Obtaining IRB approval for multi-center research: challenges and recommendations

Transcription

1 Obtaining IRB approval for multi-center research: challenges and recommendations Keith Marsolo, PhD Assistant Professor Division of Biomedical Informatics Cincinnati Children s Hospital Medical Center (CCHMC) June 24, 2013

2 Disclosure & disclaimer No conflicts of interest to disclose The views and opinions expressed in this presentation are those of the author and do not necessarily reflect those of CCHMC or ImproveCareNow

3 Outline Background & current state Approaches to facilitate approval Upfront work with IRBs Central / Federated IRBs Umbrella protocols Case studies Distributed Ambulatory Research in Therapeutics Network (DARTNet) ImproveCareNow Conclusions

4 Issues with IRB review/approval of multi-center research Request for consent form changes can increase complexity and decrease readability Increases chances that errors will be introduced Lowers study response rate Same study reviewed at different levels Duplication of effort Every center in study must have IRB approval Centers can use the same IRB

5 Current state 2011 HHS Advanced Notice of Proposed Rulemaking Revision to Common Rule Proposed: all domestic sites in multicenter study use single IRB Funders (NIH, PCORI, etc.) assign preference to applicants who can streamline IRB approval process What s an investigator to do?

6 Upfront work with IRBs Engage IRBs early when defining consent process Frequent communication Conference calls/webinars with IRB chairs/staff Provide draft protocols Raise concerns before review

7 Central IRBs (CIRBs) Not tied to a single institution Can focus on specific type of research or geographic area Examples: National Cancer Institute Children s Oncology Group Western IRB Benefits Faster review & approval Lower administrative staff burden Drawbacks Concerns from local IRBs about research integrity, review quality, local context Hesitation about use without demonstration of efficiency

8 Federated IRBs Emerged from CTSA Child Health Oversight Committee Goal: Maintain high ethical standards with reducing duplicative effort Incorporate benefits of CIRB while avoiding pitfalls 3 tiers of participation Tier 1: Reliance on lead IRB Tier 2: Facilitated or full review by local IRB Tier 3: Reliance on local IRB Local IRBs can amend consent material to comply with institutional/state requirements

9 Federated IRBs Benefits Sites choose level of participation Local IRBs can focus on local context Lower administrative burden on local sites Drawbacks Greater administrative burden & liability on lead site Approach is new little evidence

10 Umbrella protocol Protocol that covers establishment of research database Data processing Governance, data sharing Later studies added as amendments/separate protocols citing initial study Beneficial for networks, registries that will spawn multiple studies Consent form should allow this reuse

11 Case study: DARTNet Background Federated network of care centers established in 2008 Facilitate QI in primary care & compile data for CER Architecture Sites assemble a database of patient-level data Data is standardized, de-identified & linked through a web portal De-identification performed by site or Clinical Information Networks of America (CINA) Queries are broadcast to all federated databases

12 Case study: DARTNet Umbrella protocol covers creation/use of databases Site-level QI Practice-level aggregate reports Each research study requires new protocol Centers choose to participate Approved by local IRB or AAFP IRB Studies also require completion of data use agreement Language kept standard unless changes required by law Minimize back-and-forth between lawyers

13 Case study: ImproveCareNow Background 53 centers focused on improving care of children with Inflammatory Bowel Disease (IBD) Started in 2007 as QI network Variety of IRB decisions waive, obtain consent, not research Shift to QI & research after awarded grant from AHRQ Move to federated IRB New legal agreements for data sharing (research & non-research) New patients: consent; Existing patients: maybe/maybe not Note: size of network has almost doubled since grant awarded Registry architecture Before: Central web-based registry populated by double-data entry No patient identifiers (except dates) After: Mix of EHR-based data collection, double-data entry Collect patient identifiers for population management & pre-visit planning reports (non-research uses)

14 Case study: ImproveCareNow 40% rely on CCHMC (22), 60% local (31) Time spent: Aug Jan hours of project staff time spent on IRB 41 s / center (range 25-44) 22 hours of staff time / center 6 weeks total / center

15 Case study: ImproveCareNow Findings (the not so good ones) Electronic IRB submission system has created a bottleneck Only one amendment allowed in the system at a time Center PIs often rely on local IRB for guidance, local IRBs tend to choose local reliance If local IRBs know patient consent is involved, opt for local reliance Even when reliance agreements are already in place, centers still choose local reliance

16 Case study: ImproveCareNow Findings (the good ones) Engagement may be the single most important factor in moving the process along Network leadership, center leadership, IRB leadership Building trust is key Defining an administrative process/checklist has helped efficiency Better communication with IRB Targeted communication to centers Easier on-boarding / training of staff

17 Conclusions Engagement is crucial No silver bullet Process takes a significant amount of time Be sure to plan accordingly Time savings may be downstream (amendments, consent changes, etc.) Entering a new world will take time for IRBs to get comfortable with reliance

18 Questions? Funding Commissioned paper from AcademyHealth R01HS20024