UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
          °                       WASHINGTON D.C. 20460

                                                    OFFICE OF THE ADMINISTRATOR
                                                      SCIENCE ADVISORY BOARD


                                 February 28, 2007

EPA-SAB-07-003

Honorable Stephen L. Johnson
Administrator
U.S. Environmental Protection Agency
1200 Pennsylvania Avenue, N.W.
Washington, DC 20460

       Subject:  Consultation on Enhancing Risk Assessment Practices and Updating EPA's
               Exposure Guidelines

Dear Administrator Johnson:

       The EPA Office of Science Advisor requested that the Science Advisory Board
(SAB) conduct a consultation to provide input on ways to advance the Agency's human
health risk assessment practices. Additionally, the Risk Assessment Forum  requested that
the SAB provide advice on updating the 1992 Exposure Guidelines. On September 6 and 7,
2006, representatives of the Office of Research and Development (ORD) and several other
EPA offices provided informative presentations to the members of the SAB  Environmental
Health (EHC) and Integrated Human Exposure (IHEC) Committees and several board
members of the SAB. The focus of the presentations by EPA representatives for this
consultation was on advancements made in human health risk assessment and exposure
assessment.  On behalf of the committee members, we would like to express our sincerest
gratitude to the presenters for their expertise, perspectives and insights. Their contributions
greatly increased our understanding of the Agency's current policies, methods, practices and
future directions.

       The SAB was asked to comment on the relevance and priority of initiatives by the
Agency to advance human health risk assessment practices and the Agency's approach for
updating the  exposure guidelines. They were also asked to suggest other areas and
improvements that should be considered and which would be most important.  Committee
members addressed five major topics that emerged as the focus of both efforts during these
consultations; 1) Addressing Aggregate Exposure and Cumulative Risk Assessment, 2)
Addressing Populations, Groups, or Life Stages of Potential Concern, 3) Evaluating Uncertainty
and Variability, Including Probabilistic Analyses, 4) Involving Communities and
Communicating Results, and 5) Use of Data (Mechanistic, Models, Genomics, CompTox, etc.)
versus defaults.  Feedback on the charge questions was provided by committee members and

-------
summarized according to each of the five topics. A compilation of these comments is appended
to the minutes for this meeting. Highlighted in this letter are several key messages that emerged
among the Committee members as a result of the Agency presentations and discussions.

       The SAB agreed that the Agency has obviously put great effort into advancing human
health risk assessment practices in many areas. The Agency has utilized sound principles and
science, used external peer review, and is developing guidelines that should result in more
transparent assessments. The Agency presented a comprehensive conceptual framework for
human health risk assessment.  Although this framework identified many scientific and
practical needs, it did not provide an assessment of priorities or a plan for meeting those
needs. The SAB is providing a number of overarching recommendations to address both
advancing human health risk assessment and updating the exposure guidelines because the
Agency has focused on many of the same concerns with regard to both efforts. The SAB
recommends that the Agency:

   •   Develop a plan to assess and prioritize the scientific and practical needs for
       improving human health risk assessment.

Topic 1
   •   Advance cumulative risk assessment methodologies, in order to reflect real-world
       human exposure that includes multiple stressors.
   •   Integrate work completed to better characterize cumulative exposure and risk across
       age groups, and among children and the elderly.

Topic 2
   •   Consistently address early life susceptibility in assessments, using weightings for
       children, prenatal exposure,  and lifetime to pregnancy (body burden) exposure.
   •   Include the elderly subpopulation and existing health, medication, and nutrition status
       when conducting risk assessments.
   •   Determine the status of populations in terms of background exposures and disease
       factors.

Topic 3
   •   Characterize variability and  uncertainty more fully, including extending where
       scientifically feasible related quantitative analyses to the dose response and hazard
       identification parts of the Agency's cancer and noncancer risk assessments, and
       thereby identify ways to minimize uncertainty.
   •   More systematically, clarify the underlying assumptions used to build probability
       distributions for the processes and the observations on those processes.
   •   Incrementally replace the current system of single-point uncertainty factors with a set
       of distributions using probabilistic methods. (Some of the potential benefits of
       probabilistic analyses are included in Attachment 1.)
   •   Continue to develop mechanisms to evaluate both exposure and effects predictions of
       current and new human health risk assessment models.
   •   Conduct evaluations to determine whether risk assessment predictions match reality.

-------
Topic 4
   •   Convene workshops at periodic intervals (2-3 years) to gather new information,
       including new data, new techniques and new tools.
   •   Assess, and probably increase, its program of training for both assessors and
       managers to appropriately interpret, communicate, and effectively utilize probabilistic
       information in decision-making.

Topic 5
   •   Whenever possible, use data derived from humans rather than from animals.
   •   Continue to develop greater understanding of modes and mechanisms of action,
       including mechanisms of genotoxicity, to improve understanding of the relevance of
       data from animal models to humans.

       Finally, the SAB recognizes that the design and implementation of new methods will
require specialized expertise and sustained support.  We urge the Agency to provide the
necessary resources and support to ensure that continued improvements are made.  We look
forward to working with the Agency to enhance approaches for exposure and human health
risk assessment.

                                  Sincerely,
      /Signed/                                /Signed/

Rebecca T. Parkin, PhD, MPH             Granger Morgan, PhD
Chair, Integrated Human Exposure         Chair, Science Advisory Board
Committee and Environmental
Health Committee
Enclosure (Attachment 1)

-------
                                      ATTACHMENT 1

Some Potential Benefits of Probabilistic Analyses include the following:

•   In  contrast to the current definition  of the Reference Dose (RfD)1, RfDs designed to meet a
    probabilistic  goal  would allow the technical vs. policy considerations to be made explicit in
    quantitative terms—making clear how much  confidence the analysts should be able to achieve
    that risks are below some specified incidence.

•   Assessment of uncertainties quantitatively could facilitate "value of information" type analyses
    to help set research priorities toward the largest and most easily reducible sources of uncertainty.

•   A probabilistic RfD system could help reduce the potentially inaccurate implication of zero risk
    below the RfD.   The likelihood of finite risks for  some  non-cancer effects at low doses is
    highlighted by the recent example of apparently substantial mortality to vulnerable portions of
    the population from ambient levels of small airborne particles.

•   A probabilistic RfD system would provide a capability to quantify risk below or above the RfD.
    This would allow EPA to quantify benefits of exposure control measures for OMB-mandated
    juxtapositions of economic and health consequences of different policy options.  Without this
    capability, reductions in air toxics and non-cancer effects from other exposures are effectively
    not counted  in  analyses  of  benefits  in  regulatory impact analyses.   This  may  lead  to
    underweighting of efforts to abate such effects in the policy formulation process.

•   A probabilistic RfD would remove the apparent contrast in the best current assessments that are
    highly sophisticated probabilistic exposure assessments joined to simple-appearing single-point
    representations of information from the field of toxicology.

•   A  probabilistic RfD system would encourage the generation of better information because it
    would create  a clear regulatory market for such a system. As pointed out in our discussions, this
    would  improve  on the World  Health  Organization International  Programme on  Chemical
    Safety's  (WHO IPCS) data  derived  uncertainty factor procedures,  that are  not rigorously
    founded in terms  of allocation of variances  between pharmacokinetic and pharmacodynamic
    components,   or  over-constrained by  the   requirement that  default kinetic   and  dynamic
    components must multiply to the traditional factor of 10.

•   An innovative probabilistic system is more likely to  attract  the efforts of innovative researchers
    interested in  producing improved technical  information and seeing policy  responses to that
    information. Currently researchers in this area have a difficult struggle to achieve acceptance in
    place of the heritage of prior "case law" choices made from the 1954 Lehman and Fitzhugh "100
    fold safety factor" paper to the present.
  1 Reference Dose (RfD): An estimate (with uncertainty spanning perhaps an order of magnitude) of a daily oral exposure to the human
  population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime. It can be
  derived from a NOAEL, LOAEL, or benchmark dose, with uncertainty factors generally applied to reflect limitations of the data used.
  Generally used in EPA's noncancer health assessments, as defined in the glossary attached to EPA's IRIS database
  (http://www.epa.gov/iris/gloss8.htm. accessed 1/23/07)

-------
•   Examples of the richness of information that could be made  available to decision-makers by
    implementation of probabilistic methods substituting distributions based on empirical  data based
    on observations of other chemicals/drugs instead of the single-point uncertainty factors that are
    traditionally used) are illustrated in Hattis and Lynch (2007).2
  2 Hattis, D. and Lynch, M. K. "Empirically Observed Distributions of Pharmacokinetic and Pharmacodynamic
  Variability in Humans—Implications for the Derivation of Single Point Component Uncertainty Factors
  Providing Equivalent Protection as Existing RfDs." In Toxicokinetics in Risk Assessment. J. C. Lipscomb and
  E. V. Ohanian, eds., Informa Healthcare USA, Inc., 2007, pp. 69-93.

-------