ML20127E799

Informs Commission of Staff Intention to Participate in multi-agency Study in Risk Characterization Sponsored by Natl Research Council

ML20127E799
Person / Time
Issue date:	10/28/1992
From:	Taylor J NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO)
To:
References
SECY-92-367, NUDOCS 9211050390
Download: ML20127E799 (16)

Text

.}

g, s

l'.

.o s s e n e s e e o o s e e o u wb o a s i

SE!. EASED TOTriE POR f'"%,

I n/Rol9?

dam s

... u........ o.. m.. o.

y 2

o, 1

%...../

October 28, 1997 SECY-92-367 The Commiss([QfQTmation) for:

l l-From:

James M. Taylor l-Executive Director for Operations l-Sub.iect:

STAFF PLANS TO STUDY RISK CHARACTERIZATI0ft l

Puroose:

To inform the C, mission of the staff's intention to L

participate in a multuagency study in risk characterization L

sponsored by the National Research Council.

l l

piccussion:

The Commission on Behavioral and Social Science and L

Education (CBASSE) of the National Research Council has l

proposed a study on risk characterization. The proposal i

follows up one of the recommendations of the National j'

Research Council's report on Improving Risk Communication (John F. Ahearne, Chairman, et. al., National Academy Press,1989). A copy of the draft prospectus is enclosed.

L' Robert Coppock and Nancy Crowell of the CBASSE staff met with Eric Beckjord on September 13, 1992, to request NRC i

support'for the project, a 2-year study expected to get underway early in 1993. CBASSE is approaching a number of government agencies for support, including EPA, DOE, 000, and NIH, to name some of the larger ones.

CBASSE also expects contributions on the part of industry for a small

~

part of the budget.

Mr. Coppock mentioned an expected budget of $600-700K to i

~_

cover the cost of the study, a workshop in the first half of 1993, and the final report.

He has asked for a j

l Z

commitment from the NRC of funds and would like a reply by early October. We have agreed to provide a speaker on PRA

=

T at the workshop.

w NOTE:

TO BE MADE PUBLICLY AVAILABLE I

IN 10 WORKING DAYS FROM THE DATE OF THIS PAPER

Contact:

E. Beckjord, RES 492-3700 g_ -

9 A tto 50 390

%I) s to n

-J

,_ _ __..,....,...,_. ~....._.

.n.

3.g

-ej; g

-r-4-

- ~ y y:o

.s e.~

,. _ =

m-.,,._.,_o,

1. % 'r

-m t-;.y_1 p

T.-..'

s.

~

i m.

x g

, Y.

y y' 5

- )l '.

N

~

'g_'

l q ry'N' 4--

• k j

j y(, r,

'3--

,i 7

.y D' :M 4

e

"%7_? ' L,

5.? ?

r, ej I-

+ -

A g

'['qi,k'. #...

i ;

T

)

jy e

+5_'

}

a

~ '

5 g -._"

i y y.

_b o

i

{

g T'

i

-_.s

.,.7-b l

4 I

3 I

-I

~ _

' j l

y 4

'l I

a l

I f.'_

J

'I t.

't J

F d

-.L.

k. ~-

I f

f 1..

a f

g

-1 e

r

(

-, ~.._

.n

.w 2-L The Comm'i_ss_ioners I believe that this study.c'an provide.'a~ valuable synthesis

'of developments in risk characterization over the past-10 years that could improve the use and presentation of risk-information in regulatory judgments in lmany fields,-

including nuclear facility and waste' disposal ~ regulation.

-The goal of a risk characterization is to translate the quantitative results of a PRA, together with their associated uncertainties, into a form that is usable, readily understandabl(. and supports regulatory. decision-making.

The risk characterization explicitly takes quantitative and qualitative information-from the PRA and contextual- _information from outside the PRA (e.g. public-perception of risk)_into account to assist the decision makers in achieving an appropriate balance between risk:and' societal costs and benefits.

The study is more likely to have influence over the next 10 years, rather than have an immediate research application over the short term.- Our judgment-is that the-project is worthwhile for the NRC to support at a level-of about -$50,000 ; total for the 2-year study period.

Since the

~

NRC would be one of several sponsors and would have no control over the direction of the study,. our support.should not be interpreted as advocating the use of nuclear. power.

Coordination:

The Office of the General Counsel has no legal. objection to NRC taking this action.

This has been coordinated with:

NRC's representative to the FCCSET ad hoc Working Group on-Risk Assessment and he has no objection to-this approach.-

==

Conclusion:==

The Commission should note that the staff plans-to inform CBASSE of NRC's intent to support their proposed study on risk characterization: at-a level of $50,000 for a 2-year-period.

,f ylor ecutive Director for Operations-

Enclosure:

As stated DISTRIBUTION:

Commissioners SECY OGC OCAA OIG IP OCA OPA OPP EDO

-ACRS

~.

+

,p,

NATIONAL ACADEMY OF SCIENCES -

NATIONAL RESEARCH COUNCIL COMMISSION'ON BEHAVIORAL AND SOCIAL SCIENCES AND ED Study on Risk Characterization DRAFT PROSPECTUS It is increasingly clear that the way the nation handles risk often breaks down at a point in the process that can be termed " risk characterization." At this _ stage in a risk assessment is translated into a form usable by a risk manager or individual ~ d maker. Oversimplifying the science or skewing the results through selectivity can lead inappropriate use of scientific information in risk management decisions. Providing f information that includes descriptions of assumptions and uncertainties but does not addres key concerns of the audience can lead to lack of trust in the risk analysis. The Com on Behavioral and Social Sciences and Education (CBASSE) in cooperation with the Commission on Geosciences, Environment, and Resources (CGER), Commission on L Sciences (CLS), the Board o'n Agriculture (BOA), and the Institute of Medicine (IOM) proposes to undertake an assessment of opportunities to improve risk characte the aim of making scientific information on risk more accessible to non speciaHst risk managers and the public. The 7 #

" inth study will be carried out by a multi disciplinary committee of approximately 15 members. The committee will meet several times to a information, examine case studies, conduct a workshop, deliberate on issues, and prep report. The report will have a broad audience including federal, state, and local re Rev. 9/1/91 I

~

m L

~

industry, academia, and the general public. It is expected that the report will include recommendations for improving the process by which scientific information about a risk assessment is presented to expert and lay risk managers and decision makers.

ORIGIN AND BACKGROUND One of the key characteristics of contemporary public policy formulation is that decisions having complex scientific and technical components must frequently be made by non-specialists. Additionally, these decisions and their technical basis must be made cred to a largely lay public.

At the juncture of the scientific and decisional processes is the function " risk characterization." In Risk Assessment in the Federal Government (NRC,1983), risk characterization, the final step in the risk assessment process, is defined as "the process of estimating the incidence of a health effect under the various conditions of human exposure...". A risk characterization "will be used by the regulatory decision-maker when health risks are weighed against other societal costs and benefits to determine an appropriate action." Generally, a risk characterization involves commentary - often qualitative in character -- on the nature and meaning of a risk assessment. It includes qualitative and quantitative technical information on the risk assessment itself as well as information to the assessment but of contextual importance.

Since the publication of Risk Assessment in the Federal Government, a great deal of practical experience with formalized risk decision making has been accumulated. Sig efforts have been made to routinize risk assessments to provide greatcr consistency in risk

/

an.9 sit 9s 2

1 decision making and to improve the credibility of the risk assessment process. In addition, risk assessment concepts have been expanded to consider ecological consequences.. Many experts and lay persons remain skeptical, however, that risk management decisions are adequately grounded in the best available scientific taiderstanding. This skepticism derives, in part, from the process through which scientific information is incorporated into risk management.

During the risk characterization stage of the process, the information in a risk assessment must be translated into a form usable by a risk manager or individual decision maker. Determining what constitutes " usable" information is neither simple nor straightforward. In the relatively new field of ecological risk assessment, defm' ing " usable" information may also be problematic because of the range of relevant stakeholders.

De EPA memorandum Guidance on Risk Charactericationfor Risk Managers and Risk Asussors (Habicht,1992) defines risk characterization as "a combination of qualitative information, quantitative information, and information regarding uncertainties." It describes in detail the crafting of a risk characterization including the types of information to be included in the characterization and suggested descriptors for that information. From the perspective of those charged'with presenting the results of a risk assessment, the EPA memorandum provides much useful information. It is important to remember, however, that this perspective is only one side of the communication process--that of ta e,iver of information, which may differ from that of the intended recipient. People try to fit new information into the context of their existing understanding. If they have misconceptions, they may misconstrue the new information. If they have no knowledge of a topic, new i

I Rav. 9/1!?2 3

g.

i b

i ls To be usable, a risk characterization must encompass both informat on may e mean ng e s.

~

the viewpoints of the risk expert and the 'nonexpert recipient.

One of the principle findings of Unfinished Business: A Comparative Assessment of Environmental Problems (EPA,1987) underscores this point "Whether an environmental

i problem appears large or not depends critically on the type of adverse effect with which one is concerned." How does one balance the need to present cor::4*e information needed by -

decision makers with the need to address concern over various acverse effects? How does -

a one address adverse efixts on ecosystems compared to those on human health? What is the result of not matching information content to audience concerns? What determines the 5

usefulness of information?

Recent and on going research is beginning to provide answers to some of these questions. The results have direct impacts on issues of concern to federal agencies. The-case of indoor radon provides one such example. The health risks associated with indoor

.1 radon appear to be much larger than health risks associated with many other hazardous--

substances over which the public is more concerned. Further, individuals can do something about reducing their exposure to radon. Yet in spite of some general knowledge about radon-hazards, a very small percen'tage of the public even bothers to test their homes for radon,-let alone take action to reduce radon.

Ecological risk assessment relies on appropriate indicators of the environmental effects of ecological stresses. Three facets are central to what can be called " stress' ecology":

1 (1) how various components of ecosystems are exposed to stress, (2) how ecosystems respord to those stresses, and (3) how ecosystems recover and adapt to stress (Harwell et ali,-

1990). Scientists are limited, however, in their ability to predict ecological responses to -

4 Rev. 9/1/92

x.

e

'Ihe reasons are many; There is considerable variety in ecosystems and human stress.

disturbance of those ecosystems. There is a wide range of spatial,- temporal, and structural scales inherent in any ecosystem. There is sparse data on undisturbed ecosystems to use for comparison. Despite these difficulties, methods are being developed to assess ecological stress. Approaches have been developed for addressing spatial aM temporal scales, indicator species, biological monitoring, cumulative effects, and other issues (NRC,1986). The communication task is even more difficult because people may view degradation of natural endowments to be less important than economic or other uses of those resources.

There has been, however, little systematic assessment of how this complex information can be presented most effectively to nonexperts.

Work being done at Carnegie Mellon University (CMU) on? mental models" suggests ~

that providing information that fills in gaps in information and corrects misconceptions may better prepare people to cope with a hazard. Two brochures on radon prepared by CMU researchers were tested in comparison to the EPA's widely distributed Citizens Guide to Radon (U.S. Environmental Protection Agency and U.S. Depretment of Health and Human Services,1986). The CMU brochures were structured froin a decision-analytic perspective.

l Their content was based on filling in gaps and correcting flaws in the public's mental model of radon, which was derived from intensive interviews and survey data. Subjects reading the two CMU brochures performed better on a battery of measures-including a measure designed for testing the EPA brochure specifically-than did readers of the EPA brochure -

(Atman,1990; Bostrom,1990). In particular, readers of CMU brochures were better able to make inferences about information not explicitly mentioned in the brochure. They were able to give much more detailed recommendations about how to deal with a radon problem than t

I 5

an. 9et/92 I-

were EPA brochure readers. In fact 43 percent of EPA brochure readers answered " don't know" when asked what a homeowner could do to reduce high radon levels. Almost all readers of the CMU brochures (96% for one of the brochures,100% for the other) answered -

" Hire a contractor to fix the problem." The results of this study point to a possible mismatch between the information risk experts believe is important in risk characterization and the information the individual decision maker actually needs, ultimately defeating the intent of the communication.

The contentiousness in many situations involving risk today suggests that the radon brochure experience is not an isolated incident. Those characterizing risks and communicating about them all too often make assumptions about what their audience believes, knows, snu needs to know--assumptions that do not stand up to empirical testing.

In an attempt to help chemical plant managers better deal with the growing public distrust of chemicals, the Chemical Manufacturers Association put out a much lauded manual, Risk Communication, Risk Statistics and Risk Comparison (Covello et al.,1988), which included a ranking by acceptability of different types of risk comparison. A study asking lay people'to-rank these same risk comparisons found no correlation between the acceptability ratings of -

the public and those in the manual (Roth et al.,1990), even though the ratings in the manual-were logically derived from risk perception literature.

Similarly, the Food and Safety Inspection Service (FSIS) of the_ U.S. Department of Agriculture undertook a project to improve food safety education directed at food handling by consumers (USDA,1989). Because of the paucity of survey data on food handling in the home, FSIS primarily relied on the views of consumer experts as their primary measure of food safety knowledge and practice of home food handlers. The resultant approach to food.

6 n.v. 9/i/92

d i

' safety education was based primarily on the expert views, even though the study noted that -

the views of the experts were not always consistent with the consumer survey results. While-J this approach is understandable, it raises questions about the effectiveness of the resultant education program.

When problems arise in risk management, experts often decry the public's ignorance.

They seem to feel that if only the public could be " educated", problems wouldn't exist.

They therefore set about to impart knowledge to the public. Research on risk perception has demonstrated that although lay estimates of risk are subject to bias, those biases are not clearly implicated in inappropriate risk decisions. People do have a fairly stable internal risk scale (Fischhoff et al.,1992t Poulton,1989). Factors other than numerical estimates of risk are most likely involved in controversial situations. Sj6 berg and Drottz Sj6 berg (1991) found '

that the dermition a person has of risk play; a large part. Risks seem larger to people who define risk in terms of consequences than to those who define risk in terms of probabilities.

A case in point is Yucca Mountain. Despite much effort to involve the state'and local governments and to share information with the community about the testing of Yucca--

Mountain, Nevada for a nuclear waste storage facility, opposjtion to the project continues

^

unabated. Slovie et al. (1991) used a word association technique to study the images and attitudes associated with an " underground nuclear waste storage facility." Over 3,000 respondents from Nevada and its two major sources of tourists, Phoenix, Arizona and southern California gave 10,000 associations. Thirty-five percent of the responses were negative consequences deriving from such a facility and another 21 percent were other-negative concepts. It would appear that the public's definition of the risk of a nuclear waste li l

7 n.v. 9eit92

!w

l storage facility may be based on consequences rather than probabilities. This may help to explain why the public's view of the risk is so different from the experts' view.

In what other ways do the public's ideas and understanding differ from that of the experts? Kraus et al. (1992) compared the responses from toxicologists to those from the lay public on a survey dealing with chemical risks. Both toxicologists and the public (94.6%

and 93.4%, respectively) agreed that they should know as much as possible about the chemicals around them to better control the risks these chemicals pose to their health and the environment. The public was much less sensitive than the experts to considerations of dose and response, except in the case of prescription drugs. In generai, the public's responses on these sutveys showed very negative attitudes towards chemicals. Even small amounts of chemicals were deemed to cause almost certain harm. Interestingly, the public was more.

likely to accept the extrapolation of animal studies when there was evidence of-carcinogenicity than were toxicologists. This may be a result of the public's strong negative attitudes toward chemicals and their risks. It is well known that people-including experts-trust information that is congruent with their prior expectations. Expecting chemicals to be carcinogenic, they are more willing to believe a finding of carcinogenicity.

- The research seems to suggest that knowing your audience is key to effective risk communication. -However, surveys, individualinterviews, focus groups, and other means of.

testing the knowledge, attitudes, and perceptions of the audience-experts or the public-are time consuming and costly. Additionally, early and sustained involvement of_potentially affected groups has been recommended (National Research Counal,1989). -Involving ;

representatives of one's future audience in the early stages of the risk assessment may be the only way to ensure that their concerns are reflected in the analysis. How does one achieve 8

a.. 9tiiv2

early and sustained involvement? How does one turn the complex results of a risk analysis:

into usable information in a timely and cost-effective manner? The key to further progress ins this area is to improve the understanding of the risk characterization process.

PROPOSED ACTIVITY

,This study will be carried out by an appointed, multi disciplinary committee of approximately 15 experts from a variety of specialties such as risk assessment, epidemiology, toxicology, ecology, public policy, economics, social science, and law. In addition, individuals from the " user community," including regulatory agencies, industry, and environmental and citizens groups, will be sought. Committee members will be chosen so as to assure a flexible view of their charge and to provide an overall balance to the committee.-

They will meet approximately 7 times over a period of 24 months to acquire information, examine case studies, hold a workshop, deliberate on issues, and write their report.

The committee will not perform original research. The body of its report and its conclusions and recommendations will be produced from assessment of the relevant literature, review of several case studies, and the collective expertise ofits members. Case studies will be examined in an attempt to identify what approaches work best to produce effective risk characterizations. The committee will receive technical and administrative support from the staff of CBASSE who will also be responsible for assuring that all Academy and National Research Council policies and procedures are followed.

9 a.v. 9/ii92

Early in the course of the project, the committee will organize a workshop under joint sponsorship with appropriate units of CGER, CLS, BOA, and IOM. Workshop participants will be drawn from a variety of disciplines to highlight a diversity of perspectives on the issues under study and to apprise the committee of the most current research on those issues.

The multidisciplinary nature of the workshop will provide advice to the committee on how it might best approach the topics of concern in the study.

Scope Characterization of risk is commonly done in many diverse areas in which human health, safety, or ecological hazards may be present. The types of riskt vidressed and methods of assessment vary widely among fields such as carthquake engineering, medicine, public health, environmental health, ecologic systems, industrial safety, and consumer safety, _

among others, because of the inherent differences in the nature of the risk as well as the kind of scientific information t.vailable. Ultimately, however, an assessment of risk in all of these.

areas must be summarized or characterized, and that characterization will be used a; part of ~

the basis for making decisicns about how the risk should be managed. Risk characterizations are used by diverse decision-making audiences from professional decision makers responsible for regulatory actions to lay individuals making decisions regarding personal risks. It is expected that the committee will focus on a broad range of risk types and draw from experience gained in~ diverse decision-making settings in order to understand _the process that bridges risk assessment and risk management. It is also expected that the committee with examine the risk characterization process from the perspective of the user of the information 10 an. 9/1/92 -

i

.'as well as the more traditional perspective of the risk analyst.- The analysis should include =

timing of user involvement, as well. Early knowledge about the type of' information needed by the user may have important implications for fae questions asked in.'the risk' analysis, a

Study Approach Tne committee will approach its task by first identifying a range of examples of.

existing risk characterizations. These characterizations may derive from various stages of decision-making processes and are expected to differ considerably in content, tr.chnical detail, and completeness. The committee's analysis will focus on three aspects of risk characterization: the use of descriptors anct metrics to " sum-up" risk' assessments, the use of comparisons as a way of conveying magnitude and other attributes of risks, and the processes -

by which risk assessments are prepared and communicated.- The committee will examine how these concepts may need to be expanded or altered to cover ecological rish ' The committee will address ways in which risk management decisions, particularly in the federal-government, might be improved in light of its findings in these study areas or others

identified during the course of the deliberations.

PRODUCT AND DISSEMINATION The committee will issue two publications in the course of its work. The first will contain the proceedings of a workshop that it will plan and hold during the first phase afits work. The second phase of the committee's work will focus on writing its final report that' 11 an.9eti9s

3=-

k

- will be issued at the end of the 24 month study period. These reports will be subject to review in accordance with the procedures established by the Report Review Committee of the National Research Council. They wi" be made available to the public without restriction and will be prepared in sufficient quantity to ensure distribution to the sponsors, the public, and other interested panies.

REFERENCES Atman, C. J. 1990. Network structures as a foundation for risk communication: An investigation of structure and formal differences. -Ph.D. dissertation, Carnegie Mellon University, Pittsburgh, Penn.

^

Bostrom, A. 1990. A mental models approach to exploring perceptions of hazardous processes. Ph.D. dissertation, Carnegie Mellon University, Pittsburgh, Penn.

- Covello, V. T., P. M. Sandman, and P. Slovic. 1988. Risk Comm:mication, Risk Statistics, and Risk Comparisons: A Manualfor Plant Managers. Washington, D.C.: Chemical Manufacturers Association.

Fischhoff, B., A. Bostrom, and M. J. Quadrel.1992. Risk Perception and Communication.

Draft manuscript. Available from B. Fischhoff, Department of Social and Decision Sciences, Carnegie Mellon University,,Pittsburgh, Penn.

Habicht, F. H. 1992. Guidance on Risk Characterization for Risk Mangers and Risk Assessors. Environmental Protection Agency, Washington, D.C., February 26,1992, memorandum, f

12 Rev. 9/U92

g.....

^

v Harwell, M. A., C. C. Harwell, D. A. Weinstein, and J. R. Kelley. 1990. Characterizing ecosystem response to stress.

Pp.91-115 in Ecolozical Risks: Persocctives from-Poland and the United States, W. Grodzinski, E. B. Cowling, and A. I. Breymeyer, eds. Washington, D.C.: National Academy Press.

Kraus, N., T. Malmfors, and P. Slovic. Intuitive toxicology: Expert and lay judgments of chemical risks. Risk Analysis 12(2):215-232.

Lowrance, W. W.

1990. Regulatory Management of Carcinogenic Chemical Risks:

Informal Summary of a Symposium Convened by an Ad Hoc Coalition at Wrightsville -

Beach, North Carolina on May 2-5,1989. Available from W. W. Lowrance, Life Sciences and Public Policy Program, The Rockefeller University, New York, New York.

National Research Council. 1989. Improving Risk Communicarbn. Washington, D.C.:

National Academy Press.

National Research Council. 1986. Ecological Knowledge and Environmental Problem-Solvingt Concepts and Case Studies. - Washington, D.C.: National Academy Press.

National Restarch Council; 1983. Risk Assessment in the Federal Government: Managing the Process. Washington, D.C.: National Academy Press.

Pt ulton, E. C. 1989. Bias in Quonnfying Judgment. Hillsdale, N.J.: Lawrence Erlbaum.

Resovr:es for the Future. 1990. Confronting Uncertainty in Risk Management: A Guidefor Decision Makers. Center for Risk Management, Resources for the Future, Washir.Jon, D.C.

Roth, E., M. G. Morgan, B. Fischhoff, L. Lave, and A. Bostrom. What do we know about risk comparison? Risk Analysis 10(3): 375-388.

13 a.. 9tii92 r

Sj6 berg, L., and B M. Drottz-Sjaberg.1991. Knowledge and risk perception'among nuclear power plant employees. Risk Analysis 11(4):607 618.

Slovic, P., M.1.ayman, and.J. H. Flynn. 1991. Lessons from Yucca Mountain.

Environment 33(3):7-11, 28 30.

U. S. Department of Department of Agriculture. 1989. A Margin o/Sofery/ - 77w HACCP Approach to Food Safety Education. Washington, D.C.: Food Safety and Inspect Service, U.S. Department of Agriculture.

U. S. Environmental Protection Agency. 1987. Unfinished businesst A comparative assessment of environmental problems. Washington, D. C.: U.S. Environmental Protection Agency.

U.S. Environmental Protection Agency and U.S. Department of Health and Human Services.

1986. A Citizen's Guide to Radon. OPA-86-004. Washington, D. C.: -U.S.

Government Printing Office.

Viscusi, W. K. 1991. Risk perceptions in regulation, tort liability, and the market.

Regulation: Cato Review of Business and Government Fall:50-57.

l l

Rav. 9/1/92

[4

. - -,