Developing Effective Surveillance and Monitoring Systems: Future Directions and Resource Needs
Nancy E. Adler, University of California, San Francisco
John C. Bailar III, The University of Chicago
Lynn R. Goldman, Johns Hopkins University
Maureen Y. Lichtveld, Tulane University
Linda A. McCauley, Emory University
Kenneth Olden, Hunter College of the City University of New York
Linda Rosenstock, University of California, Los Angeles
David A. Savitz, Mount Sinai School of Medicine
One of the main objectives of the workshop was to examine options for building a framework for short- and long-term surveillance systems that monitor the Gulf oil spill’s effects on human health. In the final session of the workshop, the panelists considered the broad range of scientific evidence presented throughout the workshop, discussed individual participant’s suggestions, and explored options for, and components of, a public health surveillance system. David Savitz moderated the panel discussion. This chapter summarizes the discussion that took place, while also drawing on relevant content presented throughout the workshop (i.e., as summarized in other chapters of this report). The suggestions compiled here were personal recommendations made by individual panelists. These recommendations are part of the factual summary of the workshop and should not be construed as reflecting consensus recommendations or findings by the National Academies, the Institute of Medicine, or workshop participants as a whole.
SUGGESTED DIMENSIONS AND COMPONENTS OF A SURVEILLANCE FRAMEWORK
A number of presenters throughout the workshop suggested specific dimensions, components, and resources to guide development of a comprehensive surveillance system and related research activities. Based on comments offered throughout the workshop, Nancy Adler proposed six possible dimensions to consider when designing such a framework. These dimensions include key characteristics, population(s), content, processes, use of existing data sources, and unanswered questions.
Key Characteristics. Certain characteristics help define the scope of proposed frameworks for surveillance systems. Adler suggested that a surveillance framework must be long-term, flexible, multi-layered, and integrated.
Population(s). A surveillance system needs to monitor all populations in order to establish a baseline for comparison. However, a framework must also focus on particularly vulnerable or at-risk populations, such as children, specific ethnic groups, or individuals with preexisting conditions or genetic susceptibilities.
Content. Deciding what outcomes and methods of measurement best capture exposures and the effects of those exposures is imperative. Throughout the workshop, a number of participants suggested that a surveillance system must establish baseline rates for comparison. Exposures can be measured objectively and subjectively, and both are important, especially when monitoring the mental health of a population. Additional methods of collecting data include biospecimens and syndromic responses.
Processes. A successful public health surveillance system will require public input from a wide variety of stakeholders, including communities and government agencies. To set priorities, processes must be in place to interact and coordinate across different sectors. With litigation threatening to compromise or limit research and surveillance activities, there must also be a process to guard the integrity of the surveillance system.
Use of Existing Data Sources. A framework for surveillance can build on existing data sets or data-collection activities. Adler noted that a number of population surveys easily could be modified to target information relevant to the Gulf oil spill. Addition-
ally, baseline data already exists for a number of measures relevant to a surveillance system.
Unanswered Questions. Identifying knowledge gaps and pertinent questions can be equally as informative to developing a framework for surveillance systems as what is already known. Based on discussions over the course of the workshop, Adler stated that many questions still remain, such as which outcomes to monitor, which covariants to measure and control for, whether to use passive or active data collection, and who should be responsible for different areas within a surveillance system.
As briefly noted in Chapter 2, Maureen Lichtveld also proposed six components of a multi-pronged action plan (see Box 7-1) to assess exposures, identify the risk of these exposures, and communicate the results and findings to affected populations throughout the Gulf Region.
Components of a Multi-Pronged Action Plan
MAJOR OVERARCHING THEMES OF WORKSHOP DISCUSSIONS: PRINCIPLES OF A FRAMEWORK FOR SURVEILLANCE
The amount of uncertainty surrounding the Gulf oil disaster underscores a need to fill gaps in the current state of knowledge. Additional data can provide raw material for scientific discovery, observation, and theory, as well as desired information and answers most needed by at-risk populations, said Savitz. Within the context of surveillance, additional data can help identify high-risk hazards, identify the most at-risk populations, evaluate service needs and current capacities, and drive actions to better prevent or mitigate adverse health effects from future disasters. Moreover, as Nicole Lurie noted (see Chapter 1) surveillance systems and related research may have the potential to generate knowledge that could influence general public health responses and the overall delivery of health care services.
Complexity and Uncertainty
Assessing the effects of the Gulf of Mexico oil spill on human health is complex, and many questions remain about the hazards and risks posed to different populations. As elaborated throughout the workshop, there are many unanswered questions about at-risk populations, potential hazards (including exposures) to human health, the potential effects of these hazards, and how best to minimize hazards. Some of the uncertainty stems from the enormity and unprecedented magnitude and scope of the disaster. For example, as John Howard and Scott Barnhart discussed (see Chapter 2), proximity to the oil spill and response activities is a major risk factor for exposure by inhalation, ingestion, or dermal contact (see Figure 2-1), and everyone in the Gulf region, arguably even outside the region, is potentially vulnerable to feelings of anxiety and other negative psychological symptoms, conditions, or disorders.
But much of the uncertainty stems from the scarcity of scientific evidence about the types of risks to human health associated with various hazards. As Kenneth Olden remarked, “We are living in a state of toxic ignorance.” Establishing a cause-effect relationship between a specific exposure and any given outcome is rarely straightforward, and the ongoing nature of the Gulf oil spill makes surveillance design and analysis more challenging. Edward Overton explained that very few of the thou-
sands of chemical compounds in oil have been tested for toxicity (see Chapter 1), and Savitz urged further investigation into the toxicology of specific chemicals. The use of chemical dispersants and controlled burns further complicates identification of high-risk chemical exposures. Several panelists and other participants commented about the lack of data on effects of exposure to “real-world” mixtures of chemicals. Thomas Bernard (see Chapter 3) explained that much is known about heat exposure and heat stress, but more information is needed about the cumulative effects of repeated exposures on human health, including heat-related disorders and injuries.
Data on the long-term human health consequences of exposure to oil-spill-related stressors are especially sparse, but are essential, noted Nalini Sathiakumar (see Chapter 3). This type of data becomes more important given the unique nature of the Gulf oil spill, where strenuous occupational conditions and substantial losses of economic and social stability threaten the short- and long-term physical and psychological well-being of the entire Gulf region.
Psychological and Socioeconomic Health
Many participants acknowledged that the Gulf oil disaster poses known risks to the physical health of individuals in the affected regions. However, there is growing concern about the short- and long-term effects on the psychological health of the Gulf region. Past scientific research has demonstrated the seriousness of the mental health consequences of disasters. For example, Palinkas’s work following the Exxon Valdez disaster provided valuable information that identified worsening mental health as a serious problem. Linda Rosenstock opined that mental health status was of “grave concern,” even among those far removed from direct exposure to toxins. Widespread socioeconomic disruption resulting from the Gulf oil spill poses a significant risk of adverse psychological health outcomes, stated Savitz.
A lack of mental health resources only compounds the problem. Howard Osofsky (see Chapter 3) noted the growing psychological impacts of the Gulf oil spill and expressed concern about access to services. Rosenstock suggested establishing a hotline for individuals to call with questions about the spill. Just the existence of a hotline sends the message that feelings of stress or anxiety are not uncommon, even for
someone 300 miles inland. That message, in and of itself, can contribute to improved mental health status, said Rosenstock.
Many workshop participants, including members of the affected communities and Gulf state policy makers, suggested that mental health was one of the most urgent public health concerns (see Chapters 3 and 4). Although the vast majority of surveillance data that had been collected to date by the state health offices was for acute physical illnesses at the time of the workshop, all five state public health officers who attended identified the need for increased and better targeted mental health surveillance as an immediate challenge.
According to numerous participants, there is an immediate need to act based on existing data and resources and to begin new surveillance and research activities. Olden emphasized the importance of identifying evidence-based knowledge that should be applied before generating new knowledge. John Bailar agreed with Olden and emphasized that, as with the data-collection process itself, strategy development is critical but should not delay efforts to gain new knowledge. Both processes need to move forward in parallel, and both need vigorous proponents, said Bailar.
Learning from What Is Already Known
Despite the uncertainty and complexity involving assessment of the Gulf oil disaster’s impact on human health, several panelists remarked that enough information exists to build a solid surveillance framework. For example, activities related to Hurricane Katrina highlight the usefulness of electronic health records. Based on those lessons learned, Daniel Masys suggested ways that electronic health records could be incorporated into surveillance activities related to the Gulf oil disaster (see Chapter 6). As another example, Palinkas’s work following the Exxon Valdez disaster (see Chapter 3) provides information that can lead to a better understanding of what mental health is and how stressful events can affect specific measures of social and psychological health.
Additionally, Rosenstock mentioned some of the many ways that federal agencies involved in disaster response, such as the National
Institute for Occupational Safety and Health (NIOSH), have learned from past mistakes. For example, NIOSH is applying lessons learned from the 2001 World Trade Center disaster to its response to the Gulf oil crisis (e.g., by obtaining a full roster of all exposed workers). As another example, she referred to discussion among state health officers and their ongoing contact with the Centers for Disease Control and Prevention (CDC). Effective ongoing communications of that nature did not exist in the past, said Rosenstock.
New Data Collection
Several panelists called for the collection of as much data as soon as possible. As Bailar explained, people will be moving out of reach, memories will fade, and environmental contaminants will change over time. At the time of the workshop, several types of data were already being collected. As Howard stated, the CDC and NIOSH were compiling worker rosters; conducting surveillance, including worker illness and injury surveillance; and collaborating with states, OSHA, and BP to identify acute trends and potential chronic health effects (see Chapter 5). Osofsky described qualitative, mental health data from focus groups (see Chapters 3 and 6), and mentioned that the Environmental Protection Agency (EPA) was collecting environmental samples, with real-time monitoring data being posted on the EPA website (see Chapter 5). The states of Alabama, Florida, Louisiana, Mississippi, and Texas were engaged in syndromic and reporting surveillance activities (see Chapter 5).
Acknowledging that the workshop’s goal was not to come to consensus on any single issue, Rosenstock remarked that a vast majority of participants appeared to agree that the collection and banking of biological and environmental specimens were needed. During her presentation in an earlier session, Brenda Eskenazi noted several “easy” ways to collect blood specimens for biobanking, for example, as part of routine prenatal alpha-fetoprotein screenings (see Chapter 3). Several participants called for immediate archiving of tissue and environmental samples for later analyses, in the likely event that unforeseen information will be needed in the future. However, Rosenstock cautioned that, because resources are limited, “it is really important that we are very smart and strategic about which of those samples we test for what.”
Discussions also focused on the possibility of leveraging existing resources, mechanisms, or organizations, such as National Institutes of
Health (NIH) funding, to centralize research. Lynn Goldman described the National Science Foundation’s rapid grant mechanism for non-health-related research focused on disasters. The mechanism involves very stringent peer review and has a very quick turnaround. The CDC and the NIH could use similar mechanisms, suggested Goldman, encouraging the agencies to collect public input on how funding is directed.
Immediate actions based on current knowledge can also reduce occupational health hazards and increase worker health protections, noted a number of speakers, including Paul Lioy and Scott Barnhart (see Chapter 2) . Rosenstock explained that there may be gaps in jurisdiction with respect to who has responsibility for worker safety. If so, those regulatory gaps could be addressed quickly. Additionally, because different types of workers are exposed to different types of risks (see Chapter 2), some participants suggested that having a single point of contact to identify and communicate effective use of personal protective equipment would be helpful. It may be necessary to consider deployment of a federalized workforce, or some version thereof, to provide more uniform worker training, said Rosenstock.
As described in Chapter 4, David Abramson suggested that the success of surveillance and monitoring activities is dependent on how results are communicated to the affected populations and incorporated into public health practice. One way to improve health and risk communications with the public involves engaging communities. The suggestions from workshop participants focused on two primary aspects of community engagement: (1) surveillance and related research activities; and (2) risk communication.
Surveillance and Related Research Activities
A number of workshop participants expressed that community participation is key to improving surveillance and research activities be-
cause it enhances community participation and better aligns the goals of the researcher and the community. Data can provide helpful information that can directly benefit communities affected by the Gulf oil spill, as Lichtveld and others explained during the workshop. However, many Gulf communities are skeptical of research participation after Hurricane Katrina.
Recognizing practical limitations of available resources, Olden stated that the foundation of a surveillance system must be built on the priorities expressed by affected communities. He emphasized the importance of listening to the opinions expressed during the public statement session (see Chapter 4) and the need to broaden community input even more. Several other workshop participants echoed Olden’s call for engaging disaster-impacted communities early and often so that research activities are aligned with what communities want and need.
Health and Risk Communication—The Messenger
To establish effective communications between communities and the messengers, efforts to engage the public must begin as early as possible, be as factually accurate as possible, and be expressed in a manner that is relevant to affected communities, remarked Lichtveld. Savitz and others explained that, to this end, the messenger must have a strong command of the data and evidence, as well as a high degree of trustworthiness and credibility.
Unfortunately, it is not readily apparent who affected communities trust. As Abramson explained, some studies indicate that the CDC is trusted as an information source (see Chapter 4). Olden mentioned the National Cancer Institute’s toll-free number that people can dial to hear automated answers to questions. He suggested that the federal government set up a similar information service to provide information to the public.
Not all panelists agreed that having a federal source of information as the clearinghouse is necessarily the best (or should be the only) strategy. Lichtveld and other participants stated that communities in the Gulf region did not trust federal authorities. Noting that the federal government did indeed have the requisite expertise and leadership, Bailar also expressed doubt that the federal government could effectively engage the public. Savitz suggested that state health officials may be in a better posi-
tion to understand the needs of communities within individual states. Linda McCauley added that trustworthy sources may differ by topic.
Panelists also discussed possible nongovernmental messengers. Rosenstock expressed the need for an “independent, respected scientific voice that has experience working across sectors.” Lichtveld opined that centralizing communication efforts could be advantageous but that it was important to involve community leaders who can communicate in a way that will make sense to the people in affected communities. Bailar noted that health care professionals are also in a good position to serve as communicators with their patients. From a different perspective, McCauley described an opportunity to build community capacity by developing a network of youth ambassadors to help collect and disseminate information about data. Moreover, by directly participating in the solution, children could develop a sense of control over their lives, which could turn a frightening experience into an opportunity for empowerment.
Health and Risk Communication—The Message
Although panelists agreed that providing trustworthy information to the public is crucial, there was some disagreement about what to communicate with the public. Abramson stated that it is important to create a two-way dialogue between the messenger and members of the public that allows time for interpretation and deliberation (see Chapter 4). Savitz recommended creating a clearinghouse to identify information of direct relevance, stating clearly what is and is not known. Olden said that many of the questions that individuals have about the Gulf oil spill could have been answered within 2 to 3 days of the initial rig explosion through a telephone hotline.
In addition to accurate and reliable information, Abramson earlier noted that the members of the public may need tools to help them interpret complex or unfamiliar information (see Chapter 4). McCauley added that, because children perceive things differently than adults and because children use emerging technologies (e.g., Twitter), there may be a potential to do something innovative.
Health and Risk Communication—Additional Tensions
Finally, the panelists discussed additional tensions that may affect the relationship between communities and the messenger. As Adler described, “A tension exists between what communities want and what science produces.” For example, to feel at ease, communities may want conclusive evidence that a certain compound will have no adverse effect on health. However, science is designed only to offer “no evidence of an effect.” Additionally, the messenger must strike a balance between the need for prudence and caution with an eye to avoiding undue disruption to the economy, which can add to the cumulative effects on psychological and social well-being, said Savitz.
Creating a framework for surveillance requires more than filling knowledge gaps. Several panelists agreed that the Gulf oil disaster’s magnitude and complexity called for some form of high-level coordination of the various surveillance activities under way. Others suggested some form of centralization to coordinate and accelerate the initiation of research, but it is not immediately clear who should assume that role.
Rosenstock observed that the federal government has emerged as a single voice in charge, more or less, in terms of dealing with the environmental response. She opined that someone at the federal level needs to identify all the routine data being collected and determine how to put those data together to gain a more comprehensive and definitive understanding of the Gulf oil spill’s impacts on human health.
One panelist suggested that perhaps a czar or other high-level person should be appointed to coordinate surveillance across agencies. This would be especially useful as each agency is currently conducting surveys based on its own priorities, and changing existing surveillance and research agendas will require authority to trump other decisions. Savitz suggested that the federal government appoint a National Disaster Research Coordinator. One audience member then nominated Surgeon General Regina Benjamin to serve in that role. Goldman agreed that situations, such as the Gulf oil spill, call for stronger levels of organization in the response, particularly on the federal level, and that agencies need to better coordinate their efforts with respect to using existing resources and increasing the capacity to collect and manage surveillance data.
However, she felt that it was not realistic to have a new directorate because of the number of federal agencies involved in the oil spill response. Rather, she suggested prioritizing existing roles.
Although some panelists agreed that having somebody in the federal government serving in that role could be effective, others did not. McCauley agreed that there needs to be stronger coordination among federal agencies, not only to clarify agency roles and to prioritize surveillance needs and resources, but also to align surveillance system priorities with the priorities of impacted communities. Bailar expressed concern that the public’s trust and confidence in the information provided by the federal government may not be as high as their trust in information provided by a local source. He suggested asking the Gulf state governments to develop a regional consortium that could then co-opt somebody from the federal government or elsewhere to take on this responsibility. Savitz agreed that perhaps a consortium of state health officers—a group that typically interfaces with communities, government, and scientific communities—might be an appropriate level or forum for oversight.
Understanding and addressing the disaster’s potential health consequences will be an ongoing, long-term effort. Many workshop participants, such as Adler, suggested that effective surveillance systems must include long-term outcomes. This requires commitment from the federal, state, and local levels to support these activities over an extended period of time, especially if the goal of surveillance activities is to inform decisions about health care delivery, as described by Lurie in her opening remarks on the opening day of the workshop. Several panelists agreed that state-level public health agencies will require a sustained commitment and continuous funding from the federal government to build the surveillance capacity needed to respond to future, unforeseen disasters.
Investments in public health infrastructure will be a key determinant in whether surveillance activities can be sustained, said Rosenstock, echoing comments from other participants. She noted the gradual erosion in direct governmental support for local, state, and even federal public health that has been occurring since the 1980s. Public health agencies are underfunded and are functioning at a very basic, minimal level, said Rosenstock. By largely targeting resources on biosecurity and other time-specific topics, the public health infrastructure has developed around the
“disease du jour” and does not always build on the core functions that public health is supposed to deliver during disaster response, such as surveillance. When an event like the Gulf oil spill occurs, the expectation is that public health agencies will be able to ensure the safety of public health and track the consequences, said Rosenstock. However, in order to fully deliver on that perceived promise, many participants suggested that those agencies need more continuous investments so they can build the necessary capacities. While not disagreeing with the need for more federal-level resources, Lichtveld emphasized the importance of also building on existing consortia, partnerships, and mechanisms. Goldman also addressed concerns about funding. She said that making funds available for disaster-related health research should be a high priority and explained that funding could supplement existing research grants, or it could be used to launch completely new research initiatives.
Over the course of the workshop, a number of panelists and participants, such as John Hosey, stated that surveillance and research activities should drive action rather than merely generate new knowledge. Thomas Matte defined surveillance as “actionable information to make things better now,” and Bailar defined it as “service to the individual” (see Chapter 5). Accordingly, at several different times throughout the discussion, panelists commented on the importance of not only identifying individuals that need care but also providing a means for those individuals to receive care. For example, Osofsky explained how post-Katrina mental health surveillance has tracked symptoms and guided interventions, services, and resource availability (see Chapters 2 and 5). Rosenstock also expressed a desire to identify areas where additional mental health services are needed. For example, a hotline not only could collect information about potential exposures and adverse outcomes, but also could guide individuals toward available health care services.
Surveillance, Research, and Legal Liability
Although the Department of Health and Human Services (HHS) did not ask the Institute of Medicine (IOM) to consider how the threat of litigation may complicate surveillance and research design and implementation, panel members echoed concerns from workshop participants that litigation may compromise the reliability and the longevity of short-and long-term surveillance and research activities. Certain states have been reluctant to protect research data, explained Goldman. For example, she continued, by allowing research subjects to be recontacted, attorneys
have compromised the integrity of research results. Summarizing comments from workshop participants, Savitz stated that certificates of confidentiality may prevent the misuse of data by attorneys, litigants, government officials, or other interested parties. From another perspective, some speakers such as Wilma Subra (see Chapter 4) expressed concern that fear of litigation persuaded some employers to discourage employees from reporting or seeking treatment for adverse health effects. Bailar encouraged state health officers to work with their respective state attorneys general and federal government representatives to better protect research participants and their data.
Understanding and addressing the potential health consequences of the Gulf disaster will be an ongoing, long-term effort. This workshop was only an early step in that process, said Adler. Several participants commented on the importance of revisiting these questions and issues in the future through additional activities and collaborations.
Although it is difficult to predict the full magnitude of the Gulf oil disaster’s impact on human health, there is an opportunity to help the communities whose well-being is in jeopardy and to prevent or mitigate similar outcomes in the future, according to many workshop participants. Touching on some of Bernard Goldstein’s remarks about the connection between the environment and human health (see Chapter 1), Goldman identified the need for better health-impacts analyses of the policy decisions that led up to this disaster. She noted that the depth of the Deepwater Horizon oil break was not accidental; that is where the oil is. In the future, when additional oil reserves are needed, she opined that drilling decisions may entail entering even riskier environments. Goldman encouraged the health community to become more involved in energy decision making and policy development.
In conclusion, Adler echoed the opinions of other participants, encouraging federal, state, and local governments; academia; private industry; and community networks and programs to coordinate and share their expertise. By including the public in the development of monitoring and research activities and by protecting the integrity of data collection and analysis, surveillance systems could be developed to accurately inform decision makers and the public about the real risks to the physical and psychological health related to the Deepwater Horizon disaster.