A PHE might be the only opportunity to collect clinical safety and efficacy data in humans; therefore, it is important to consider what questions should be answered (see Chapter 3) and leverage existing systems and infrastructure as much as possible to collect informative and actionable data on MCM use, said Runnels. Higgs added that three critical goals of conducting clinical research during a PHE are mitigating mortality, ending the outbreak, and developing rigorous, regulatory-level data for future outbreaks. Meeting these goals requires increased efficiency, expediency, and adaptability compared to routine clinical drug development. For studies conducted rapidly during a PHE, it is more difficult to answer questions about how the product fits into the larger paradigm of care, what combination of therapies might be effective, how the product is best administered and dosed, or which subpopulations may benefit the most, said Peppercorn. Participants throughout the workshop considered these issues, including how operations for threat response could be adapted and how the current clinical trial infrastructure could be leveraged to facilitate rapid monitoring and assessment of MCM use in a PHE. This chapter is organized conceptually into the following three areas:
- Potential elements of clinical research on MCM use;
- Data and information sharing; and
Although an MCM may be developed and potentially approved for an intended indication (or indications), the specific situation in which it will be used remains unknown until the time of a PHE. This uncertainty makes it difficult to prospectively craft protocol designs and determine elements of clinical research that will be adaptable to a broad range of situations, said Proestel. Individual workshop panelists and participants discussed potential elements of predefined protocols and other considerations for clinical research on MCMs, such as comparator arms, prescreening for contraindications, and leveraging existing clinical networks.
Predefined, Prepositioned Protocols
Proestel noted the importance of having protocols and site agreements prepared in advance of a PHE; however, some aspects of trial design cannot be expedited, said Higgs, including that IRB review and considerations on sharing data can cause delays. Proestel also said that working with a research site that does not have prior experience in PHEs has challenges, compared to working with a site that can rapidly receive and implement a protocol with minimal training. When possible, working with sites that are already up and running is most effective, Higgs added, though lessons can be learned from successful use cases in rapidly standing up clinical research sites in PHEs (see Box 4-1).
Elements to consider when prospectively designing trial protocols for use in monitoring and assessing MCM use during PHEs could include the following:
- Applying Bayesian statistical approaches to data analysis (Higgs);
- Closely involving an independent data and safety monitoring board that regularly reviews data for safety signals (Higgs);
- Availability of the MCM (Higgs); and
- Considering the impact of disaster scale on the appropriateness of the protocol design; for example, a large simple trial versus case reports (Grabenstein).
Higgs referenced an independent panel review of the Ebola response1 that identified lessons learned from the recent outbreak and six areas for improvement, including MCM development. Accordingly, one interagency deliverable currently being addressed is the development of disease- and
1 See https://www.phe.gov/Preparedness/responders/ebola/Documents/EbolaIP.pdf (accessed August 23, 2017).
pathogen-agnostic interagency protocols for use during future outbreaks or pandemics. Although many stakeholders agree that having a protocol in advance of a PHE is valuable (see Box 4-2), there is some debate about the usefulness of this exercise, she said, given that the end products will need to be further adapted to each specific scenario. Higgs also emphasized that this exercise is for international, not domestic scenarios, but it provides important lessons on the need for leadership and responsibility during PHEs.
When designing clinical protocols for products approved under the Animal Rule, it can be difficult to define a comparator for establishing efficacy, said Proestel. Even when there are well-collected, standardized data on the use of an MCM in an emergency, it is difficult to determine the effectiveness of an MCM without a comparator. However, a placebo arm may not be a viable option in a PHE. Two options for control arms, both notably imperfect, could be individuals who declined administration of the MCM or trial arms based on data collected on patients who were administered the MCM from time of exposure to the hazard to first MCM dosage, said Proestel.
Based on her work within different biopreparedness programs at GSK, Peppercorn noted that having an appropriate control group for MCM
clinical studies is essential to interpreting the data. For example, the MCM antibody to the anthrax toxin was approved under the Animal Rule and has not yet been administered to patients infected with anthrax. Assessment of safety and effectiveness in a field study are among the post-approval commitments for this product, she said. If and when the antibody is deployed from SNS, whether for a confined local or mass exposure event, GSK is prepared to collect safety and efficacy data. As previously noted, however, she stated that such a study is difficult to design when the location and the scale of the event are unknown.
Prescreening for Contraindications
In considering how to safely deliver MCMs to affected populations, it is important to know what medical screening is necessary to safely deliver an MCM to the public, said Lance. Previously performed pre-clinical or clinical research should provide as much information as possible on potential contraindications. These and other potential risks should be built into an algorithm to support safe administration of an MCM. For example, during the 2003 smallpox vaccination campaign, CDC reported that data gathered through its Vaccine Adverse Event Reporting System indicated increased risk for adverse cardiac events in response to vaccination. At the time, CDC noted a need for increased surveillance to confirm a relationship between the vaccination and adverse event.
Leveraging Existing Clinical Networks
Cobb recalled a point made by Lushniak about the need for connectivity throughout MCM monitoring and assessment, creating a feedback loop from the local to state to national levels (see Chapter 2). From a research operations standpoint, Proestel emphasized the importance of using existing clinical trial networks to achieve this connectivity. Setting up a new study site takes a long time, he said, especially with regard to training new site staff. In an emergency, predesigned protocols that could be sent to predetermined, experienced clinical sites, would be the most efficient approach, he said (see earlier section in this chapter on Predefined, Prepositioned Protocols and the Leveraging Existing Resources to Supply Data Needs section in Chapter 5).
Maher said FDA and others have been working to understand how clinical trial networks could be used most effectively in a PHE. One potential challenge would occur if the affected population is not located near a site where the clinical trial network has prepositioned protocols. What will it take to shift the protocol to a new site? A workshop participant suggested that the clinical research capacities of NIH grantees should be catalogued to be leveraged in the event of a PHE.
Lushniak emphasized the need to accelerate data sharing. He asked whether CMS and the Office of the National Coordinator for Health Information Technology (ONC) could accelerate sharing capabilities by the EHR vendors, particularly as CMS is paying for these systems.
Higgs suggested creating a historical database of the best IT development practices that have allowed EHR systems to share information that is traditionally in the domain of public health data or monitoring data. What is the best architecture to make a data system that is extensible, so that it is independent of the original domain?
Leremy Colf of HHS pointed out that the Paperwork Reduction Act prevents federal agencies from giving local jurisdictions a list of questions and information to be collected in a disaster. Any forms, questions, or surveys must be approved by the Office of Management and Budget before they can be disseminated to prevent them from being overly burdensome on the public. The approval process generally takes about 8 months, he said, which is not suitable in a disaster scenario. HHS has been working to decrease the regulatory time it takes to get local jurisdictions the information they want, and the 21st Century Cures Act, passed in December 2016, allows for waivers of the Paperwork Reduction Act in a PHE.
Communication and Outreach to Affected Populations
Higgs emphasized that the rules of clinical science do not change in an emergency. Safety and efficacy must be determined using data that are collected in a systematic way through rigorously designed studies, and there is an ethical obligation to communicate to affected populations the inherent risk of the proposed research.
Amanda Fuller Moore of Public Health Preparedness and Response, Division of Public Health, North Carolina Department of Health and Human Services, said every resident of a state is a stakeholder. After North Carolina authorities recognized that segments of the population were being missed during public health messaging campaigns, the state began looking at how to best get information to individuals. For example, newspapers had fallen out of favor as a public health communication medium until it was learned that a large segment of the population of North Carolina receive their news from print newspapers. The state has also taken on a project to engage individual community leaders and community groups as potential partners in sharing information with their neighbors.
It is important to understand needs and communicate across stakeholders at all levels (federal to local) to ensure not only that everyone is receiving a message, but receiving the correct message for their situation, said Moore. She added that information shared with other states is not necessarily the same information needed by local jurisdictions where the event is happening.
Cooper stressed the importance of returning the results of any data analysis back to the community. Communities will engage, but it is reasonable for them to expect to get something back from that investment of their time and resources.
Barriers to Information Sharing
Deven McGraw, deputy director, Health Information Privacy, Office for Civil Rights (OCR), HHS, raised the issue of data security. Data collection creates targets for cyberattacks, and it will be important to ensure that data are stored and accessed under reasonable security measures, she said. Data breaches reduce public trust in sharing their data and affects future data collection efforts. It is not possible to have zero risk, she said, but there are approaches that can reduce the risk of a data breach considerably. Cobb suggested there will need to be a culture shift in how health information is handled that addresses patient protection and privacy. He noted that other nations have addressed this issue through a unique patient identifier.
Balancing Patient Privacy and Public Health Efforts
Scott Needle, member of the American Academy of Pediatrics and a primary care pediatrician, remarked that there is sometimes a tension between maintaining patient privacy and furthering public health efforts that results in decreased sharing of information. For example, the Florida state surgeon general released a statement during the Zika epidemic saying that there were seven babies born in the state who had been neurologically affected by the Zika virus, he said. However, the state was steadfast against releasing further details, such as disclosing the region in which the cases were identified and whether the cases were contracted locally or were travel related. Access to this kind of information would help providers better care for their patients and better assess and communicate risks to patients, said Needle. In order for health care providers to assist in public health efforts, he said, there needs to be a culture of increased data sharing.
Health Insurance Portability and Accountability Act
Workshop participants discussed the potential implications of HIPAA2 on sharing patient-level medical data on MCM use during PHEs (see Box 4-3). As the entity responsible for enforcing HIPAA, OCR regulates most of the potential data sources for monitoring and assessing MCM use, McGraw said. Most hospitals, physicians, and all health plans are covered by HIPAA rules that govern how they use and share data, and the vendors they work with are also largely covered as business associates. However, HIPAA provisions governing the sharing of information for public health practice are fairly permissive. Furthermore, once the data are shared, the entity that shared them is not responsible for any subsequent use, so long as they provided the data in a way that was HIPAA compliant. There are pathways for data disclosure, as well as regulations, standards, and methodology for de-identification of data, if necessary, that stakeholders should understand to ensure that information sharing or analysis is consistent with the law, said McGraw. Other HIPAA issues discussed by individual workshop participants included the impact of state laws, consumer-shared data, and emergency department data.
State law HIPAA governs many of the entities that would contribute or use data for MCM monitoring and assessment; however, state law also plays a strong role, McGraw said. First, HIPAA does not preempt stronger state privacy laws (e.g., if state laws governing particular types of health
information place greater restrictions on sharing without the expressed consent of the patient). Second, state laws must conform with the public health permission clause in HIPAA, which allows entities covered by the rule to use and share information for public health purposes. This provision is very much reliant on the scope of authority of the public health entity involved, McGraw said. If the public health authority in a state is fairly constrained, then a sophisticated compliance officer will question whether there is enough permission in the law for them to be able to use or disclose protected health information. Less sophisticated staff, who might not read the regulations to the letter, might readily interpret the public health permitted use to mean they can release the information. This variety of interpretation of the law creates a dilemma for HHS with regard to enforcement if, in fact, the authority for the public health department to collect and use that information is not in the law, she said. However, the intent is for entities to
be able to share information for public health purposes, and it is unlikely that such a case would be pursued. Nevertheless, if state laws can provide that authority, and the pathway for disclosure is clear, it does reduce delays and allows information to flow without any uncertainty.
Consumer-shared health data McGraw pointed out that data collected by consumer-facing technologies such as wearable health devices and data from social media and social networking sites would not be covered by HIPAA. Their policies regarding sharing of user information are set in their user agreements. User agreements are enforced by the Federal Trade Commission and are expected to be transparent with regard to how the data are accessed, used, and shared. In response to a suggestion by Jeff Coughlin, senior director, Federal and State Affairs, Healthcare Information Management and Systems Society, in order to empower patients to voluntarily contribute their information, McGraw noted that HIPAA provides a pathway for the individual contribution of health data: patients have the right to obtain a copy of their medical data and to have it sent directly to the entity of their choice, which could include a public health authority.
Emergency department data Needle said it can be time-consuming for providers to obtain information from local emergency departments. Though such sharing is permitted under HIPAA, providers must first obtain a release from their patient to be able to receive their information.
Leadership and Collaboration
In addition to clinical science expertise, leadership and responsibility during PHEs are needed moving forward, said Higgs. Establishing leadership is currently being done for international responses, but not domestic responses, she added. The Ebola response exemplified the necessity of identifying a lead U.S. government agency for a given crisis and providing clarity regarding the respective responsibilities of government agencies for both preparedness and response efforts. As previously noted in this section, an independent panel review following the Ebola crisis identified different government agencies as leads for future international PHEs, including tasking NIH with leading biomedical research during crises. Higgs also stressed the importance of working with local emergency operations systems, listening to the local experts, and establishing partnerships based on trust and agreement on the principles for the conduct of the research.
Keralis said the human tendency is to want to work with people who speak the same technical language, and she emphasized the need to bring
people together across silos for planning and exercises in advance of an emergency. Peer-to-peer sessions are also very helpful, she said, and may be more effective than top-down communication approaches. For example, state legal departments can be brought together to share cautionary tales, solutions, and best practices from past experiences. Marc Overhage, chief health informatics officer at Cerner, said that information silos often result in limited levels of education and awareness at the local level. He reiterated the importance of practicing response plans before a PHE to ensure that stakeholders are aware of regulations and their implications. Lee agreed with the importance of bringing the key stakeholders to the table, but she added that it can be logistically difficult to assemble such a large group, and meetings tend to be hours long. In addition, stakeholders are very enthusiastic at the table, but are often frustrated when they return home and try to implement ideas. Several weeks later they may have given up, she said, not because they have given up on the issue, but they simply do not know how to move forward.
Cobb added that in addition to filling leadership gaps, there is also a need to address funding. He said that none of the groups in the network of networks he described (see Chapter 2) exist solely to conduct preparedness research. Each group conducts research in their particular domain; however, they are able to pivot rapidly to initiate pre-defined clinical trial protocols, if needed. A potential concern is that these networks are not exercising the preparedness systems on a regular basis due to a lack of leadership, funding, or both. The investigators and professional organizations encompassed in these networks are committed to moving forward if funding were available, Cobb added, but at present it is not justifiable to dedicate 20 to 40 percent of their time to work on an initiative that is unfunded. David Reddick, chief strategy officer and co-founder of Bio-Defense Network, pointed out that the new CMS Emergency Preparedness Rule requires all CMS providers to have participated in a major community preparedness exercise.3 Perhaps that will be the impetus for clinical networks to begin their preparedness exercises, said Reddick.
Funding is a perennial challenge, said Barishansky. Most health departments are stretched financially, and future funding for some critical public health preparedness grants will only remain stable or will be reduced. He added that many state and local health departments fund their public health preparedness programs entirely with these grants. Moore added that
3 See https://www.cms.gov/Medicare/Provider-Enrollment-and-Certification/SurveyCertEmergPrep/Emergency-Prep-Rule.html (accessed August 23, 2017).
preparedness funding is also used to support everyday activities that are key systems in a PHE as well. For example, data from routine surveillance systems and inpatient monitoring systems can provide potential signals of emerging events or product safety issues.
Clinical Research Training
Cairns suggested that including a research component in disaster preparedness and response training could introduce disaster responders to clinical research in disaster settings and increase their comfort level with collecting data for research. As there is limited time for disaster response training, Cairns said, the intent would not be to develop responders into researchers, but to provide a broader perspective. He recalled comments by Chip Hughes of the National Institute of Environmental Health Sciences at an IOM workshop regarding the challenges of collecting data from the point of initial contact in environmental disasters (e.g., the Gulf oil spill or hurricanes). Cairns recalled that Hughes mentioned the availability of a standardized disaster response training module for chemical environmental disasters and had proposed adding a research module. Higgs noted there are discussions about integrating basic training on research into disaster training and preparedness at the global response level within the Office of U.S. Foreign Disaster Assistance.
Chris Kratochvil, a principal investigator for the National Ebola Training Education Center, said the center conducts training, education, and site visits for regional Ebola treatment centers across the United States, with funding from CDC and ASPR.4 Recent efforts have focused on setting up a research infrastructure within that network, including a central IRB, case report forms, and a data repository. They are also developing training protocols for the nurses at the Ebola treatment centers to be able to support clinical research. The initiative is now being expanded beyond Ebola to encompass highly infectious diseases in general. Although this initiative is on a small scale, including just the 10 regional treatment centers, it is a model that could be useful more broadly.
Colf said relevant training programs are available, and he referred participants to the National Center for Disaster Medicine and Public Health, which conducts education, training, and research around disaster medicine and public health preparedness.5 Another example is the Department of the Interior, which trains researchers on how to follow the incident command system in a disaster so they can better collect data without hindering responders.
A participant added that responses at the federal public health level and the local health care facility level are entirely different (e.g., Ebola and pandemic influenza). Being prepared for one does not necessarily translate to being prepared for the other. Organizationally, the right kind of training needs to be provided at the right time to support local health organizations. Communities cannot be expected to be well-prepared to respond to every type of challenge that federal public health agencies are responsible for responding to.
Public health preparedness and response has become quite robust over the past 15 years, Barishansky said. However, there is an “administrative preparedness” angle to MCM monitoring and assessment that is often not as robust as it should be, he said. How do existing administrative functions (e.g., policies and authorities) affect response and the ability to collect the necessary information? Moore agreed that administrative preparedness is a concern and is often not thought of until an event occurs and information is not being shared as needed. For example, the public health system in North Carolina is decentralized, she said, and local health directors have
the primary authority in response and surveillance efforts. As such, there are often delays and confusion around sharing of data among hospitals, local jurisdictions, and state agencies. She emphasized the need to prepare in advance of the next PHE to determine how administrative preparedness can be addressed to inform how information can best be collected and shared. Cobb highlighted what he called a “healthy tension” between obtaining information needed during a health emergency and protecting the privacy of patients (see Box 4-4).
Barishansky said that a state department of health serves its constituents and visitors to the state and strives to ensure that they remain healthy. In this regard, the Commonwealth of Pennsylvania Department of Health considers MCMs to be a critical component of public health preparedness. When considering the ramifications of a PHE, he asks how will it impact the state, the neighboring states, or a certain area of the state (county, town), with the intent of understanding how communities might be made more resilient from a public health perspective. Barishansky remarked that many states lack the necessary statutory authority to conduct MCM monitoring and assessment. He suggested that a template could be developed to assist states in developing statutory language.
This page intentionally left blank.