical Sciences Professor of Medicine Melinda Owen Bass Chair in Medicine
IRBshare Overview: June 2011-October 2014 Background and history Key decisions and IRBshare s evolution Current progress Hearing firsthand from IRBshare users
Preface: Approaches to reducing duplicate reviews Central IRB (CIRB): a hard-wired entity comprising one IRB to review all applicable studies using one-way reliance agreements with institutions (e.g., NCI CIRB; NeuroNEXT; StrokeNET) ( do not scale well) Reliance agreement: an agreement between institutions allowing the IRB of Record / Reviewing IRB to rotate between the institutions IRBs (can scale well) Ceded reliance: Relying IRBs relinquish all regulatory responsibility to the Reviewing IRB (e.g., The New England (formerly Harvard Catalyst) reciprocal agreement; UT IRB reciprocal agreement; UC BRAID reciprocal agreement) Shared reliance: Relying IRBs complete the review (shared review) adding local context allowing the Relying IRB to maintain regulatory oversight (e.g., IRBshare) IRB Authorization Agreement (MOU): an agreement between two institutions whereby one institution s IRB serves as the IRB of Record for a study or studies
IRBshare development timeline NCRR/NCATS-funded conference series R-13 grant (PI Bernard) Goals Reduce duplication of effort in IRB review Accelerate IRB approval and study start up Maintain if not improve human subjects protections perhaps by increased transparency of IRB reviews, shared best practices Regulatory Expertise OHRP, AAHRPP, WIRB, VA Institutional Representatives CTSA Leadership, IRB Directors, Lawyers, Clinical Trial Experts IRBshare Meeting 1 June 15, 2011 Sponsors Eli Lily, NIH, VA
IRBshare Evolution Since Meeting 1 Key Decisions Made The Model: Shared vs. ceded reliance model The Scope: New/initial study review only End Users: IRBshare is a tool for IRBs to streamline study review IRBs identify multisite studies for IRBshare IRBshare should minimize changes for IRBs (e.g., do not change local submission processes) IRBshare s two-year pilot revealed the needed changes to the model, scope, and end users.
The Shared Review Model A Refresher
IRBshare Shared Review Model LEAD/REVIEWING IRB 1. Local investigator submits study to local IRB 2. Study reviewed by convened IRB upload IRBshare System IRB Approved Documents LOCAL/RELYING IRB 1. Local investigator submits study to local IRB (perhaps abbrev forms) 2. Study reviewed by local subcommittee (>1 IRB member) using IRBshare documents download Committee Responsibility: 1. Determine whether study meets all federal regulatory guidelines for approval of research (e.g., 45 CFR 46.111) IRB application Consent form(s) Protocol + IB/Device Manual Meeting minutes (redacted) Determination letter Subcommittee Responsibility: 1. Verify Reviewing IRB s determination, AND 2. Review for site s own local context issues 3. Assume study oversight
Using IRBshare: User Survey Highlights Need to Clarify Operations Survey sent to all 45 Participating Institutions in October 2014 Response Rate: 84% (38 of 45 sites) IRBshare Feature User Understanding When and who can use the IRBshare How to use IRBshare IRBshare can be used with two-site studies. 97% AAHRPP accreditation is not required to join IRBshare. 97% CTSA affiliation is not required to join IRBshare. 91% IRBshare is free to join and use. 91% IRBshare can be used with federal, industry, or investigator-initiated studies. 91% No IRB Authorization Agreement is required. The IMA satisfies the regulatory requirement. The Relying IRBs local context DOES NOT have to be submitted to or reviewed by the Lead IRB Relying IRBs inherit the expiration date of the Lead IRB 58% IRBshare System access is limited to IRB Personnel only 58% 86% 69% The local context checklist is an optional tool for subcommittee use (not required) 47%
Key Decision: The Shared Review Model Background, Lessons Learned, and Evolution
User Survey Highlights Need to Clarify Advantages of IRBshare Reasons IRBshare Sites listed for NOT USING IRBshare yet (n=28 of 38 sites) Need SOPs 5% Do not see advantage 22% Can Only Rely 3% IRB Staff Resistance 3% Recently Joined 15% IRB Member Resistance 0% Staff Turnover 7% IRB Transitions 10% Awaiting Cont. Reviews/Amends. 25% Competing Initiatives 10%
Why establish a Shared Review Model? Prefer a review model that reduces duplicative reviews with minimal overhead for the Lead IRB and minimal process and procedural changes for all IRBs
Shared vs. Ceded Reliance Model Difficult issues faced by the Ceded IRB Process Significant new work burden for IRBs to serve as the IRB of Record Many IRBs prefer not to cede reviews to other institutions or have institutional policies actually prohibiting it Lengthy contract negotiations for individual reliance agreements Most reliance agreements only apply to a small percentage of multisite studies at an institution Ceded reliances remove the local IRB from the review process, which removes the opportunity to allow shared expertise and standardize best practices IRBshare solution Lead IRBs have no additional responsibilities IRBs maintain local autonomy and control Single master agreement open to all institutions and all studies Increased transparency and access to shared expertise
Shared Model vs. Central IRBs (CIRBs) CIRBs minimize regulatory oversight for local IRBs and reduce redundant investigator effort (duplicate applications); however Review times may not be faster; Initial start-up and ongoing costs are significant; Scope is generally limited to a specific funder/disease type; and Local IRBs have to make significant modifications to local IRB systems, processes and databases
IRBshare: A Lean Reliance Model IRBshare reduces duplicative reviews and leaves local issues to the local IRB.
IRBshare Evolution: Can we offer multiple reliance models? Lessons Learned: Some institutions prefer to completely cede regulatory responsibilities Funders continue to call for ceding regulatory reviews and CIRBs Can one reliance fit all institutions, all studies, all investigators all the time? Can we provide multiple reliance models in one system and allow IRBs to CHOOSE what works best? IRBchoice: More to come after lunch!
Key Decision: Rely at Initial Study Review Only Target new/initial study review because it poses the greatest IRB barrier to study start up
IRBshare Evolution: Expanding opportunities to use the Shared Review! What we ve learned IRBs prefer to use Shared Review at all phases of review Evolution Requested and recently received OHRP and AAHRPP acknowledgement to use the Shared Review at all phases of review IRBshare Expansion: More to come in the next session! Reasons IRBshare Sites HAVE NOT used IRBshare (n=28 of 38 sites) Need SOPs 5% Do not see advantage 22% Can Only Rely 3% Awaiting Cont. Reviews/Amends. 25% IRB Staff Resistance 3% Recently Joined 15% Staff Turnover 7% IRB Transitions 10% Competing Initiatives 10%
Key Decision: IRBshare as a tool for IRBs (only) 1. IRBs can use IRBshare for all multisite studies by checking the website to determine whether to upload or rely
IRBshare Evolution: Requests to use IRBshare need to come from outside the IRB What we ve learned Too many multisite studies to check IRBshare.org for each one Timing issue: by the time studies reach the IRB, many other IRBs have started their review 61% of IRBshare sites report they will use IRBshare if requested Evolution Increasing outreach to investigators, institutional leadership and funders WHAT MIGHT INFLUENCE IRBSHARE SITE USE IN THE FUTURE General Support 10% To gain experience 10% IRBshare Helps Coordinate Use 9% Familiarity with Sites 10% Funder Requests 18% PI Request 25% Inst. Ldrs Request 18% More to come in the last session!
Key Decision: IRBshare as a tool for IRBs (only) 2. IRBs will be more likely to adopt IRBshare if we minimize the changes required to implement (e.g., do not change local submission processes)
IRBshare Evolution: Need to incentivize use for investigators What we ve learned IRBs prefer to use local processes (because they are in place) unless requested otherwise IRBs change processes if requested by investigators or institutional policy Investigators want faster IRB turnaround times AND to reduce duplication in the submission process Evolution IRBshare must offer some way to streamline submission processes where possible More to come in the last session!
IRBshare Utilization: 2 Year Pilot October 2012 to October 2014
IRBshare Network (n=45) Baystate Health* Tufts Medical Center Boston University Medical Center* Tufts University Children s National Medical Center* Tulane University* Columbia University* University of Alabama Birmingham* Duke University* University of Arkansas for Medical Sciences* Louisiana State University A & M University of California San Diego Louisiana State University HSC* University of Illinois Chicago* Maine Medical Center* University of Iowa* Marshall University* University of Kansas* Medical University of South Carolina* University of Kentucky* Meharry Medical College University of Miami Mississippi State University* University of Minnesota* Mount Sinai Medical School* University of North Carolina at Chapel Hill * New York University School of Medicine* University of Pittsburgh* North Shore LIH Health System University of Southern California* Northwestern University University of Texas HSC at San Antonio* Pennington Biomedical Research Center* University of Texas Southwestern Medical Center St. Claire Regional Medical Center University of Washington Seattle Children s Hospital Vanderbilt University* Stanford University* Virginia Commonwealth University* The Scripps Research Institute Washington University (St. Louis) * The University of Texas HSC at Houston* Xavier University (Louisiana) The University of Utah* States (n=23): Alabama Arkansas California Florida Illinois Iowa Kansas Kentucky Louisiana Maine Massachusetts Mississippi Minnesota New York North Carolina Pennsylvania South Carolina Tennessee Texas Utah Virginia West Virginia Washington (Washington DC) *AAHRPP accredited CTSA institution
Preliminary Results Median of 20 days from IRB submission to approval 3 studies approved in <1 week Median of 3 days from IRB review to approval 3 studies approved on the same day as review 10 sites have used IRBshare 8 sites have uploaded their approval 7 sites have relied 5 sites have uploaded and relied Study Reviewing IRB Relying IRB # days from submission to approval # days from review to approval 1 ISCHEMIA Trial Vanderbilt* Duke 6 1 2 MENDS II Vanderbilt Baystate 42 6 3 Vitamin D & Type 2 Diabetes Tufts MUSC 27 2 4 Neonatal Abstinence Syndrome Tufts Vanderbilt 14 0 5 A-5315 HIV UNC Vanderbilt 20 0 6 A-5326 HIV UNC Vanderbilt 22 8 7 Resistance Training in Older Adults 8 9 TCAP UK UAB 1 0 Vanderbilt UK Baystate 10 CHAP UAB UT Houston 7 4 23 90 20 7
Why have IRBs uploaded approval? (n=7 of 8) To Pilot IRBshare 20% Familiar with Sites 7% Funder Req. 6% Ldrship Req. 7% Investigator Requested 20% General Support for IRBshare 27% IRBshare helped coordinate use 13%
Why have IRBs relied via IRBshare? (n=6 of 7) To Pilot 17% Familiar with Sites 0% Funder Req. 17% Leadership Requested 0% General Support for IRBshare 17% IRBshare Helped Coordinate Use 8% Investigator Requested 41%
Hearing from IRBshare Sites Joining IRBshare Leigh Lamonica, Pennington Biomedical Research Center Uploading Approval Leslie Cooper, UAB Using the Shared Review Audrey Ester, UT HSC Houston Streamlining Processes Amy Kolasa, UK