The barriers that IRBs present to researchers following a disaster are broad and complex and often can delay a study getting off the ground. Participants explored strategies to balance human subject protections while enabling timely IRB review of research protocols, discussed options to obtain informed consent in emergency situations, and considered the ethics of data collection for special populations in disasters.
In 2011, the National Preparedness and Research Science Board (NPRSB)1 at ASPR produced a report on science preparedness disaster research, discussing what would be needed to bring the ongoing efforts by NIH, CDC, ASPR, and others together in a more cohesive and organized way.2 Among its recommendations, the NPRSB called for the creation of a PHERRB, an IRB that would be able to rapidly convene to assess research protocols while maintaining very robust protections for human subjects. Diane DiEuliis, deputy director of the Office of Policy and Planning at ASPR, said that ASPR has been working with NIH to develop the PHERRB. It was established in 2012 by HHS and is main-
1Formerly the National Biodefense Science Board.
2Available at http://www.phe.gov/Preparedness/legal/boards/nprsb/Documents/nbsbrec14.pdf (accessed September 8, 2014).
tained by NIH through the Office of Human Subjects Research Protections (OHSRP).
Michael Gottesman, deputy director for intramural research at NIH and the institutional official at NIH responsible for OHSRP, said NIH has experience in conducting research under disaster conditions (e.g., the GuLF Study discussed by Birnbaum). In addition, NIH has an extensive IRB system, many reliance agreements and cooperative relations with the IRBs throughout the country, and numerous experts at its disposal. As host of the PHERRB, NIH provides staff, develops operating procedures, and maintains operations.
The PHERRB is a central IRB and serves as the single IRB for human subjects protections review of HHS-conducted, -supported, or -regulated research studies addressing public health emergencies (e.g., natural disasters, biohazards including anthrax, chemical and radiological emergencies, oil spills, pandemic influenza or other infectious diseases, and other mass casualty events). Gottesman explained that the PHERRB would provide human subjects protections/regulatory review only under existing applicable federal regulations (including 45 CFR 46 and/or 21 CFR 50 and 56). The PHERRB can serve as the IRB of record for any institution (this can be a state health department, an academic medical center, a community hospital, etc.) that is engaged in the conduct of the protocol and that executes a reliance agreement with NIH for PHERRB review. At present, use of the PHERRB is generally encouraged but not required, so institutions conducting public health emergency research could choose to use a local IRB (or multiple IRBs) for human subject protections review. The PHERRB would conduct initial reviews, continuing reviews, review of amendments, review of unanticipated problems, and local context review. Gottesman noted that review by the PHERRB does not replace other institutional oversight responsibilities (e.g., principal investigator training, adequacy of local resources, ancillary reviews, IRB office functions).
An IRB Authorization Agreement (reliance agreement) between NIH and the institution conducting the research is required. The reliance agreement allows the PHERRB to conduct human subjects research protections review of a public health emergency research protocol in the place of an IRB in a local community, under the NIH Federal-wide Assurance.3 The agreement provides a road map for the life of the study and the reliance relationship. It also helps institutions consider all issues
3See http://www.hhs.gov/ohrp/assurances/assurances/filasurt.html (accessed November 10, 2014).
in advance and describes the division of responsibilities between the PHERRB and the local institution (e.g., conflict of interest, investigator human subjects protection training, confidentiality and privacy, compensation of participants, other committee reviews that may be needed). Gottesman noted that NIH currently has 12 IRBs, and any of them can be designated to serve as the PHERRB IRB of record. In addition, NIH is prepared to assemble an IRB as needed to deal with multicomponent disasters (i.e., additional expertise could be brought into any of the basic IRBs). In general, principal investigators would be able to access the presubmission application, frequently asked questions, and other information about the PHERRB process on the OHSRP website. Upon submission of a protocol, NIH will determine if it is within the scope of PHERRB review.
One issue yet to be resolved is whether and with whom reliance agreements should be prenegotiated. Another question is whether individual academic institutions would want to have these open-ended reliance agreements. Could protocols be developed in advance of an emergency to accelerate initiation of research activities, and if so, would these protocols be developed by NIH, or would NIH solicit submissions? Gottesman noted that the concern with soliciting submissions is that there could be hundreds or even thousands of protocols, and the IRB system would come to a crashing halt. NIH could perhaps solicit protocols from a limited number of institutions, those that are part of a clinical research network or are highly likely in some other way to have to deal with emergencies.
Nell Allbritton, the IRB director for the Louisiana Department of Health and Hospitals, said her experience shows that a state agency has a set of concerns that are very different from those of researchers and universities. A key issue for a state health department is patient privacy. The state errs on the side of caution when allowing researchers access to patients, she said. In addition, transitional staff and space during disaster response can compromise the security of confidential data and the privacy of subjects. The first priority for a public health agency is to respond and to mobilize quickly so that patients are removed from harm and continue to receive services. Monitoring or assisting in research
efforts can be a strain on the resources of the health department and its ability to be responsive to the community’s needs. Another concern for the health department is that research participants will confuse research for clinical services (referred to as therapeutic misconception). Communities decimated by disaster are also vulnerable to exploitation and additional harm. Research can be designed to pose minimal risk to participants, but could still compromise the community’s ability to respond and recover. Finally, maintaining the participation of the population being studied is a challenge for disaster research, as survivors relocate and might lose access to prior means of communication.
Recent Proposals in Louisiana
Allbritton shared several examples of disaster research proposals that her IRB has dealt with in recent years. The first example from 2005–2006 involved a Louisiana State University study to evaluate rapid-needs assessments conducted by the state mental health services provider. Around 250,000 people were displaced to Baton Rouge following Hurricane Katrina, and more than 6,000 people who sought treatment at the Capital Area Human Services District were surveyed with informed consent. However, the study had not been reviewed or approved by the IRB, Allbritton said. The health department learned of the study prior to publication and would not let its publication proceed until the IRB had reviewed the manuscript. (She clarified that while it is possible that the study was approved by the university IRB, the state IRB is the IRB of record because a state mental health services population was the study population.) Another Louisiana State University research project, examining the efficacy of a group treatment protocol for posttraumatic stress disorder, was able to modify its existing protocol for use with Hurricane Katrina survivors. The IRB turnaround was rapid, as this was a modification of research already being done.
In 2009, Tulane University sought to identify the most effective therapy model for stress reduction of substance-abusing adults in disaster-prone regions. The intent was to study clients who entered addictive disorders clinics, correlated to periods of natural disaster that occurred in the region. The instruments and the methodology were submitted and approved by the IRB prior to the commencement of the study period, Allbritton said. As it turned out, there was no hurricane activity in 2009 in the region, so while this is a model of preapproval of a protocol, the research did not occur. More recently, the IRB approved a
Tulane University study comparing mental and physical health outcomes over a 4-year period in 1,800 women exposed in varying degrees to the Deepwater Horizon disaster.
Lessons in Reviewing Disaster Proposals
Drawing on her experience with these four studies as well as others, Allbritton shared some of the lessons learned by the IRB. The oil spill study, in particular, raised the issue of therapeutic misconception, she said. The IRB ensured that the consent forms clearly defined what would be done with the blood and saliva samples so that none of the participants would mistake consenting to give samples for being able to receive clinical services. Investigators are responsible for assessing individual decisional capacity and the possible effects of the research on participants, and Allbritton added that monetary incentives can be coercive. In a disaster situation, investigators or support staff should be trained to identify participants who are distressed and need aid, and refer them for care. There should also be a plan for care of researchers exposed to emotionally difficult situations. Including community leadership in the IRB review of a protocol is the best way to get community buy-in before the protocol is deployed, Allbritton said. Preapproval of the protocol and methodology expedites deployment of the study, and provisions for confidentiality and privacy should be an explicit part of the research plan. Ground-level program staff must also be prepared to redirect researchers to the IRB for approval of any deviations from the protocol.
With regard to reliance agreements, Allbritton noted that there is currently language in the Louisiana state rule that prohibits reliance agreements with other IRBs and would therefore prevent reliance agreements with the PHERRB. However, she agreed to return to her state and bring up the need to consider working with the PHERRB or a centralized IRB.
While there is not a separate or unique application process for disaster protocol IRB review, there are extra questions that the IRB for the Louisiana Department of Health and Hospitals recommends researchers consider in developing their protocol and methodology (see Box 4-1). Often, the IRB might approve a protocol with stipulations that certain changes be made, and this delays the process as researchers must then return to the IRB with the answers to the questions. Where app-
licable, including answers to these questions in the protocol will hopefully reduce the revisions that the IRB requires, Allbritton said.
Disaster takes us out of routine, Allbritton concluded. As a state agency, day-to-day operations change completely, and employees may also be directly impacted. Maintaining services is challenging, as even normal activities require additional oversight. The research component of the department, unfortunately, does not take priority in times of crisis, but it is important not to discourage or prevent this kind of work from being done. Therefore, Allbritton said, the state IRB must employ additional measures to ensure proper protections for public health consumers participating in postdisaster research and safeguard the ability of the department to maintain the standard operating level during catastrophic events.
Questions to Address for Institutional Review Board Review of a Disaster Protocol from Nell Allbritton
- Does the design make it easy to collect data immediately after the disaster occurs?
- Do the researchers contribute their time in relief efforts in addition to collecting data?
- Are the researchers prepared to convince disaster relief workers that the study is beneficial to the survivors in order to gain access to participants?
- Does the protocol account for continuing research and recruitment during and after the relocation of survivors?
- Will researchers ask participants for contact information to be shared from agencies as part of informed consent?
- What alternative resources will the researchers provide to participants?
- Do consent forms account for diversity in age, culture, and education?
- How will the researchers determine who may give consent for minors to participate?
- Has the community given consent?
SOURCE: Allbritton presentation, June 13, 2014.
Given the priority of meeting the immediate medical and mental health needs of survivors of and witnesses to a disaster, the issue of deciding when and how to conduct mental and behavioral health research with these populations is logistically and ethically challenging, said Holly Taylor, core faculty at the Johns Hopkins Berman Institute of Bioethics. With funding from the Preparedness and Emergency Response Center (a CDC-funded entity) at the Johns Hopkins Bloomberg School of Public Health, Taylor set out to describe and consider the ethical challenges encountered by IRBs and investigators in the review, oversight, and conduct of postdisaster research.
The study is ongoing, but Taylor was able to share some of her preliminary findings. She noted that the IRBs managed reviews of disaster research in many ways. In some cases, a unique committee or the IRB subcommittee was created to review the study prior to the IRB review, and it appeared to look at issues unrelated to human subjects research. Some IRBs put the disaster protocols into the routine review process, while others prioritized them.
Among the concerns raised by the IRB members she interviewed were the quality of the research proposed and the feeling that researchers were “following the money.” The level of harm to which subjects may be exposed was also a concern (however, Taylor added that the current literature strongly indicates that people do not experience any additional harm as a result of being interviewed). The IRB members also raised concerns about the burden on subjects, potential for therapeutic misconception, and safety of research staff. Challenges encountered by the IRBs interviewed included the speed required, the multiple stakeholders involved (including labor unions and employers when workplaces are involved, as well as law enforcement and the legal system if it is a crime scene), lack of coordination among institutions and the IRBs, turf battles with affiliated institutions, multisite research, and the value of certificates of confidentiality (especially relative to federal court cases).
The principal investigators interviewed reported positive experience with the IRBs (with the exception of navigating multiple IRBs). Investigators reported that the IRBs had issues with the quality of the research they proposed, asked them about the level of harm to which subjects need to be exposed, and stressed the importance of having
appropriate referrals in place. Procedural challenges highlighted by investigators included funding, additional layers of review, lack of communication among federal sponsors, and access to populations. Substantive challenges mentioned by investigators were the availability of appropriate services for referral (which are often lacking in a disaster) and the potential vulnerability of the subjects.
Some unique challenges were noted. For example, there is community-wide impact, and the affected population may include the study teams. There are also infrastructure issues impacting the retention and follow-up of displaced subjects. Local staff at local agencies are overwhelmed, and it is challenging to train them to adopt new interventions while they are managing their own personal and professional priorities. The number of medical service providers, social service providers, agency staff, and other research teams in the field created confusion for the subjects. Taylor explained that this became a problem for one research team when they were mistaken for a different research team that had apparently made promises they did not fulfill. DiEuliis added that after Hurricane Sandy, one issue raised by investigators was the potential for “survey fatigue” among participants, given the large number of research projects being done concurrently. It was suggested that CDC, NIH, FEMA, ASPR, and other relevant agencies could coordinate somehow so that all of the investigators could see each other’s survey populations and questions. She noted that ASPR, working with NIEHS, has been building a web space for the grantees that will have a map and listing of shared research populations. The next step is working with the investigators to determine the best way to help them coordinate with each other. She suggested that the PHERRB may be able to provide assistance in this area.
DiEuliis reported on the top challenges and issues identified by participants (see Box 4-2). She indicated that there was broad recognition of the usefulness of having a national-level, centralized IRB, such as the PHERRB. Research during a disaster is different from general research, and the use of a central IRB in a disaster is not commonplace.
Another issue raised was the need for coordination among in-stitutions, the IRBs, and the federal entities funding the research. It
Addressing IRB Barriers to Health Research Implementationa Challenges and Issues
- Recognition of the usefulness of the PHERRB, but also that there would be same standards for disaster research.
- Coordination among institutions/the IRBs/federal support entities, particularly with regard to “survey fatigue” in affected populations.
- Quality of research and asking the most important questions.
- Differentiation of acute needs (at beginning of an outbreak) versus longer term, more structured clinical studies; are definitions needed?
- Reliance agreements and how to approach them. Acknowledgment that states may not recognize these.
- Confusion of “research” versus “services.”
Critical Partnerships and Collaborations
- States, localities, research institutions, principal investigators, federal agencies.
aThe challenges, opportunities, and partnerships listed were identified by one or more individual participants in this breakout panel discussion. This summary was prepared by the panel facilitator and presented in the subsequent plenary session. This list is not meant to reflect a consensus among workshop participants.
SOURCE: Plenary session summary of breakout panel discussion as reported by panel facilitator Diane DiEuliis.
was suggested that the absence of coordination could lead to survey fatigue in affected populations, with many different investigators approaching the same people or populations for different study protocols. During the discussion, a participant suggested that the central IRB could maintain a list of all disaster-related protocols and make the list available to all researchers and the IRB reviewers. Several participants highlighted the importance of understanding what is acceptable to the community in terms of human subjects protection (an example given was a tribal entity that had very different views of who should have access to their data and
how that data could be profiled and published). This may mean involving the community in developing the protocol for the IRB submission.
With regard to approval and coordination of the many potential research studies, a participant stated that it can be difficult to make the distinction between an investigator who is clearly taking advantage of a situation (i.e., “following the funding”) versus one who sees a real connection to his or her ongoing work. DiEuliis pointed out that a stipulation of the Hurricane Sandy supplemental funding was that any outside institutions (outside of the hurricane-affected states) seeking funding needed to demonstrate that they had relationships in the communities and with the affected populations. Taylor suggested looking at how the IRBs handle international collaborations for lessons that could be applied to working with investigators from outside the disaster-affected area (e.g., requiring documentation that researchers are collaborating with a particular local organization and have the IRB review from the local level). A participant referred others to the World Association for Disaster and Emergency Medicine updated disaster research and evaluation frameworks (Birnbaum et al., 2014), which offers five frameworks to apply to studying and comparing disasters and evaluating interventions. The critical piece of these frameworks is the inclusion of standardized definitions throughout.
In the same vein, some participants emphasized the need for attention to the quality of the research proposed. Are we asking the most important questions when looking back over the history of clinical research done during disasters, DiEuliis asked, or are we asking the same questions over again? There was also discussion of differentiation between acute needs (at the very beginning of an outbreak or event) versus long-term, more structured clinical studies. Several participants discussed the possible need for definitions or a rubric for the different kinds of research that would be needed during and after a disaster, and the associated IRB needs.
Much discussion took place about how best to approach and broker reliance agreements. It was reiterated that some states may not recognize the type of reliance agreements that would be done at the federal level directly with institutions. Several participants stressed that reliance agreements should clearly spell out the roles and responsibilities of the entities involved (e.g., HIPAA responsibility). Finally, DiEuliis said, the issue of therapeutic misconception was discussed further, including ensuring that study participants understand research versus services and why the research they might be part of is of value to recovery.