Free PDF AccessFree Resources |
||||||||||||||||
|
||||||||||||||||
|
||||||||||||||||
|
| ||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 4
BACKGROUND
In 2003, the Boston University (BU) Medical Center was awarded a $128 million
grant from the National Institutes of Health (NIH) to build one of two national maximum-
containment laboratories for research on biological pathogens. The National Emerging
Infectious Diseases Laboratory (NEIDL) is part of the National Institute of Allergy and
Infectious Diseases efforts to provide physical infrastructure for the conduct of
biodefense and emerging-infectious-disease research to develop new and improved
approaches to treating, preventing, and diagnosing a variety of bacterial and viral
diseases. Diseases to be studied include biothreat agents and emerging novel pathogens,
such as those which cause Ebola, Marburg, plague, dengue fever, Lassa fever, shigellosis,
and unusual virulent influenzas. The facility will include a biosafety level-4 (BSL-4) and
several BSL-3 containment laboratories housed in a 223,000-ft2 building. Under the
National Environmental Policy Act (NEPA), NIH reviewed the potential impacts of the
NEIDL at its location 3 in Boston's South End. The review concluded that the facility
would not pose a risk to the community. However, the location of the facility on Albany
Street in Boston's South End, which includes environmental justice communities with
large low-income and minority populations, is controversial, and there have been
numerous contentious public meetings about the plans for the facility. Three legal actions
have been filed to stop the funding and construction of the NEIDL.
NIH prepared a document, “Draft Supplementary Risk Assessment and Site
Suitability Analyses” (DSRASSA), regarding the siting and operation of the BU NEIDL
in response to comments from the federal court presiding over a NEPA lawsuit. The
DSRASSA was prepared to supplement NIH's previous assessments of the potential risks
posed by the NEIDL at its current location in Boston.
In 2007, the Massachusetts Executive Office of Energy and Environmental
Affairs (MEOEEA) asked the National Research Council to establish a committee to
provide technical input on the NIH DSRASSA to the MEOEEA. Although the
DSRASSA was prepared in response to comments that arose in federal litigation pursuant
to the NEPA process, the MEOEEA requested a review because it expected the
DSRASSA to be an integral part of the material that would be submitted to it by BU in
fulfillment of Massachusetts Environmental Policy Act (MEPA) requirements.
The National Research Council Committee on Technical Input on the NIH's Draft
Supplementary Risk Assessment and Site Suitability Analyses reviewed the DSRASSA
and discussed its methods and analyses to address the specific questions posed by the
MEOEEA (see below). In November 2007, the committee released its letter report
answering these questions. The committee's letter report was critical of the DSRASSA,
finding that it was not sound and credible, did not adequately identify and thoroughly
develop worst-case scenarios, and did not contain the appropriate level of information to
compare the risks associated with alternative locations. The letter report also raised
specific concerns about agent selection, scenario development, modeling methodology,
consideration of environmental justice issues, and risk communication.
3
Construction of the laboratory building is nearly complete. The remaining issue is whether the BSL-4
component will become operational.
4
OCR for page 5
In March 2008, NIH announced that additional steps would be taken to address
judicial requests and public comments on risks associated with the siting and operation of
the NEIDL (see http://nihblueribbonpanel-bumc-neidl.od.nih.gov/roster.htm for a list of
blue ribbon panel members.) Specifically, NIH established a blue ribbon panel of outside
experts to advise NIH on how to respond to comments by the courts and the public
regarding possible risks associated with the siting and operation of the NEIDL. An early
task of the panel will be to advise NIH on the development of a statement of work for any
risk analyses that may be necessary later. Given prior National Research Council
comments on the DSRASSA, NIH also asked the Research Council to reconvene the
Committee on Technical Input on the NIH's Draft Supplementary Risk Assessment and
Site Suitability Analyses to obtain additional insights on scope and methodologies for
future risk analyses from the NRC Committee.
INTRODUCTION AND COMMITTEE’S CHARGE
The report prepared by the committee and released for publication on November
29, 2007, was a review of a document prepared by NIH (now called the DSRASSA but
also called the NIH study and the DSER in the November 2007 report) for the MEOEEA.
The committee was asked by Massachusetts to carry out a technical review of the
scientific adequacy of the DSRASSA and to address three specific questions:
• Are the scientific analyses in the DSER sound and credible?
• Has the NIH identified representative worst case scenarios?
• Based on comparison of risk associated with alternative locations, is there a greater risk
to public health and safety from the location of the facility in one or another proposed
location?
These three questions were not developed by the committee but rather were negotiated as
part of the statement of task agreed on between the National Research Council and
Massachusetts to guide the committee’s work.
In its November 2007 report, the committee addressed the three questions and
concluded that the DSRASSA had significant deficiencies in scientific adequacy. The
committee described the deficiencies in relation to the three questions. It did not focus
attention on how the deficiencies might be remedied, whether they were limited to the
single work product it reviewed, or whether they reflected problems in previous NIH
work products to assess the potential impacts of the NEIDL.
In the present report, the same committee is responding to a request from NIH to
provide input and assistance on the scope and design of any additional studies that may
be needed to assess the risks associated with the siting and operation of the NEIDL. The
committee’s new statement of task is as follows:
The NRC Committee on Technical Input on the NIH's Draft Supplementary Risk
Assessments and Site Suitability Analyses (DSRASSA) for the Boston University (BU)
National Emerging Infectious Diseases Laboratories (NEIDL) will be reconvened to
provide input on the scope and design of any additional studies that may be needed to
assess risk associated with the siting and operation of the NEIDL. The original NRC
Committee was appointed to provide technical input on the DSRASSA as requested by the
5
OCR for page 6
Massachusetts Executive Office of Energy and Environmental Affairs. The Committee's
letter report, which was released in November 2007, was critical of the NIH's draft
document, with specific concerns raised about agent selection, scenario development,
modeling methodology, consideration of environmental justice issues, and risk
communication. The NIH has now appointed a Blue Ribbon Panel to advise NIH on
responding to judicial and public concerns about the siting and operation of the BU
NEIDL and to recommend any additional risk assessment studies that may be needed.
Given prior NRC comments on the DSRASSA, the NIH is asking the NRC Committee for
input on any further supplementary risk assessments that NIH might undertake. The
reconvened NRC committee will prepare a brief letter report summarizing its views on
the scope (e.g., worst case scenarios, alternative sites, BSL-3 and BSL-4 facilities,
selection of agents, etc.) and methodological approaches to be taken to improve any
additional risk assessment studies NIH prepares and will discuss these views with the
Blue Ribbon Panel in a meeting or conference call after the letter report is delivered to
the NIH.
As in its first report, in addressing this charge from NIH, the committee did not
review the content of previous documents (such as the original environmental impact
statement or environmental impact report) or the scope of what has already been done to
address risk and community concerns. The committee restricted its comments to
suggestions based only on its review of the DSRASSA and on improving the risk
assessments presented therein as input to any additional studies that may be needed to
assess risk associated with the siting and operation of the NEIDL.
The committee prepared this report largely on the basis of the analysis and
discussions that went into the preparation of its November 2007 report, discussions that
were expanded on in a series of conference calls held in April 2008. Additional input
from outside the committee was not solicited beyond the standard National Research
Council review process.
As noted in its previous report, the committee acknowledges here—and wishes to
emphasize—the need for biocontainment laboratories, including BSL-4 laboratories.
These laboratories can conduct valuable scientific research. The committee also
recognizes that BSL-4 facilities are being operated safely in both urban and rural areas.
However, the committee’s view remains that the selection of sites for high-containment
laboratories should be supported by detailed analyses and transparent communication of
the available scientific information regarding possible risks.
COMMITTEE’S SUGGESTIONS AND RECOMMENDATIONS
Risk assessment can and should be used to address both the probability and the
consequences of adverse events, such as the release of human or animal pathogens from a
biocontainment facility that leads to morbidity and mortality. Risk assessment is
generally an appropriate approach for characterizing risk and, when performed well and
directed at answering the right questions, can assist in decision-making (such as siting
decisions) and in addressing public concerns. It provides a framework for organizing
information about a situation that may be highly complex and involve uncertainties with
respect to matters on which experimental data are sparse or absent. Risk assessment does
not generally produce a precise quantitative risk value, but it can be used to summarize
6
OCR for page 7
whatever information is available and provide insights to improve understanding and
suggest new research that is needed. Such understanding, in turn, can be used to design
appropriate mitigation and response strategies. The risk assessment process should be
transparent, and it should inform the parties who have decision responsibility so that they
are better able to make decisions, in this case, about measures to ensure the safe siting,
design, and operation of the laboratory. The communities of professionals in risk analysis
and infectious disease, working together, can provide specific guidance in these fields,
and NIH should seek to use the best knowledge and talent available in the two
communities in any future risk assessments.
Scientifically sound documents can help NIH address the public’s concerns and
provide information requested by the courts about site comparisons. Reviewing the scope
and content of previous project documents is not within the committee’s scope of work,
but the committee is pleased to make suggestions about approaches for the blue ribbon
panel to consider. The committee cannot comment on the cost of such measures or on
what resources are needed.
The committee has elected to structure its suggestions for the blue ribbon panel
around a small number of overarching questions (Kaplan and Garrick, 1981) about the
risks associated with operating the NEIDL:
• What could go wrong? That is, what might be the sequence of events that could
cause an infectious agent to escape the laboratory, set up a chain of transmission,
and cause infectious disease in the surrounding community?
• What are the probabilities of such a sequence of events?
• What would be the consequences of such a sequence of events?
What Could Go Wrong? Scenarios of Release of an Infectious Agent
The committee is aware that the courts asked for a description and evaluation of
“worst-case scenarios” and reiterates that the question of whether NIH had provided
representative worst-case scenarios in the DSRASSA was specifically posed to the
committee by the MEOEEA. However, the committee does not endorse an exclusive
focus on the development of worst-case scenarios as an appropriate procedure for
carrying out risk assessments for the NEIDL or for other facilities of this type. Rather,
the committee suggests two phases of analysis. The first phase is risk assessment
based on a variety of plausible scenarios designed to allow a realistic assessment of
risks associated with the NEIDL in general and to illuminate the comparative risks
to the communities at the three sites evaluated in the DSRASSA. This analysis would
not represent worst-case scenarios; rather, it could lay out realistic situations, such as
protective features in place, public health mitigation strategies in place, and training and
standard operating procedures followed. In a second phase, a highly unlikely but still
credible high-consequence event could be analyzed. This might be referred to as a
worst-case scenario, although the committee encourages NIH to define clearly what it
means if it uses this term. This phase of the analysis could examine possible sequences of
post-release events to explore the magnitude of the possible consequences of a release,
7
OCR for page 8
perhaps by considering such details as highly effective transmission (large R0 4 ) and a
long latent period during which infectious symptoms are nonspecific or not evident. The
effects of limitations in the public health and emergency response systems could also be
analyzed. Any future risk assessments should incorporate sufficient meaningful
biological data in the scenarios to make it possible to understand how the results of the
analyses were reached.
Although engineering and design—and hence safety—of high-containment
biological laboratories have undoubtedly improved greatly with contemporary practices,
accidental releases due to human error or maintenance failures certainly can still occur.
Recent such events include the infection of workers with Brucella at one of Texas A&M
University’s BSL-3 laboratories in 2006; a 1-hour power outage in 2007 at the new BSL-
4 facility of the Centers for Disease Control and Prevention in Atlanta, before work with
pathogens began, wherein the main and backup power systems both failed and the
negative-air-pressure system, a key element of pathogen containment, shut down; and,
also in 2007, a release of foot-and-mouth disease to livestock on farms near the Pirbright
high-containment laboratory in the United Kingdom due to a damaged and leaking
drainage system at the facility (GAO 2007). Scenarios for evaluating the risks posed by
the NEIDL should systematically include potential realistic means of biological-agent
escape and should describe the various safeguards to protect laboratory workers and the
surrounding community. The committee recommends that discussions of potential
agent release include
• Procedural or work-practice failures, including those which lead to worker
exposures and infections.
• Biocontainment-system and equipment failures.
• An appropriate array of malevolent actions.
Within these categories, one could consider contamination of the waste stream from the
laboratory, the effects of power outages, unintentional or malevolent infection of
laboratory workers, and unintentional or malevolent release of laboratory animals or pests
(such as insects capable of serving as disease vectors).
Designing scenarios in this way may also highlight where additional measures
might prove useful for enhancing laboratory safety. The DSRASSA assumed, for
purposes of providing an initial case for modeling, that a release occurred. Scientifically
accurate scenarios that include probabilistic evaluation (see next section for discussion of
probabilistic evaluation) of how a biological agent could be released could lead to
enhanced preventive measures. For example, an assessment might highlight the
importance of laboratory-worker training or of occupational health surveillance. Or it
could lead to the recommendation of interventions instituted in other laboratories, such as
working with vectorborne agents during seasons when the vectors are not circulating in
the community.
4
Theoretically, R0, the basic reproduction number, is defined as the average number of secondary cases
generated by a single primary case during its entire period of infectiousness in a completely susceptible
population (Diekmann and Heesterbeek, 2000).
8
OCR for page 9
In addition to laboratory-related interventions to minimize the occurrence of such
events (that is, prevention measures), risk assessments should address the capabilities of
the medical and public health systems to respond to untoward events (that is, mitigating
measures) at the South End and alternative sites. These measures are especially important
to consider in the context of environmental justice, potentially unequal access to health
care among the three sites, and other factors of importance to the communities. Without
the discussion of preventive and mitigating measures, scenarios do not reflect how the
laboratory is intended to be operated and managed, and risks are obscured to the
detriment of decision-making. Basing scenarios on as much factual information as
possible will make them more relevant and ensure that they portray more accurately the
hazards associated with work in high-containment (BSL-3) and maximum-containment
(BSL-4) laboratories.
What Could Go Wrong? Agents to Consider for Risk Assessment
The characteristics of a particular infectious agent may make it more or less likely
that the agent could lend itself to the establishment of a chain of transmission that leads to
the spread of infection in the community. The DSRASSA analyzed the potential for
disease spread by four pathogens, but all four were of low transmissibility and not likely
to spread beyond the persons initially infected. As noted by the committee in its
November 2007 report, “Because the probability of transmission of disease from one
person to another was set to be low, infections die out, rather than propagate. As a result,
for all four of the agents considered, the risks calculated from the two models are small.”
The committee believes that many of the agents mentioned as expected to be studied at
the NEIDL (Klempner, 2008) are candidate agents with higher transmission rates that
could be addressed in risk assessments regardless of the biosafety level at which they will
be studied.
Including both BSL-3 and BSL-4 agents in any future risk assessments is
appropriate because the reasons for studying a biological agent under BSL-3 vs BSL-4
conditions include factors other than the risk associated with release of an agent (BMBL
2007). These factors include, for example, risk to laboratory workers and whether or not
the agent is endemic. BSL-3 laboratories are used to study biological agents that are
potentially lethal and that are transmissible by the aerosol route. It is thus possible that
BSL-3 agents have greater transmissibility than some BSL-4 agents. BSL-4 agents may
produce higher mortality and lack treatment options, but morbidity is also important in
evaluating risk. In addition, engineered controls are greater in BSL-4 facilities, and it is
possible that risks of human error are greater in BSL-3 laboratories.
The committee recommends that for any future assessments NIH select a
variety of agents with appropriately diverse transmission characteristics
(bloodborne, transmitted on fomites, spread by aerosol, and/or requiring vectors
and the potential for maintenance in existing reservoir species). In addition to portal
of entry into the host, such aspects of transmission as high or low R0, latency, and
incubation periods should be thoroughly addressed. Furthermore, NIH should
describe why specific agents were ultimately selected for the analysis. The committee is
aware of the degree of complexity involved in this task, but it is a cornerstone of
assessing and communicating biological risk reliably and realistically.
9
OCR for page 10
The committee believes that it may be helpful for NIH to clarify for the
public and the courts what agents and forms of agents will not be researched at the
NEIDL for reasons that are likely to apply in the future. Examples may include the
virus that causes smallpox and dry, powdered agents that are more easily spread in
the air. A sound and well-documented rationale could be provided to substantiate why
particular agents or forms of agents will not be studied. The rationale may include legal
or treaty constraints and prohibitions, the fact that government agencies other than NIH
are charged with missions involving work with particular agents and forms of agents, and
circumstances surrounding the acquisition of agents. For example, NIH might clarify that
no offensive biological weapons research will be conducted at the NEIDL, because it is
prohibited by the biological weapons convention (Convention on the Prohibition of the
Development, Production, and Stockpiling of Bacteriological (Biological) and Toxin
Weapons and on Their Destruction). This treaty prohibits signatories from developing,
producing, stockpiling, or otherwise retaining microbial or other biological agents or
toxins, whatever their origin or method of production, of types and in quantities that have
no justification for prophylactic, protective, or other peaceful purposes.
What Are the Probabilities?
Risk assessment addresses both the probability and the consequences of adverse
events. The scenarios and agents discussed above should be used in any future risk
assessments to analyze and communicate the probabilities of adverse events.
The committee recommends that discussions of potential agent release
include probabilistic statements regarding the three categories of release discussed
above:
• Procedural or work-practice failures, including those which lead to
worker exposures and infections.
• Biocontainment-system and equipment failures.
• An appropriate array of malevolent actions.
The development of these probabilistic statements should draw on information
that already exists (for example, Johnson, 2003a, 2003b, 2004) and other risk assessment
documents despite the fact that inherently the information is not comprehensive. NIH
could also update previously generated quantitative measurements of safety records for
its own and other contemporary BSL-3 and BSL-4 laboratories over the last 20 years,
including consideration of recent accidents and exposures to inform the process. Such a
quantitative analysis could include estimates of person-hours worked, numbers of
laboratory-acquired infections, outcomes of infections in workers and the community,
biological agents involved, and other measures relevant to biocontainment work. The
historical experience of biocontainment facilities—both those associated with NIH
activities and the many similar facilities around the world—is that releases of disease
pathogens have been rare. There have been laboratory-acquired infections, but the
resulting diseases have mostly been confined to the facilities’ workers and, in a few
cases, members of their immediate families or health care providers (Harding and Byers,
2006). As noted above, contemporary BSL-3 and BSL- 4 facilities minimize the
probability that a release will occur with extensive equipment and design features,
10
OCR for page 11
laboratory protocols for safety, and rigorous occupational health programs. In addition,
specialized patient isolation facilities are generally available at local hospitals in the event
that workers become ill after an inadvertent exposure.
An infectious agent release could have a variety of consequences, and an
assessment should account for them. These consequences can be conceptualized as a
continuum that ranges from few or no adverse outcomes (requiring minimal or no public
health response) to amplified disease transmission resulting in a public health emergency.
To illustrate the continuum in more detail, the committee has described four possible
scenarios that are points along it. The committee has provided examples for each
scenario. Although the examples represent public health events that have been
documented in the literature, the committee emphasizes that they are not based on
releases from BSL-4 laboratories.
• No subsequent transmission, following a small initial pool of infection.
The agent may fail to establish a productive chain of transmission after
only a few people are initially infected. An example is the 2003
monkeypox outbreak in the United States, which is thought to have been
related to contact with pet rodents.
• Little or no subsequent transmission, following multiple exposures. The
agent may fail to establish a productive chain of transmission after
multiple initial exposures. An example is the intentional contamination of
food with Salmonella that infected hundreds of consumers but failed to
spread in the community.
• Limited transmission that is contained by public health measures. The
agent may establish a successful chain of transmission but be controllable
by public health measures (tens to perhaps hundreds or thousands of
people infected). An example is the SARS outbreak observed in 2003
(Lipsitch et al., 2003).
• Amplified transmission. The agent may establish a chain of transmission
that amplifies rapidly and is not controlled by public health measures,
which may be ineffective or overwhelmed (say, 10,000 people infected).
Examples are the outbreaks of influenza, smallpox, and poliomyelitis
before the availability of effective vaccines for these agents.
A basic risk assessment should begin with these four possible outcomes and assess
how the characteristics of agents that might be studied in the NEIDL influence the
likelihood of each outcome in the event of a release. This basic approach should be a
minimal requirement for risk assessment. A qualitative approach to this assessment might
consider actual events, taking into consideration important differences, such as
metropolitan settings and circumstances, and qualitative consideration of transmissibility
(R0) and the proportion of transmission that occurs before onset of symptoms. R0 is a key
quantity in estimating transmissibility of infectious diseases, and the proportion of
11
OCR for page 12
transmission that occurs before the onset of overt clinical symptoms can affect the
success of public health measures (Fraser et al., 2004).
Even a qualitative analysis of potential outcomes should consider impact of
local characteristics (for example, population density, vector availability, and public
health infrastructure) on the probability of the various outcomes.
More complex approaches to predicting outcome, such as modeling, if pursued,
should be rigorously justified and should be designed to build on this basic analysis (see
next section).
What Would Be the Consequences?
The consequences of a release of an infectious agent from a high-containment
laboratory depend on numerous factors, such as the characteristics of the agent, the
pathway by which it is spread, and the size and characteristics of the population that is
exposed to it. The major concern is the potential for community outbreaks of disease,
taking into account both morbidity and mortality.
The previous section discussed the need for an assessment of agents and the
probability of different outcomes in the event of a release. This section discusses
modeling, which is of course, another way of assessing how the disease caused by an
agent may spread. Modeling may also be an important tool in devising appropriate
mitigating strategies.
Calculating the outcome of a release of a biological agent with models is
extraordinarily difficult. The basic test of a model is whether it can replicate the various
types of outcomes that are known to happen, but our understanding of any individual
agent is incomplete, to say the least. Furthermore, the biology of agents within
experimentally infected animals or infected humans is much better understood than the
process of transmission, about which relatively little is known although it is a major
parameter in determining the results of a release. For example, the observation that there
are “superspreaders”, a small proportion of hosts that account for a large portion of the
amplification of an epidemic, makes estimates of average transmission rates highly
questionable. Likewise, it is difficult to estimate the number of contacts between people
although recent estimates of age-specific contact rates from surveys that are relevant for
respiratory spread of infectious diseases have become available for some populations
(Mossong et al., 2008). The ability of a single model to simulate accurately both the
transmission of an aerosol-transmissible agent and that of a fomite-transmitted agent is
questionable. These uncertainties and complexities compound as the number of model
parameters increases.
There is no consensus on an approach to model all, or even many, infectious
diseases. In the absence of an accepted approach, simplicity has advantages: the behavior
of simple models is relatively well understood, and the effects of changing inputs are
relatively transparent. More complexity and detail may not add to confidence or accuracy
of model results, particularly if the data used to develop input are scant and there are
many uncertainties.
In short, although mathematical models of infectious diseases at the population
level may provide results that can give us perspective and insight as to how and why
infectious diseases cause epidemics, there is great complexity in using them and in
interpreting their results. The use of models cannot make up for what is often a deficiency
12
OCR for page 13
of biological and other data, so it is essential that the judgment of epidemiologists,
infectious disease specialists, and microbial risk assessors be applied to the interpretation
of model results. If NIH decides that there is a compelling rationale for the use of
mathematical modeling in any future risk assessments, the modeling must be done
credibly, transparently, and to professional standards by an experienced team of
epidemiological modelers and microbial risk assessors. The results should be
interpreted in light of the strength of the data used to develop them.
If modeling is deemed necessary to study the effects of an infectious agent release
into a community, the type of model used should be considered case by case. If the
objective is to evaluate epidemic characteristics—such as size, peak, and duration—
dynamic compartmental epidemic models based on differential equations can be useful
(Anderson and May, 1991). Most mathematical models used in the literature to date are
simple compartmental models of various levels of complexity, such as those used to
study the SARS epidemic (see, for example, Lipsitch et al., 2003). Dynamic models
based on differential equations are tractable for systematic uncertainty and sensitivity
analyses. In contrast, large-scale agent-based models are increasingly used to assess the
role of specific control interventions in specific settings. However, these large-scale
agent-based models are typically difficult to calibrate and require large-scale computing
resources.
Independently of the type of approach used, the model-building procedure and the
procedure for assigning values to parameters need to be clearly laid out and justified. For
example, which parameter values are supported by the literature, which are
estimated from empirical data, and how estimates were derived need to be
transparent and clearly presented. The level of detail in a model should be defended
with appropriate empirical data and reference to appropriate scientific literature.
The infectious disease transmission potential and uncertainty of transmission must
be quantified to determine the disease related impact on the population of a release of an
infectious agent. Any modeling exercise should be accompanied by thorough
uncertainty and sensitivity analyses. As pointed out in the committee’s November 2007
report, assessing the uncertainty of parameter values and the sensitivity of model outputs
to them is crucial. Uncertainty analysis includes assessment of the uncertainty in
epidemic size, peak, and duration as parameter values vary within plausible ranges. It is
especially important to consider the impact of values used for infectious disease
transmission potential. Because each set of plausible model values is not equally likely,
values can be drawn from appropriate probability distributions with simple random
sampling or Latin hypercube sampling (see, for example, Blower and Dowlatabadi, 1994;
Chowell et al., 2004). Similarly, sensitivity analysis should be conducted to assess the
effects of changes in parameter values on specific model outputs, such as those described
above. A sensitivity analysis will help to rank parameter values according to the size of
their effect on model output.
As discussed in the qualitative description above, modeling approaches should
also consider the impact of local conditions (for example, population density, vector
availability, and public health infrastructure) on the consequences. It would be useful to
consider the possibility that different disease spread outcomes have different implications
for the population immediately surrounding the laboratory (see the next section).
13
OCR for page 14
Including Community Characteristics
The characteristics of the surrounding community—such as its racial, ethnic, and
socioeconomic composition; its access to health care and health services; and the
environmental stressors it faces—should be taken into account in the risk assessment and
analysis. Urban communities often face environmental and other stressors that wealthier
communities do not face. These factors are important because communities, such as the
South Boston neighborhoods that surround the NEIDL, face challenges that could affect,
among other things, the transmission of infectious disease, the health consequences, and
the scope and deployment of public health resources required for response. It is also
important to include these factors in an analysis because they form the basis of many
community and environmental justice concerns about the siting of the NEIDL. Site
selection can contribute to the probability of various possible outcomes. The potential for
various outcomes to have different effects on sites is noted above on page 13.
If modeling is used, these factors could be incorporated into the modeling
exercise (see, for example, Halloran et al., 2008). If another approach is chosen, or if a
modeling approach that does not accommodate the inclusion of environmental justice
concerns is used, the risk assessment should adopt another quantitative or qualitative
technique that reflects the community’s attributes.
Improving Communication of Risk
In its November 2007 report, the committee discussed risk communication
aspects of the DSRASSA. The report noted that particularly in cases where there is strong
public interest, such as this siting decision, it is important to develop presentations and
documents that are transparent and complete and that clearly address the concerns of
affected and other interested parties. There are many information resources on risk
assessment and risk communication, and NIH should use the wisdom accumulated in the
published literature on effective communication of risk. Although the committee has not
described the specifics of risk communication in this report, it notes that a recent article
by Race (2008) analyzes public review processes and risk communication with respect to
a number of high-containment laboratories recently built or under construction. Many of
the laboratories that generated serious controversies had key issues in common, including
concerns about trust, transparency, and the reporting of accidents. Lofstedt (2002) and
Fell and Bailey (2005) also discuss risk communication in connection with laboratory
siting. Finally, the committee refers the blue ribbon panel to the risk communication
concepts discussed in the National Research Council reports Improving Risk
Communication (1989) and Understanding Risk: Informing Decisions in a Democratic
Society (1996).
14
