[Senate Hearing 109-1048]
[From the U.S. Government Publishing Office]
S. Hrg. 109-1048
THE SCIENCE AND RISK ASSESSMENT BEHIND EPA'S PROPOSED REVISIONS TO THE
PARTICULATE MATTER AIR QUALITY STANDARDS
=======================================================================
HEARING
BEFORE THE
COMMITTEE ON
ENVIRONMENT AND PUBLIC WORKS
UNITED STATES SENATE
ONE HUNDRED NINTH CONGRESS
SECOND SESSION
__________
July 19, 2006
__________
Printed for the use of the Committee on Environment and Public Works
Available via the World Wide Web: http://access.gpo.gov/congress.senate
__________
U.S. GOVERNMENT PRINTING OFFICE
47-635 PDF WASHINGTON : 2009
For sale by the Superintendent of Documents, U.S. Government Printing
Office Internet: bookstore.gpo.gov Phone: toll free (866) 512-1800;
DC area (202) 512-1800 Fax: (202) 512-2104 Mail: Stop IDCC,
Washington, DC 20402-0001
COMMITTEE ON ENVIRONMENT AND PUBLIC WORKS
ONE HUNDRED NINTH CONGRESS
SECOND SESSION
JAMES M. INHOFE, Oklahoma, Chairman
JOHN W. WARNER, Virginia JAMES M. JEFFORDS, Vermont
CHRISTOPHER S. BOND, Missouri MAX BAUCUS, Montana
GEORGE V. VOINOVICH, Ohio JOSEPH I. LIEBERMAN, Connecticut
LINCOLN CHAFEE, Rhode Island BARBARA BOXER, California
LISA MURKOWSKI, Alaska THOMAS R. CARPER, Delaware
JOHN THUNE, South Dakota HILLARY RODHAM CLINTON, New York
JIM DeMINT, South Carolina FRANK R. LAUTENBERG, New Jersey
JOHNNY ISAKSON, Georgia BARACK OBAMA, Illinois
DAVID VITTER, Louisiana
Andrew Wheeler, Majority Staff Director
Ken Connolly, Minority Staff Director
(ii)
C O N T E N T S
----------
Page
July 19, 2006
OPENING STATEMENTS
Boxer, Hon. Barbara, U.S. Senator from the State of California... 6
Carper, Hon. Thomas R., U.S. Senator from the State of Delaware.. 10
Clinton, Hon. Hillary Rodham, U.S. Senator from the State of New
York........................................................... 9
DeMint, Hon. Jim, U.S. Senator from the State of South Carolina.. 7
Inhofe, Hon. James M., U.S. Senator from the State of Oklahoma... 1
Jeffords, Hon. James M., U.S. Senator from the State of Vermont.. 3
Lautenberg, Hon. Frank R., U.S. Senator from the State of New
Jersey......................................................... 8
Voinovich, Hon. George V., U.S. Senator from the State of Ohio... 4
WITNESSES
Lieberman, Hon. Joseph, U.S. Senator from the State of
Connecticut, prepared statement................................ 47
Gray, George, Assistant Administrator for Research and
Development, U.S. Environmental Protection Agency.............. 11
Prepared statement........................................... 47
Responses to additional questions from:
Senator Inhofe........................................... 50
Senator Voinovich........................................ 51
Greenbaum, Daniel, President, Health Effects Institute........... 33
Prepared statement........................................... 54
Responses to additional questions from Senator Voinovich..... 58
McClellan, Roger, Advisor, Toxicology and Human Health Risk
Analysis....................................................... 28
Prepared statement........................................... 60
Responses to additional questions from Senator Voinovich..... 66
Smith, Anne, Vice President, CRA International................... 32
Prepared statement........................................... 68
Responses to additional questions from Senator Voinovich..... 81
Stephenson, John, Director, Natural Resources and Environment,
U.S. General Accountability Office............................. 13
Prepared statement........................................... 86
Responses to additional questions from Senator Voinovich..... 90
Thurston, George, Associate Professor, New York University,
School of Medicine, Department of Environmental Medicine....... 30
Prepared statement........................................... 92
Responses to additional questions from Senator Voinovich..... 95
ADDITIONAL MATERIAL
Letters:
Children's Health Protection Advisory Committee.............. 163
Undersigned by Bart Ostro and Michael Lipsett (regarding
support for revising particulate matter standards)......... 188
Undersigned by National, Regional, State, and Local
Organizations
(regarding supprt for revising particulate matter
standards)................................................. 193
Undersigned by Senators (regarding funding for the State and
Local Air Quality Management Program)...................... 202
Statements:
CRA International............................................96-162
Fran Pavley, Assembly California Legislature................191-192
Jonathan Borak & Company Inc................................171-187
THE SCIENCE AND RISK ASSESSMENT BEHIND EPA'S PROPOSED REVISIONS TO THE
PARTICULATE MATTER AIR QUALITY STANDARDS
----------
Wednesday, July 19, 2006
U.S. Senate,
Committee on Environment and Public Works,
Washington, DC.
The committee met, pursuant to notice, at 9 o'clock a.m. in
room 628, Dirksen Senate Office Building, Hon. James Inhofe
(chairman of the committee) presiding.
Present: Senators Inhofe, Voinovich, Chafee, DeMint,
Isakson, Jeffords, Boxer, Carper, Clinton, and Lautenberg.
Senator Inhofe. Consistent with our policy of starting on
time, we will call this meeting to order.
OPENING STATEMENT OF HON. JAMES M. INHOFE, U.S. SENATOR FROM
THE STATE OF OKLAHOMA
I would like to tell our members, since we have quite a
number of people here and of course we have the WRDA bill on
the floor, which I have to manage so I am not going to be able
to stay very long, we are going to try to keep everyone to
about 3 minutes. I think we put a notice out if we could
encourage both sides to do that, and then anyone who arrives
after the conclusion of our opening statements would have to
just submit the statements for the record, if that is
acceptable.
I am managing the WRDA bill, along with my colleague
Senator Jeffords, on the floor so I am going to have to go down
and prepare for that and won't be able to stay here very long.
I am going to ask that Senator Voinovich, who is kind of the
master of all air issues anyway, would handle chairing this
committee.
Last week, the Air Subcommittee examined the impacts of
tightening particulate matter standards on our Nation. Although
EPA failed to analyze the regulatory impact on the Nation,
these impacts will be enormous. I was particularly struck by
the testimony of Harry Alford, president of the National Black
Chamber of Commerce, who testified that the greatest health
threat to minorities is access to health care and a tightened
standard which would threaten the paychecks that cure that
threat.
Today, we are examining the science underlying the
particulate matter review. The estimated risk today is less
than it was estimated in 1997 under Carol Browner when the
current standard was set. So while I feel EPA's proposal to
tighten the daily standard to 35 micrograms is overly
stringent, I am pleased EPA proposed to retain the existing
annual standard.
The rationale to tighten this standard is weak. The EPA
cherry-picked what studies it was relying on, downplaying many
key studies that shed light on the health effect of particulate
matter, some of which are listed on this chart which is
difficult even for me to read this close. But anyway, these are
the studies.
It also cherry-picked what information it provided to the
Clean Air Science Advisory Committee, an important document
that seriously skewed the review. A lot of times, CASAC is not
given their assigned task. CASAC, the Clean Air Scientific
Advisory Committee, is one of a set of some 21 scientists that
are supposed to be giving advice and we are supposed to be
listening to that advice. They were never told by the EPA that
the estimated risk from PM exposure is now considered lower
than the risk level estimated during the last review.
EPA's process for this review is also radically different
from every previous scientific assessment, calling into
question the credibility of the entire review. CASAC is
supposed to review relevant science and the public is supposed
to provide input. This time, EPA had a cutoff date of April,
2002 which meant that CASAC's assessment does not include
almost 4 1/2 years of new studies. Only after I asked EPA to
collect the newer studies did it do so. Now, it plans to issue
a final rule without an opportunity for public review of how it
is assessing those studies.
Worse, the General Accounting Office report being released
today shows EPA has failed to follow the National Academy of
Sciences recommendations to examine the health effect
associated with different kinds of particulates and to
incorporate the range of particulate toxicity assumptions into
its uncertainty analysis. If we don't know the types of
particles causing the health effects, we can't really
accurately assess those.
The system that is in place today only considers the size
of the particulate matter. I think we all know that the health
effects of a molecule of asbestos is much more dangerous than
would be a molecule of dust. These things should be considered.
Senator Jeffords.
[The prepared statement of Senator Inhofe follows:]
Statement of Hon. James M. Inhofe, U.S. Senator from
the State of Oklahoma
Last week, the Air Subcommittee examined the impacts of tightening
particulate matter standards on our nation. Although EPA failed to
analyze the regulatory impact on the Nation, these impacts will be
enormous. I was particularly struck by the testimony of Harry Alford,
President of the National Black Chamber of Commerce, who testified that
the greatest health threat to minorities is access to health care and a
tightened standard would threaten the paychecks that cure that threat.
Today, we are examining the science underlying the particulate
matter review. The estimated risk today is less than what was estimated
in 1997 under Carol Browner when the current standard was set. So while
I feel EPA's proposal to tighten the daily standard to 35 micrograms is
overly stringent, I am pleased EPA proposed to retain the existing
annual standard.
The rationale to tighten the standard is weak. EPA cherry-picked
what studies it relied on, downplaying many key studies that shed light
on the health effect of PM, some of which are listed on this chart.
It also cherry-picked what information it provided to the Clean Air
Science Advisory Committee in important documents, seriously skewing
the review. For instance, CASAC was never told by EPA that the
estimated risks from PM exposure is now considered lower than the risk
level estimated during the last review.
EPA's process for this review is also radically different from
every previous scientific assessment, calling into question the
credibility of the entire review. CASAC is supposed to review relevant
science and the public is supposed to provide input. This time, EPA had
a cut-off date of April 2002, which meant CASAC's assessment doesn't
include almost 4 1/2 years of new studies. Only after I asked EPA to
collect the newer studies did it do so, and now it plans to issue the
final rule without an opportunity for public review of how it is
assessing those studies.
Worse, as the General Accountability Office report being released
today shows, EPA has failed to follow the National Academy of Science's
recommendations to examine the health effect associated with different
kinds of particles and to incorporate a range of particle toxicity
assumptions into its uncertainty analyses. If we don't know what types
of particle cause health effects, we may well spend billions of dollars
on pollution controls while doing little to improve health.
Thank you.
Senator Jeffords. Mr. Chairman, thank you.
Senator Inhofe. Let me just repeat what we said earlier to
our new arrivals here. We are going to try to adhere to 3
minutes in opening statements, and then cutoff statements after
those who are present have concluded.
Senator Jeffords.
OPENING STATEMENT OF HON. JAMES M. JEFFORDS, U.S. SENATOR FROM
THE STATE OF VERMONT
Senator Jeffords. Mr. Chairman, thank you for having this
hearing on the EPA's proposed standards for particulate matter.
There is perhaps no more important environmental standard than
the national ambient air quality standard for particulate
matter. Tens of thousands of Americans die prematurely each
year from particulate matter.
The NAAQS are the cornerstone of the entire Clean Air Act.
Their fundamental purpose is to tell us when the air is safe to
breathe. For more than 30 years, these standards have been set
solely on health considerations, using the latest scientific
evidence. Congress specifically chose not to allow
consideration of cost when setting these standards.
Those who call for a cost-benefit analysis would have us
set a standard that fails to protect the health of all
Americans. We cannot afford the human costs of such an
approach.
Consider this, if the cost-benefit analysis was used to set
the NAAQS and the compliance costs of the standards are
estimated at $1 billion, the EPA would be required to set a
standard that would allow up to $1 billion worth of people to
die from air pollution. So just how do we measure $1 billion in
human life. Setting the NAAQS in that way was unconscionable.
Telling people their air is safe to breathe when it is actually
not safe is unacceptable public policy.
We need to revise the existing standards set in 1997.
Although the EPA itself has recognized this fact, EPA's
proposal falls short of what the scientific evidence requires.
The Clean Air Scientific Advisory Committee recommended the EPA
revise both the annual standard and the daily standard, but the
EPA disregarded this advice, forcing the Scientific Advisory
Committee to reconvene and reiterate its advice to the EPA.
The committee has made it clear that the EPA's proposal for
the coarse particle standard was outside the bounds of
scientific evidence. If the EPA were to listen to the advice of
its science advisors, tens of thousands of lives could be
saved. The standard proposed by the EPA would result in a 22
percent reduction in deaths. Standards set within the range
recommended by the Scientific Advisory Committee would save up
to 48 percent more lives.
Particulate matter kills more people than HIV/AIDS and more
people than drunk driving. It is a big killer and we need to
reduce it as soon as possible. That is what the science shows.
Thank you. I look forward to hearing from the witnesses.
[The prepared statement of Senator Jeffords follows:]
Statement of Hon. James M. Jeffords, U.S. Senator from
the State of Vermont
Mr. Chairman, thank you for having this hearing on the EPA's
proposed standards for Particulate Matter.
There is perhaps no more important environmental standard than the
National Ambient Air Quality Standard for Particulate Matter. Tens of
thousands of Americans die prematurely each year from particulate
matter.
The NAAAQS are the cornerstone of the entire Clean Air Act. Their
fundamental purpose is to tell us when the air is safe to breathe. For
more than 30 years, these standards have been set based solely on
health considerations, using the latest scientific evidence. Congress
specifically chose to not allow consideration of costs when setting
these standards.
Those who call for a cost-benefit analysis would have us set a
standard that fails to protect the health of all Americans. We cannot
afford the human cost of such an approach.
Consider this: If cost benefit analysis is used to set the NAAAQS
and the compliance costs of a standard are estimated at $1 billion, the
EPA would be required to set a standard that would allow up to $1
billion worth of people to die from air pollution. So just how do we
measure $1 billion in human life? Setting the NAAAQS in that way is
unconscionable. Telling people their air is safe to breathe, when it is
actually not safe, is unacceptable public policy.
We need to revise the existing standard set in 1997. Although the
EPA itself has recognized that fact, the EPA's proposal falls far short
of what the scientific evidence requires. The Clean Air Scientific
Advisory Committee recommended that the EPA revise both the annual
standard and the daily standard. But the EPA disregarded that advice,
forcing the scientific advisory committee to reconvene and reiterate
its advice to the EPA.
The Committee also made it clear that the EPA's proposal for the
coarse particle standard was outside the bounds of the scientific
evidence. If the EPA were to listen to the advice of its science
advisors, tens of thousands of lives could be saved.
The standard proposed by EPA would result in a 22 percent reduction
in deaths, but a standard set within the range recommended by the
scientific advisory committee could save up to 48 percent more lives.
Particulate matter kills more people than HIV/AIDS and more people than
drunk driving. It is a big killer and we need to reduce it as soon as
possible. That is what the science shows.
Thank you, and I look forward to hearing from our witnesses.
Senator Inhofe. Thank you, Senator Jeffords.
Senator Voinovich.
OPENING STATEMENT OF HON. GEORGE V. VOINOVICH, U.S. SENATOR
FROM THE STATE OF OHIO
Senator Voinovich. Thank you, Mr. Chairman. I appreciate
the fact that you bumped this hearing from my subcommittee to
the full committee. It shows how important this matter is. To
sum up last week's hearing, we do not know the impact of
revising the particulate matter standards, and let's review
what we do know.
We know that the air is significantly cleaner; that the EPA
has proposed to move the goal posts in the States as they
implement the current standards. We believe that the Agency is
underestimating the number of non-attainment counties. We know
that the current non-attainment designations threaten highway
funding and jobs and increase energy prices and that revised
standards would exacerbate the situation.
We also know this decision has a profound impact on the
health and well being of many, as the National Black Chamber of
Commerce President Harry Alford stated, ``The biggest health
risk to African Americans anywhere is poverty.'' I will never
forget when Tom Mullen from the Catholic Charities came here
several years ago and testified that in setting ambient air
standards and environmental policy, that we should take into
consideration its impact on the poor and the elderly, and the
Clean Air Trust named him villain of the month because he had
the audacity to raise the issue.
Today's hearing focuses on the science and risk assessment
behind the standards. I make three points. First, according to
the recent proposal and risk assessment, we will hear today the
EPA's estimates, the risk from exposure to fine particulate
matter has declined from the level of risk estimated in setting
the last standards.
Second, the health benefits of EPA's proposal are not fully
understood. A 2002 National Academy of Sciences report made 34
recommendations to improve our understanding of the estimation
of the public health benefits. According to GAO that we asked
for, EPA has fully implemented less than 25 percent of
recommendations made back in 2002 by the National Academy of
Sciences. For example, the NAS, along with the Office of
Management and Budget and EPA's Inspector General, has urged
the Agency not to assume that all fine particulate constituents
have the same potency. As stated in the IG's report, otherwise
some facilities may install unneeded controls, while some
needed controls may go uninstalled. Ultimately, compliance may
be further delayed and more costly.
Unfortunately, the EPA has largely not implemented the
recommendations. The standards should not be revised until we
have adequate information that tells us with greater certainty
the health benefits and whether we are targeting the most
harmful constituents of particulate matter, especially
considering the negative impacts on this country's economy. Air
quality will continue to improve through implementation of the
current standards and the Federal clean air rules.
I know you want me to cut this short. I will just basically
say, Mr. Chairman, it is important that everyone understand
that science can only take us so far, but ultimately it is a
policy decision whether or not to change a standard. That is
why the Clean Air Act states the air quality standards are to
be set in the judgment of the Administrator. Given this
judgment discretion, how can EPA revise the particulate matter
standards when the public health benefit is not fully
understood?
Thank you, Mr. Chairman.
[The prepared statement of Senator Voinovich follows:]
Statement of Hon. George V. Voinovich, U.S. Senator from
the State of Ohio
Mr. Chairman, after our hearing in the subcommittee, I appreciate
this important issue being elevated to the full committee.
To sum up last week's hearing, we do not know the impact of
revising the particulate matter standards. Let's review what we do
know:
We know that the air is significantly cleaner and that EPA has
proposed to move the goal posts on States as they implement the current
standards.
We believe that the Agency is underestimating the number of
nonattainment counties and that Federal clean air rules will only
mitigate this problem.
We know that the current nonattainment designations threaten
highway funding and jobs and increase energy prices and that revised
standards would exacerbate the situation.
We also know that this decision has a profound impact on the
health and well-being of many, as National Black Chamber of Commerce
President Harry Alford stated: ``the biggest health risk to African
Americans anywhere is poverty.''
Today's hearing focuses on the science and risk assessment behind
the proposed standards. I will make three points.
First, according to the recent proposal and risk assessment, we
will hear today that EPA's estimate of the risk from exposure to fine
particulate matter has declined from the level of risk estimated in
setting the 1997 standards.
Second, the health benefits of EPA's proposal are not fully
understood. A 2002 National Academy of Sciences (NAS) report made 34
recommendations to improve our understanding of the estimation of the
public health benefits. According to the Government Accountability
Office report that Chairman Inhofe and I requested, EPA has fully
implemented less than 25 percent of them.
For example, the NAS along with the Office of Management and Budget
and EPA's Inspector General has urged the Agency not to assume that all
fine particle constituents have the same potency. As stated by the IG:
``Otherwise, some facilities may install unneeded controls, while some
needed controls may go uninstalled; ultimately, compliance may be
further delayed and more costly.''
Unfortunately, EPA has largely not implemented the recommendations.
The standards should not be revised until we have adequate information
that tells us with greater certainty the health benefits and whether we
are targeting the most harmful constituents of particulate matter--
especially considering the negative impacts on our economy. Air quality
will continue to improve through implementation of the current
standards and the Federal clean air rules.
EPA claims that they will address more of the NAS recommendations
when the final rule is issued. At this time, we are told that there
will also be a more complete Regulatory Impact Analysis and new science
will be considered. It is unacceptable for the public and this
committee to get critical information only after the final decision has
already been made.
Third, testimony that we will hear today indicates that EPA
selectively used study results in developing the proposal.
In conclusion, it is important that everyone understand that
science can only take us so far, but ultimately, it is a policy
decision whether or not to change a standard. That is why the Clean Air
Act states that air quality standards are to be set ``in the judgment
of the Administrator.'' Given this judgment discretion, how can EPA
revise the particulate matter standards when the public health benefit
is not fully understood?
Mr. Chairman, I again thank you for holding this hearing and look
forward to hearing from the witnesses. Thank you. _
Senator Inhofe. Thank you, Senator Voinovich.
Senator Boxer.
OPENING STATEMENT OF HON. BARBARA BOXER, U.S. SENATOR FROM THE
STATE OF CALIFORNIA
Senator Boxer. Thank you, Mr. Chairman, for holding this
hearing. I ask that my full statement be placed in the record.
Senator Inhofe. Without objection, so ordered.
Senator Boxer. I will read part of it.
This hearing is important because protecting the air that
Americans breathe is important. It is a critical responsibility
of this committee. We have made great strides in this country,
but unfortunately EPA's proposed standards for controlling
toxic soot and dust drastically depart from this path.
Now, I won't get into politicizing science, but what I want
to say here is EPA has rejected advice from its own Clean Air
Science Advisory Committee on health-based standards for toxic
soot and dust. This, despite the fact that the American Lung
Association and other public health groups say the children,
the elderly and people with heart disease, diabetes or
respiratory diseases are especially vulnerable to the adverse
impacts of such pollution.
Let me be more specific. Particulate pollution causes
premature death, whether you are in Oklahoma, New York, New
Jersey, Vermont, California or Ohio. Particulate pollution
causes premature death. It exacerbates asthma, cardiovascular
disease, including heart attacks and strokes, and it increases
hospital admissions.
I have to say, sometimes in this committee we talk about
things as if they were very interesting scientific issues. But
the bottom line is what we are talking about here is the
longevity or lack of such of the American people and the
quality of their lives, be they children or adults. More than
160 national, State, local, environmental, religious groups,
and I want to point out religious groups are getting involved
in this, and public interest groups have urged EPA to increase
these clean air protections. EPA's Children's Health Advisory
Committee also urged EPA to revise these standards to protect
children.
I would ask unanimous consent that these letters be placed
in the record at this time. Mr. Chairman, I would like to place
these letters from the religious groups and others into the
record at this time.
Senator Inhofe. Without objection.
[The referenced letters can be found on pages 193-201.]
Senator Boxer. EPA has also proposed a protective standard
for coarse particulate matter. The Agency proposal ignores
pollution by exempting some industries, exempting them,
industries such as mining. In addition, EPA proposes such a
limited amount of monitoring that pollution will go undetected
in many of our medium size and small cities.
So EPA must stop this politicizing. They are politicizing
scientific decisions, and I resent it as a U.S. Senator from
the largest State in the union, where we are deeply affected by
this. A particularly egregious example was identified by Mr.
Bart Ostro, chief of California's Office of Environmental
Health, regarding the particulate pollution rule. On February
3, 2006, Mr. Ostro testified before EPA's Science Committee
that ``last minute additions of edits and opinions by OMB and
others circumvented the entire peer review process.'' That is a
condemning statement.
The Clean Air Act has it right. Sound science and the
protection of public health should guide the establishment of
Clean Air Act standards and sadly, Mr. Chairman, I am very
worried about the fact that we seem to be veering away from
science and making politics the key ingredient in these
decisions.
Thank you.
Senator Inhofe. Thank you, Senator Boxer.
Senator DeMint.
OPENING STATEMENT OF HON. JIM DEMINT, U.S. SENATOR FROM THE
STATE OF SOUTH CAROLINA
Senator DeMint. Thank you, Mr. Chairman. I am very
interested in the results today. I appreciate our witnesses.
Last week, we looked at the possibility of a rule change.
As we all know, the Clean Air Science Advisory Committee has
recommended that EPA lower its annual standard for particulate
matter. What I find interesting is that the air today is
significantly cleaner than it was years ago, and even more
intriguing is the risk assessment of health complications due
to long-term exposure, PM2.5 ambient air, have been
reduced.
Studies have shown that the risk assessment for
PM2.5 have either stayed the same or been reduced.
Then why is the EPA receiving a recommendation from the
committee to tighten the standard? Obviously, I am very
interested in clean air. I just want to make sure that we find
the source and really do something that cleans the air, not
just do something to do something.
The mechanisms that were used to direct the committee to
recommend more stringent annual standards are actually
disturbing. The EPA staff memo directed the committee to study
three specific scientific studies. One of the studies gave
mixed results of the human risk of high levels of
PM2.5; another was nullified by the scientists who
performed the study because the data constantly changed based
on their assumptions; and the third one was based on cities
outside the United States over a 6-year process that showed a
drop in ambient air quality.
So I am very interested in the testimony today. We want to
know about good science. We want to know how to clean the air.
I am just afraid we are looking at doing something just to look
like we are doing something. We will end up costing this
country a lot of jobs and prosperity, and not leave the air any
cleaner.
So I am here to learn, and I thank you, Mr. Chairman.
Senator Inhofe. Thank you, Senator DeMint.
Senator Lautenberg.
OPENING STATEMENT OF HON. FRANK R. LAUTENBERG, U.S. SENATOR
FROM THE STATE OF NEW JERSEY
Senator Lautenberg. Thanks, Mr. Chairman, for giving us
another opportunity to examine the need to protect our air.
This week gives us an urgent reminder about what we have to
do. The red air days are oppressive. I felt it myself, not
enough to make me ill, but enough to make me realize that that
air wasn't particularly clean.
Parents are warned not to let children play outdoors if
they have respiratory problems because of increased danger of
asthma attack or other complications. We know that particulate
matter pollution triggers asthma attacks and affects the health
of people with other respiratory and lung conditions.
Today, we are going to hear how it also poses threats to
people with heart conditions. But we will hear a lot of talk
about uncertainty. We are not sure. It is like saying, well,
the fire is only in the basement; let's not get excited
upstairs.
I heard the same thing 20 years ago when I wrote the law to
ban smoking on airplanes. The tobacco industry said there was
too much uncertainty about the effects of secondhand smoke.
There is no doubt anymore about it. Today, we know that there
is real danger from secondhand smoke and we are so grateful
that we did something when we did it.
We have heard the same excuse about uncertainty for years
in the discussion of global warming. Look outside; feel it. The
claim of uncertainty is the smokescreen that diverts attention
from real problems. It is merely an excuse to avoid taking any
action that might cost an extra few bucks to accommodate our
need.
We are here to represent the 25 million American adults
with heart disease and 6 million children with asthma, one of
whom is my grandson, Alexander. There is no uncertainty about
whether they deserve clean, healthy air. This week, we learned
of a report on particulate matter which had been requested by
OMB. That report suggests that particulate matter is actually a
worse problem than we previously thought. So much for
uncertainty.
So we ought not to let a handful of particular interests
pressure us into doing anything that will permit more harm to
millions of Americans. I think that this is a matter of urgency
and I think we have to get on to making sure that the standards
that we have are the highest and are the toughest than we get
to meet.
And not without consideration for either jobs or
investment. But I will tell you, when we talk about jobs and I
read the reports about what is happening, where things are
being made today, they are not being made in India and other
places, or Bangladesh because of environmental issues. They are
being taken away because of cost issues. So we shouldn't
confuse the facts.
I thank you, Mr. Chairman, once again, for the opportunity
to air this problem.
Senator Inhofe. OK. Senator Clinton.
OPENING STATEMENT OF HON. HILLARY RODHAM CLINTON, U.S. SENATOR
FROM THE STATE OF NEW YORK
Senator Clinton. Thank you, Mr. Chairman.
Although we are making opening remarks, I want to begin by
welcoming someone whom I have a great deal of regard for, Dr.
George Thurston, to the committee. Dr. Thurston is on the
faculty at New York University School of Medicine, where he
conducts investigations into the human health effects of air
pollution. He is truly one of the leading scientists in this
field. I am pleased that he will be testifying on the second
panel.
I have particular regard for Dr. Thurston because when I
was raising questions about the quality of air after 9/11 and
the Administration was assuring everyone that there were no
problems in the air, and in fact everything was fine, Dr.
Thurston was one of those who said this is just not squaring
with the scientific work I have done.
Well, now nearly 5 years later, we have hundreds and
hundreds of firefighters, police officers, construction
workers, residents and others who are suffering from
respiratory distress. Now, that was an intense experience
because the air was so contaminated with everything that came
from the World Trade Center. We can look at that and we can
draw some conclusions about the impact of such a horrible
event.
But it is also true that after 9/11 when all the planes
were grounded, there are a lot of pictures showing how clean
the air was. We are facing a very serious question here. This
is an important issue, not just in New York, but across our
country. Since the 1997 revisions to the particulate matter
standards, numerous scientific studies have been published.
These have not been partisan studies. These have not been
ideological studies. They have ranged from scholarly academic
studies like Dr. Thurston, to the American Lung Association.
These studies have overwhelmingly strengthened the links
between particulate pollution and a range of adverse health
outcomes, including asthma, heart attacks, hospital admissions,
adverse birth outcomes, and premature death. Some recent
studies have even linked long-term particulate matter exposure
to increased risk of lung cancer.
In addition, our understanding of the biological mechanisms
at work helps explain these links better than we understood
back in 1997. So I was dismayed when the EPA disregarded the
advice of its Clean Air Science Advisory Committee.
You know, when CASAC recommended what it thought needed to
be done, the EPA decided not to lower the annual PM standard in
its proposed revisions. This action prompted a letter back to
the EPA from CASAC, which I think is unprecedented. That letter
reiterated the CASAC's finding. All of these scientists who
have no particular ax to grind, ``the epidemiological evidence
supported by emerging mechanistic understanding indicates
adverse PM2.5 health effects below 15 micrograms per
cubic meter.''
Now, I think the science is clear. One of the reasons the
air is cleaner today is that for more than 30 years we have
been cleaning it up. So we have a little bit in the bank, but
that is not going to stand.
So Mr. Chairman, I hope that we will go with the lower
standard, and I ask unanimous consent to submit the full
statement to the record.
Senator Inhofe. Without objection, so ordered.
Senator Inhofe. Senator Carper.
OPENING STATEMENT OF HON. THOMAS R. CARPER, U.S. SENATOR FROM
THE STATE OF DELAWARE
Senator Carper. Thanks, Mr. Chairman.
I want to follow up on what Senator Clinton has said, if I
may, and also just kind of respond to what Senator DeMint has
said. EPA doesn't come up with these recommendations out of
thin air. They have this CASAC committee that Senator Clinton
just referred to. They made their recommendations for advising
the annual and the daily limits on this particulate matter.
Good people, they are all scientists. They have no ax to
grind. What they try to do is to use good science and to give
us their best recommendations, and also to give those
recommendations to EPA. EPA has reviewed, I am told, over 2,000
scientific studies I believe since 1997, and they found an
association between particulates in ambient air and all kinds
of health problems. We have heard a couple of them, a
recitation of them from Senator Clinton. I want to mention them
again: aggravated asthma, chronic bronchitis, reduced lung
function, irregular heartbeat, heart attacks, and premature
death in people with heart or lung diseases.
In just nine cities that EPA looked at, particulates would
cause an estimated almost 5,000 premature deaths unless current
standards are strengthened. EPA could have come in and frankly
followed the advice of CASAC. They didn't. They erred the other
way. For some of us on this committee, we were disappointed
with that result.
I do want to commend Administrator Johnson for
acknowledging that the science does support strengthening the
current particulate matter standards. I am concerned that the
proposed standard is higher than the range recommended by EPA's
CASAC. I look forward to hearing today EPA's justification for
that decision.
Finally, I would like to follow on a point made during last
week's hearing by a number of Senators, as well as Bill Wehrum
of EPA. To do so, let me just quote Mr. Wehrum's testimony from
last week. Here is what he said: ``between 1970 and 2005, gross
domestic product increased by 195 percent. Vehicle miles
traveled increased 178 percent. Energy consumption increased 48
percent. The U.S. population grew by 42 percent. During the
same time period, total emissions of the six principle air
pollutants dropped by 53 percent.''
In other words, since the inception of the Clean Air Act,
our country has made significant increases in our energy usage,
mobility and an ever-growing economy, and at the same time we
have seen our air pollution decrease. Our Nation's air has not
become cleaner on its own. The improvements are a direct result
of specific air regulations and the implementation of the
National Ambient Air Quality Standards. The arguments we are
hearing today that claim that tighter air regs will hurt our
economy and cost people jobs are not new. We have heard these
claims every time EPA has proposed a new regulation over the
past 30 years.
However, as Mr. Wehrum and others on the committee
correctly pointed out last week, those claims have not proven
to be true. We can have cleaner air. We can grow our economy
and we can do those things today, as we have in the last 30
years.
Thank you, Mr. Chairman.
Senator Inhofe. Thank you, Senator Carper.
That will conclude our opening statements. We have on our
first panel Dr. George Gray, Assistant Administrator for
Research and Development for the EPA; and John Stephenson,
Director, Natural Resources and Environment for the United
States GAO.
We will start with you, and try to confine your remarks to
5 minutes, if you could. Dr. Gray, you are recognized.
STATEMENT OF GEORGE GRAY, ASSISTANT ADMINISTRATOR FOR RESEARCH
AND DEVELOPMENT, U.S. ENVIRONMENTAL PROTECTION AGENCY
Mr. Gray. Thank you, Mr. Chairman, members of the
committee. I am George Gray, the Assistant Administrator for
the Office of Research and Development at the Environmental
Protection Agency.
I want to thank you for the opportunity to discuss EPA's
evaluation of the scientific evidence for potential health
effects of airborne particulate matter known as PM. Last week,
the subcommittee heard from my colleague Bill Wehrum, the
Acting Assistant Administrator for Air and Radiation and the
EPA's review of the National Ambient Air Quality Standards, we
call them the NAAQS.
During the testimony, he explained to you the crucial role
that science plays in helping to inform our decisions about the
NAAQS. Today, what I would like to talk to you about in more
detail is the science, how it is prioritized, how it is
developed, and how it is synthesized and integrated.
As the Assistant Administrator for the Office of Research
and Development, I am responsible for both the development of
new scientific information targeted at specific Agency needs,
and for the evaluation, synthesis and integration of the
world's peer reviewed literature and to a document that informs
EPA decisionmakers.
As we characterize the current state of our collective
scientific knowledge, we are careful to point out the strengths
and weaknesses of this large body of information so that
informed decisions can be made. It is clear that scientists and
staff of EPA play a crucial role in the development and
evaluation of the world's scientific information to inform
their review of the NAAQS.
A very important function of EPA is the synthesis and
integration of these thousands of individual acts of science to
provide a clear characterization of our knowledge and the
degree to which we are still uncertain about some aspects of PM
health and environmental effects.
We have a scientifically rigorous process by which we
evaluate and interpret this important body of knowledge and we
ensure that our interpretation of them is complete,
transparent, unbiased, and consistent with an array of views in
the scientific community.
A fundamental step in the review of the NAAQS is the
evaluation of scientific evidence in the preparation of
scientific assessments known as criteria documents by the
National Center for Environmental Assessment, which is part of
the Office of Research and Development. The development of
criteria documents involves a review of thousands of peer-
reviewed research publications, evaluation of those studies
that are most relevant to the review of air quality standards
and the integration of scientific information across
disciplines.
The body of evidence must be reviewed, evaluated, weighed
and then accurately and objectively described to help inform
our decisions about National Ambient Air Quality Standards. For
the current PM review, EPA evaluated research studies that
addressed a wide range of issues, including PM toxicology,
epidemiology, atmospheric chemistry, human exposure and other
areas of environmental effects. Thousands of studies were
reviewed and over 2,000 were referenced in the final criteria
document, many of which were conducted or funded by EPA's
Office of Research and Development.
Considered together, these new studies significantly
advanced our understanding of PM's potential effects on public
health and welfare and reduced the uncertainties associated
with some important aspects of the science. Drawing on the
evaluation of studies reviewed in the PM criteria document
about health effects and dose response, as well as information
about exposures to PM, EPA also completed a risk assessment to
estimate the degree to which various approaches for revising
the standards would potentially affect the public health risks
that are posed by PM.
Further, the Agency prepared a document known as a staff
paper that utilized the evaluation and characterization of
scientific evidence in the criteria document, together with the
results of the risk assessment, to help inform policy judgments
that are required in making decisions about the NAAQS.
In summary, the Bush administration is committed to the
development and use of the highest quality scientific
information to inform decisionmaking. The mission of the Office
of Research and Development is to develop, evaluate and
communicate relevant scientific information to the
Administrator and to assure that the Administrator is well
informed of the nature, the strengths, and the limitations of
this information. EPA has sponsored a targeted and effective
research program on particulate matter, and I am pleased to
convey to you and others the value of this investment.
We have made a great effort to evaluate and characterize
the existing and new scientific results available on
particulate matter and I am personally pleased to share with
you my views on this work. I look forward to addressing any
questions that you may have, and I respectfully request that my
full written testimony, as submitted, be inserted into the
record.
Senator Voinovich. [Presiding] Thank you, Mr. Gray.
Mr. Stephenson, before you give your testimony to the
committee, I am going to ask that this report by the GAO be
made a part of the record. I look forward to your testimony.
[The referenced document can be found on pages 86-90.]
STATEMENT OF JOHN STEPHENSON, DIRECTOR, NATURAL RESOURCES AND
ENVIRONMENT, U.S. GENERAL ACCOUNTABILITY OFFICE
Mr. Stephenson. Thank you, Mr. Chairman and other members
of the committee. I am pleased to be here today as the
committee considers the science and risk assessment supporting
EPA's proposed revisions to the national air quality standards
for particulate matter.
A large body of scientific evidence over the past several
years links exposure to particulate matter to serious health
problems, including asthma, chronic bronchitis, heart attack
and premature death. EPA, as part of its authority under the
Clean Air Act to periodically review and revise as appropriate
the air quality level at which to set national standards,
proposed revisions to the particulate standards. It issued a
draft regulatory impact analysis in January, 2006 of the
revisions' expected costs and benefits.
As you know, EPA's estimates of the expected benefits from
its air pollution regulations have in the past often been
controversial and the methods the Agency used to prepare these
estimates have been questioned. As a result, at the direction
of the Senate Appropriations Committee, the National Academy of
Sciences evaluated EPA's overall methodology and approach for
estimating the health benefits of all proposed air regulations,
and in 2002 issued a report that made recommendations to EPA
to, among other things, conduct more rigorous assessments of
uncertainty, increase the transparency of how it estimates
benefits, conduct more detailed analysis of exposure, and
estimate the benefits of each regulatory option under
consideration.
Mr. Chairman, you asked GAO to determine the extent to
which EPA has implemented the academy's recommendations and our
report that you just mentioned being released today provides a
detailed discussion of each recommendation and EPA's efforts to
respond to them.
In summary, we found that while the National Academy
generally supported EPA's overall approach to estimating
benefits, it made 34 specific recommendations for improvements.
EPA is making progress in responding to the National Academy's
recommendations, but in the case of its January, 2006 analysis,
EPA had applied 8 of the recommendations, partially applied 14,
and did not apply 12.
For example, in response to the Academy's recommendations,
EPA evaluated how benefits might change, given alternative
assumptions, and discussed sources of uncertainty not included
in the benefit estimates. Although EPA applied an alternative
technique for evaluating one key uncertainty that concerning
the causal link between exposure to particulate matter and
premature death, the health benefits analysis did not assess
how the benefit estimates would vary in light of other key
uncertainties, as the Academy had recommended.
Consequently, EPA's response represents a partial
application of some of the recommendations. EPA told us that
insufficient resources impeded its progress in applying some of
the Academy's recommendations, citing in particular the limited
availability of skilled staff, time and other resources to
conduct the recommended analyses. EPA also stated that in some
cases, it did not believe the state of the scientific knowledge
was sufficient to implement other academy recommendations. One
such area has been mentioned today, is for EPA to determine the
relative toxicity of particulate matter components. EPA intends
to pursue research and development so that additional academy
recommendations such as that can be incorporated in future air
regulations.
We believe that continued commitment and dedication of
resources will be needed if EPA is to fully implement the
improvements recommended by the National Academy. In
particular, EPA will need to ensure that it allocates resources
to needed research on emerging issues such as the relative
toxicity of particulate matter, and to assessing which sources
of uncertainty have the greatest influence on the benefit
estimates. EPA officials said that they expect the final rule,
due September 27, will better address other academy
recommendations pertaining to uncertainties associated with the
health benefits estimates.
We should all be closely examining EPA's final rule because
a more robust analysis of the remaining uncertainties, not
addressing the proposal, will be critical if decisionmakers and
the public are to better understand the likelihood of actually
attaining the health benefit estimates.
Mr. Chairman, that concludes my summary. I will be happy to
answer questions.
Senator Voinovich. [Presiding.] Thank you very much.
In our last hearing, we talked particularly about asthma. I
would like to read a statement to both of you and have you
comment on it in terms of what impact do you believe that
increasing the particulate role would have on the issue of
asthma.
According to the Centers for Disease Control and
Prevention, from 1980 to 2001, the number of Americans with
asthma tripled to 20.3 million. From 1975 to 2000, the number
of asthma attacks also tripled. At the same time, EPA data
indicates the emissions of the six principal air pollutants
dropped by 53 percent since 1970. Thus, while the number of
people with asthma and the incidence of attacks have increased,
air pollution has decreased substantially over the same time.
Senator Lautenberg stated at the hearing, ``But during the
summertime when most people are enjoying their vacation playing
outside, children with asthma often have to stay indoors, and
the reason that they have to stay indoors is the air is just
too dangerous for them to breathe.''
However, staying indoors is exactly the opposite of what
children should be doing, according to a report on the indoor
air pollution. EPA ``studies of human exposure to air
pollutants indicate the indoor levels of many pollutants may be
25 times and occasionally more than 100 times higher than
outdoor levels. These levels of indoor air pollutants are of
particular concern because it is estimated that most people
spend as much as 90 percent of their time indoors. The poor
quality of indoor air and the amount of time spent indoors
seems to explain the increased incidence of asthma better than
outdoor air which is greatly improved.''
Furthermore, according to NIH, air pollution is one of the
many things that can bring on asthma symptoms. They talk about
animal dander, dust mites, cockroaches, pollen and mold,
cigarette smoke, air pollution, cold air, strong odor, scented
products, strong emotional expression and stress, medicines,
sulfites in food, or beverages.
Will you comment on how you think, if we improve this
particulate matter, it is going to make a difference in terms
of something like asthma, that is always the subject matter
here before this committee?
Mr. Gray. Senator, as many of you have stated, it is very
clear that the issue of particulate matter and the ambient air
quality standard is one of the most important decisions that
EPA makes. Particulate matter is linked with a variety of
adverse health outcomes, exacerbation of asthma being one of
them. That is, no one is suggesting that in fact particulate
matter causes asthma, but that it may be something that helps
to trigger an attack or make them worse.
That is part of the evidence, along with the other health
effects that we can consider. They were part of the rationale
behind our proposed rule that you can read about, where we
suggested a particulate matter ambient air standard. We also
asked for people's views. We asked for input from the
scientific community and from the public about that. The
information we are going to get is going to be an important
part of the ultimate decision we make. We are trying to
understand the range of scientific information on all of the
health effects of particulate matter, and the role that that
might play in the ultimate judgment of an appropriate standard.
Senator Voinovich. Mr. Stephenson, the fact that your
report says that the recommendations of the National Academy of
Sciences in terms of the things that you need to do to measure
whether or not particulate matter has an impact on health have
not been carried out by the Environmental Protection Agency. Do
you believe that they have enough information now that they can
intelligently, objectively make a decision in terms of what
this rule's impact will have on public health?
Mr. Stephenson. Well, one of the problems is the rulemaking
process doesn't lend itself to a broad picture of air quality
in general. We are looking at particulate matter in this case.
Our concern was that many of the uncertainty analyses were
done, but they were never rolled up into a more robust analysis
of all the uncertainties such that you could determine the
ranges of exposure and project the health benefits in that way.
So we think they have data. It is a question of how you
analyze that data.
Senator Voinovich. Do they have enough data to
intelligently, objectively decide whether or not for public
health, and I am not talking about weighing all the other sides
of this. In other words, they are not supposed to weigh the
impact on the economy and jobs and poor people's energy costs
and all the rest of that. They close their eyes to that. But do
they have enough information to intelligently, objectively make
a decision in terms of what impact a new rulemaking it more
stringent would have on public health?
Mr. Stephenson. In some cases yes and in some cases no.
They need more exposure data. They need more data to determine
how particulate matter affects various populations, how people
inhale the particulate matter, etc. There is another category,
which you mentioned, on the components of particulate matter,
what makes up particulate matter. As Chairman Inhofe mentioned,
asbestos is certainly more dangerous than dust. There is more
research that is needed in that area in order to determine if
there is in fact a difference between different types of
particulate matter.
Senator Voinovich. Thank you.
At the request of the Ranking Member of the subcommittee,
he asked that Senator Boxer be next.
Senator Boxer. Thank you very much. I appreciate that.
I really have one question that I want to ask to both of
you. Assistant Administrator Gray, EPA's proposal to reduce
dangerous levels of coarse particulate matter or toxic dust
relies on air pollution monitoring to detect potential
violations. However, the President's proposed budget for fiscal
year 2007 cuts these monitoring funds by $17 million. State and
local Clean Air Act administrators believe this will severely
weaken monitoring programs and likely result in significant
staff cuts across the country, so we will not have the people
necessary to tell us what this monitoring is showing us. And
what the monitoring is showing us is very important because at
certain levels, we know our most vulnerable populations, our
children, our elderly, our sick, our weakest, are impacted.
So I don't quite get why we see that kind of a cut, $17
million is not a lot in the scheme of things, when you think of
what we are spending in Iraq every week. It is in the billions.
We need the $17 million to be restored. It just boggles the
mind that would be gone.
Let me tell you the impact of it that we are being told. We
are being told that as a result of this, dangerous pollution
will go undetected in rural areas and cities, such as
California's San Joaquin Valley where we have our farms and
workers, farm workers; Missoula, MT; Carson City, NV; the
Raleigh-Durham metro area in North Carolina; the Fort Worth-
Arlington area in Texas.
I am just giving you examples of cities and regions that
don't meet the population set out here. So I am wondering if
you think that people in these areas deserve less protection
than people in larger areas. To Mr. Stephenson, given the same
set of facts, you have testified before that EPA fully or
partially implemented only two thirds of the National Academy
of Sciences' recommendations, but that Agency officials
complained of a lack of funds hampered their ability to do so.
Your testimony states that continued commitment and dedication
of resources will be needed if EPA is to fully implement the
improvements recommended by the National Academy.
This is an Administration, and my colleagues on the other
side and we join them in this. We want to do everything that is
based on science, and the National Academy of Sciences puts
forward the plans, and then we don't have the funding. So this
$17 million cut that I am coming back to, and then there is $1
million in cuts to EPA's budget for science and research for
developing and implementing clean air standards.
So I would like both of you to react to these budget cuts,
and how do you, first of all defend them, if you do, Mr. Gray
and Mr. Stephenson. Can you just give us some straight talk
about what it really means to our people?
Mr. Stephenson. Well, let me just say that we have in the
past recommended in other reports on other air issues that more
monitoring is needed. We are a data-based organization. We
support the need for real monitoring data. That is the
cornerstone of scientific research. So if there is a cutback in
the amount of data you actually collect, we would see that as a
problem.
Senator Boxer. And Mr. Gray, I assume you don't think it is
a problem?
Mr. Gray. I want you to know that the EPA is committed to
having the information that they need to make good science-
based decisions. In fact, it is our network of thousands of
monitors that we operate, together with our partners in the
States, that helps give the public information Senator
Lautenberg referred to, when there is a Code Red day. That
information comes from monitors that are like those or even
perhaps from those that are maintained by the Agency. But it
also helps to support the research that we are talking about.
Senator Boxer. Wait a minute. Excuse me for interrupting
you.
Mr. Gray. Yes.
Senator Boxer. We agree that monitoring is important. I am
asking you to tell me why, then, would you be cutting that
budget?
Mr. Gray. First of all, we have invested over $100 million
since 1998 in monitoring and specifically in monitoring
speciation, questions that we are being asked to address by the
National Academy, by this committee and by others. In addition,
we believe that we can find efficiencies. We can do our
monitoring in better ways. We are all being asked to do things
more efficiently and we are looking for efficiencies.
Senator Boxer. So you don't believe that the local people
who are telling us that as a result of these cuts there will be
pollution undetected in rural areas and cities, including for
example, San Joaquin Valley in California; Missoula, MT; Carson
City, NV; Raleigh-Durham metro. We have a whole list of them.
You think that is an overreaction and that these people will in
fact have their air monitored? Is that what you are telling us?
Mr. Gray. I am telling you that I am sure that they may be
concerned, but that we are committed to making sure that we
have the information that we need to monitor the air, to make
sure that everyone has an adequate level of protection.
Senator Boxer. So they may be concerned, but they shouldn't
be concerned because you are going to monitor their air. Is
that what you are saying?
Mr. Gray. I am saying that.
Senator Boxer. Yes or no? Are you going to monitor the air
in those areas and those regions?
Mr. Gray. Most of the monitoring that is done is not done
by the Agency.
Senator Boxer. Yes or no? Are you going to monitor the air
in those regions?
Mr. Gray. We are going to do what we need to do to make
sure that we have the information.
Senator Boxer. Mr. Chairman, they are not going to monitor
the air in these areas. You can't get a yes or no answer. It is
very disturbing. Yes or no, can't get it.
Thank you.
Senator Voinovich. Thank you, Senator Boxer.
For the record, I would like insert a letter signed by 33
members of the U.S. Senate to the Appropriations Committee
asking that they restore this $33 million to your budget, $17
million of which is to be used for monitoring. I think it is
outrageous that some of the domestic budgets of this Government
have not been given the dollars they need to get the job done
that they are being asked to do. The non-defense discretionary
budget is being clobbered, and we just stick our heads in the
sand and ignore it.
If we ask you to do a job, you ought to have the money to
get the job done and we are not giving you the money and the
Administration should wake up to the fact that we have a
problem not only in the EPA, but across the board. I am sorry.
Senator Isakson.
Senator Isakson. Thank you, Mr. Chairman.
On that subject, just to help educate me, this is a
prospective cut that is being discussed based on last year's
funding and this year's appropriation. Do you feel like you
have sufficient funding at this year's level to do the
monitoring necessary to make a reasonably informed
determination on your standards?
Mr. Gray. The information that we use to make judgments
about our standards is based on a range of scientific
information. It includes not only this monitoring.
Senator Isakson. No, no, no. Excuse me for interrupting.
But in terms of funding for monitors, which was the question, I
am just wondering if you think you have enough funding at
current levels, not contemplating the future proposed cut, to
make reasonable determinations?
Mr. Gray. We believe that we have the information that we
need to make the determinations that we need to make.
Senator Isakson. OK. Second, the GAO report, and excuse me,
but I will read this so I will get it correct, in referring to
the unexecuted 12 recommendations where you all cited
insufficient resources, the GAO report says, the Agency will
need to ensure that it allocates resources to needed research
on emerging issues such as the relative toxicity of particulate
matter components and to assessing which sources of uncertainty
have the greatest influence on benefits.
This portends that you have some discretion in allocation
that you could use to move moneys to complete all the 34
recommendations that GAO recommended. Is that correct?
Mr. Gray. We take the recommendations from the National
Academy of Sciences very seriously. Characterizing the science
well in a balanced way and characterizing uncertainty are very
important to the Agency. I, in fact, take some comfort in the
fact that we have achieved, at least partially two thirds of
those, and when our final RIA comes out in September, I think
you will find we have made even more progress.
Some of the recommendations will require additional work.
They will require us to do methods development. They will
require us to do research. At this point, for example, we have
committed over $150 million for research to help address the
question of speciation, one of the very highest priorities that
we have in addressing National Ambient Air Quality Standards.
So we are setting our priorities to help to meet many of those
recommendations that came from the National Academy.
Senator Isakson. Should you implement a rule before you
have reached all the conclusions you need to reach based on the
studies of the science?
Mr. Gray. I think that the recommendations from the
National Academy are important for us in the way we do our
Regulatory Impact Analyses. Those are very important documents
that help the Nation, that help you understand the costs and
benefits of the actions that we take. We cannot consider those
in setting our National Ambient Air Quality Standard.
Senator Isakson. You cannot consider them?
Mr. Gray. We cannot consider the costs in setting our
National Ambient Air Quality Standard. For that reason, we
evaluate the scientific evidence. We consult with the
scientific community. We have our panel of experts that we use,
and that is the information that goes into the setting of the
National Ambient Air Quality Standard.
Senator Isakson. One final question, and I want to thank
the Chairman for holding this hearing and the hearing we held I
guess last week, wasn't it?
At the hearing last week, we had a county commissioner from
Georgia, Walker County, GA, that is in non-attainment; entered
into a compact with EPA; has from what I can understand from
EPA and them, done everything that was asked of them within
their control. But they find themselves the recipient of
pollution from other continents, as well as larger cities near
them, and also Mother Nature with the Bermuda High which over
the southeast traps so much of this stuff for about 5 months
out of the year.
I know everybody has heard this before, but just one more
time. It seems to me like the restrictions and the punishment,
if you will, that somebody who has done everything admittedly
within their control to do, that we are punishing the wrong
person. There ought to be some attention to the unintended
consequences of them receiving this pollution without any
ability to stop or divert it after they have taken care of all
of the things at the local level they need to do.
So I know there was some testimony by one of the gentlemen
testifying that there is a way out through appeal to EPA? I
want to just perfect the record from last week. Walker County,
at their own expense, did extensive studies to do that, and
demonstrate the point source of this pollution, but still was
unable to get any breathing room from the EPA restrictions.
So I think we do, I am very much for clean and cleaner air,
and I am very much for doing everything we can in terms of
standards to implement that. But I also am for reasonable
common sense on the application, particularly of these non-
attainment standards when there is no control by the people who
are being restricted, and in my case it is 63 percent of my
State.
Thank you, Mr. Chairman.
Senator Voinovich. Thank you, Senator.
Senator Lautenberg.
Senator Lautenberg. Yes, thanks, Mr. Chairman.
Mr. Gray, the EPA is proposing to attain an annual health
standard of particulates that is much less protective than that
which California has adopted. Can you explain why?
Mr. Gray. In the case of our proposed rule, first, it is
important to note that we are committed to making sure that our
National Ambient Air Quality Standards fulfill the Clean Air
Act goal of protecting human health. Setting a standard, as we
have already discussed, is a judgment of the Administrator. The
job of my office, and a job that I think we have done very
well, is to present the Administrator with a picture, a
complete picture of the science, its strengths and its
limitations, and the uncertainties of what we know and don't
know, and to help inform that judgment.
In our proposed rule, we did describe the basis for the
level of 15 that we proposed, but we did ask for comments on
other levels. I want you to know that those comments are going
to be taken seriously and they will be an important factor in
the Administrator's decision. We are in the middle of an open-
ended, deliberative process that is going to result in a final
standard in September. It will fulfill that goal of protecting
human health, taking into account the science that is
available, the advice of our advisory committees like CASAC,
and public comment.
Senator Lautenberg. Do you think that we are at a point in
time where we think that there ought to be any making the
standards more rigid, to raise the standards for particulate
matter that we ought to move on with? Have we done enough
research, in your judgment, Mr. Gray, to say that there are
significant problems with the ambient air quality as we know
it? We ought to, particularly as it affects particulate matter,
should we strengthen those standards, make them more rigid than
they are? Or should we reduce them?
Mr. Gray. Again, this is an important judgment that has to
be made. National Ambient Air Quality Standards for particulate
matter are some of the most important decisions that are made
by the Environmental Protection Agency and the decisions are
not made lightly. They are made with as much scientific
information, as much scientific advice, as much scientific
input as we can muster. That information is put before the
Administrator in a balanced way, in a careful way, describing
the strengths and limitations to help inform the judgment about
whether the standard needs to be retained, needs to be raised,
or needs to be lowered.
Senator Lautenberg. Do you make recommendations when you
submit it to the Administrator as to the direction that EPA
ought to take? Or is this simply, I've done my job and here it
is? Or do you do all the decisionmaking as to the
Administrator?
Mr. Gray. My office does not do that. We prepare what we
call the criteria document, thousands of pages of scientific
information summarizing what we know and don't know. There are
recommendations that come from the staff in the Office of Air
and Radiation in EPA that make suggestions to the
Administrator. CASAC makes suggestions to the Administrator.
All of that informs their judgment.
Senator Lautenberg. You are aware of those recommendations?
Mr. Gray. I am very aware of them, yes.
Senator Lautenberg. What is the conclusion that you would
come to, hearing their recommendations? Do you think that the
standards ought to be reduced, lowered?
Mr. Gray. I think that when we put out our proposed rule
and our proposed standards, we very clearly articulated the
thinking behind them, the science behind them. We asked for
public comment, scientific information that might be out there
that we might find useful. All of that will inform our
judgments.
Senator Lautenberg. I understand the mechanics, but there
must be some conclusions out of this that say we could do
better. We certainly ought not to do worse. The Chairman talked
about a statement I made about keeping kids indoors. So I think
I can come up with a conclusion that you have to choose your
poison. In other words, if a mother unwittingly says, well, I
am going to bring my child inside, or keep my child indoors
because I think that the air quality is better. Here we come up
with a conclusion that no, that there is more danger indoors
than there is outside.
So that is quite a dilemma. I think that what we ought to
do is understand people's emotion, and understand what people's
thinking is. In the house, air conditioned, etc., and you think
that the air quality is better. I know it's easier to breathe
when you are in a house that has some air quality action going
on.
So maybe we ought to meet outside in the interest of safety
for the Senators and the visitors.
[Laughter.]
Senator Voinovich. Thank you, Senator.
Senator Chafee.
Senator Chafee. Thank you, Mr. Chairman.
Senator Isakson was talking about his county in Georgia,
Walker County, and he mentioned particulate matter coming from
other continents. In your studies, are there any other
countries that are similar to us in our research? I don't want
to put you on the spot, but that we can work with as we try and
wrestle with reconciling the health benefits toward some of the
regulations on particulate matter?
Mr. Gray. There are two parts here. One is that we, as EPA,
review the scientific information that is available to us to
help us understand the appropriate levels that inform the
judgments about the appropriate levels of air quality
standards, we use information from around the globe. We use
research that is done in the United States, Europe, and North
America. We tend to focus on certain areas, partly for
technical reasons. There is better monitoring data in the
United States than in Canada, or monitoring is done differently
than it is in other parts of the world and interpretation can
become a little tricky.
We work with international experts. We work with
international data to try to inform our decisions, to bring the
best information to bear. In implementation, we do have to
think about the fact that some of these pollutants can travel a
very long way, when we think about how we control them.
Senator Chafee. Any countries in particular really leading
the way, that we work more closely with? Or are we more on the
forefront?
Mr. Gray. Well, I don't know that I can speak to that very
authoritatively, though I will say that in many cases we do
think about international transport of pollutants, and we have
efforts under way in EPA to help reduce the pollution that
comes out of power plants in China because we know that that
can affect air quality in the United States. So we are thinking
about the global environment and trying to transfer some of the
practices that we have to other countries not only to benefit
them but to benefit us.
Senator Chafee. Again, I don't want to put you on the spot,
but I am just curious. Are there any other countries that have
stricter standards than we do?
Mr. Gray. There are countries that have mixes of standards
some parts of which are stricter than ours, and some they are
not. In some parts of the world, even in some parts of this
country, they may have standards that are very strict, but the
rules for implementation are not as strict. When we set a rule,
we mean it and it is going to mean something. That is not
always the case; for example, the World Health Organization has
set a standard that is stricter than ours, but it is completely
aspirational and has no teeth to it.
Senator Chafee. Can we learn from some of the economic
impacts from those stricter standards?
Mr. Gray. Again, because many of them are not being
implemented, because they are aspirational, it is hard to
learn. If there are opportunities to learn, we will use them.
As I said, we take our requirement for doing a Regulatory
Impact Analysis very seriously. We want to make sure that we
inform the Nation and you and anyone else who is interested in
it, of the costs and the benefits of these actions that we
take.
Senator Chafee. OK, thank you.
On another subject, does the science dictate that the
Administrator, I think this is following up on Senator
Lautenberg's question, dictate that the Administrator choose a
specific number in setting the particulate matter standards?
Does the Administrator have some leeway?
Mr. Gray. It is my understanding--and again I am a
toxicologist, not a lawyer--that in the Clean Air Act
recognizes that in fact, setting a National Ambient Air Quality
Standard is a judgment on the part of the Administrator that
considers the science, scientific advice, and a variety of
other factors. So in fact setting a national ambient air
quality standard is a judgment. In the Office of Research and
Development, we do our very best to make sure that he has an
accurate, unbiased characterization of the science to inform
that judgment.
Senator Chafee. Thank you, gentlemen.
Thank you, Mr. Chairman.
Senator Voinovich. Senator Clinton.
Senator Clinton. Thank you, Mr. Chairman.
Mr. Gray, in your testimony you said that EPA has spent
$500 million since 1997 to improve our scientific understanding
of the health impacts of particulate matter. Now, it is my
understanding, based on all of this research that has been
conducted since 1997 and the extensive review that the research
has received from outside experts, as well as the public, that
there is a consensus.
Let me ask you, does EPA have significantly more confidence
now that current levels of PM are responsible for very serious
health effects, including premature death?
Mr. Gray. When we look at the science that is available to
us, especially that which has been developed since the last
evaluation of the standard in the late 1990's, I can say that
there is better causal evidence linking particulate matter to a
range of adverse health effects. The understanding of the
levels at which that happens is much less clear.
Senator Clinton. Right, but we actually know more today
about the causal effects than we did in 1997. Is that correct?
Mr. Gray. Based on what we have learned from epidemiology,
from studies in people, from toxicology, it is believed we have
a better understanding of the causal relationship, yes.
Senator Clinton. What I am confused about is that given our
understanding of the causal effects, which is greater than it
was in 1997, given the recommendations from CASAC and others,
in your proposed rule, EPA itself said that risk assessment was
too uncertain for standard-setting purposes. What is the basis
for that assertion?
I know that we are searching for the best way to implement
a rule that does take into account the scientific evidence,
particularly the causal connections, so how did EPA reach a
different conclusion than all of the other experts who have
reviewed this matter?
Mr. Gray. I think an important point is that EPA doesn't
come to a different conclusion on this issue of causality, is
there a relationship between particulate matter and adverse
health effects. Where there is uncertainty and where the Agency
felt there was too much uncertainty to rely upon the risk
assessment for setting a standard is in the ultimate
understanding of the strength of those relationships and the
levels at which the effects occur.
Senator Clinton. Well, then may I ask you, because we are
limited, this is a very frustrating experience asking difficult
question in 5 minutes, but may I ask you then, why wouldn't our
Government err on the side of health? If there is an
uncertainty of the mechanism, why wouldn't we go with the
stricter standard as recommended by the weight of scientific
evidence?
Mr. Gray. I think I would echo your final sentence, that is
the importance of the weight of scientific evidence. What we
have done in our proposed rule is lay out what we think the
scientific evidence is for the standards that we proposed. We
asked for further comments, recognizing that others, reasonable
minds can differ on this very large and very complex body of
information. We have asked for comments. We have asked for
further information, and that is all going to play a role in
the final judgment of the Administrator for the final standard
that will be coming out in September.
Senator Clinton. Mr. Gray, may I also ask you, have you or
anyone on your staff had contact with or from anyone in the
White House concerning this proposed rule?
Mr. Gray. In the preparation of the science here, the work
that we do in the Office of Research and Development has been
done in a completely public and open way through the CASAC
process. All of our CASAC meetings are open to the public. All
of their deliberations are public.
Senator Clinton. I know, but have you or anyone on your
staff had any contact with or from anyone in the White House
with respect to this proposed rule?
Mr. Gray. I can speak for myself and say that I have not.
In terms of the science here, I cannot speak for my staff. I
don't know.
Senator Clinton. OK. But when you say in terms of the
science that is a blanket denial of any contact whatsoever from
or through the White House?
Mr. Gray. It is important to recognize, and I have said
this----
Senator Clinton. Can you answer that yes or no?
Mr. Gray. These are very important standards and that there
are interagency reviews.
Senator Clinton. Mr. Gray, I just want a yes or no answer.
Mr. Gray. Our scientific information is in our hands.
Senator Clinton. I know, but you haven't answered my
question. Have you or anyone on your staff had any kind of
contact with from or through the White House with respect to
this proposed rule?
Mr. Gray. In respect to the proposed rule, yes, of course
we have.
Senator Clinton. OK.
Mr. Chairman, could we get further information about this?
I asked because, again, I go back to the very difficult
experience I had after 9/11 when it became clear that the White
House was directly interfering with the EPA scientific
assessments and the public information provided to my
constituents. I think that this committee deserves to know
what, if any, kind of contact, pressure or other relationship
exists between the White House and the EPA with respect to this
proposed rule.
Senator Voinovich. I would like to comment that I was just
concerned about that as you are. I would really like to discuss
that because I was not aware of the fact that the White House
had had anything to do with whatever they came out with after
that. I know one thing, that when Joe Albaugh came before this
committee, no one seemed to know what was going on, and that
was one of the things that disturbed me. That is why our
legislation is so important. The President can move in and
immediately determine what folks are exposed to.
I do not see that there is any problem with it, that we
look into that. I would be surprised if there wasn't. I suspect
once you are done with this, don't you have to submit it to
OIRA to look at the cost-benefit analysis at OMB?
Mr. Gray. As I said, with our proposed rule, there is
clearly contact with the other parts of the executive branch.
Senator Voinovich. I think if you talk about the official
contact, I think it is reasonable that we get an answer to that
question.
Mr. Gray. There is an interagency process that we use with
all of our rulemakings.
Senator Voinovich. Why don't you share it with us, OK? Then
I think we will feel a lot better about it.
Mr. Gray. Certainly.
Senator Voinovich. Respond to Senator Clinton's concerns.
Senator Carper.
Senator Carper. Thanks, Mr. Chairman. Mr. Chairman, let me
ask a first question of you. Do you recall several years ago
when, maybe 3 or 4 years ago when Governor Whitman was EPA
Administrator that you and I collaborated on a proposal calling
for the establishment of a position within EPA?
Senator Voinovich. Here it is.
Senator Carper. The Deputy Administrator for Science and
Technology. There you go. You may have been talking about this
when I was out of the room.
Senator Voinovich. No, we didn't talk about it. Why don't
you talk about it?
[Laughter.]
Senator Carper. We had this notion, silly notion that maybe
we ought to have somebody who was pretty senior within EPA who
would be part of the top dog for science and technology. It is
an idea that was not warmly embraced by EPA and maybe by the
Administration. We introduced it as legislation and not a whole
lot has come of it, but I am still interested. I know Senator
Voinovich is still interested in making sure that we are making
decisions that are based on good science.
Let me just follow it up by asking you to talk a little bit
more with us about CASAC, this Clean Air Scientific Advisory
Committee. Just start out by saying, who are these people?
Mr. Gray. CASAC is actually called for in the Clean Air
Act, and its base is a seven member panel of experts with the
requirement for one member to be a physician, one member to be
from the National Academy of Sciences, and one member to be
from a State pollution control Agency. Then it is filled in
with nationally recognized experts to help to provide advice to
the Agency on ambient air quality standards.
On any specific standard, whether it is particulate matter,
ozone, lead or something else, the base group of CASAC is
augmented by topic-specific experts, again to help us in our
evaluation, characterization, and presentation of the science.
Senator Carper. Who selects or appoints these folks?
Mr. Gray. They are selected by the Administrator.
Senator Carper. What is the criteria used, just roughly?
Mr. Gray. Roughly, it is relevant expertise and knowledge.
Senator Carper. They serve a period of several years?
Mr. Gray. I don't know the exact period. Yes, they serve
for several years.
Senator Carper. OK, thank you.
Why should EPA, why should the Congress, frankly why should
anybody else heed the recommendations of the CASAC?
Mr. Gray. Well, the EPA certainly values the advice of the
CASAC. It is very, very important to us. The current proposal
on National Ambient Air Quality Standards for particulate
matter, for example, our proposal for a daily standard is
something that reflects very clearly the advice of CASAC.
Advice from CASAC, again, is one of the important factors that
are weighed by the Administrator in making the sort judgments
that are made.
Senator Carper. What are the other factors?
Mr. Gray. There are a range of other factors, comments from
outside scientists, comments from others, the range of science
that is presented to the Administrator. All of these are going
to be before the Administrator. When we proposed our standards
in every case, whether we were matching the recommendation of
CASAC or not, we asked for comment. We asked people to advise
us on levels that included everything that CASAC had
recommended across a wide range of different potential
standards. All of those comments and all of that information
will be an important part of the factors that are weighed by
the Administrator in making a final judgment.
Senator Carper. Just to understand this, the current
standards that exist are for annual, and there is a standard
for daily exposures. If my life depended on it, I don't know
that I could well explain what a microgram is, but I understand
that the current PM standard on an annual basis is 15
micrograms per cubic meter. On a daily basis, it is 65.
Just explain that. As a toxicologist, I am glad that you
aren't a lawyer for this question, but as a toxicologist and
someone who studied some science himself, explain that so that
the lay person can understand it.
Mr. Gray. The numbers you are citing are the concentration
of particles that are in the air. These are particles that are
all around us right now. We can't see them. In our monitoring
system, we suck air into a collector, put it on a filter, and
measure it, and that is how we learn about how much is there.
That information plays a role in our doing epidemiology,
comparing the health effects in populations with higher levels
of exposure to those with lower levels of exposure. It helps us
to understand what happens when we give animals exposure to
this, to help us understand the mechanisms by which adverse
effects might occur.
So these standards, these numbers are simply ways of
identifying a certain amount of this particulate matter that is
in the air.
Senator Carper. Again, drawing on your own expertise and
training as a toxicologist, could you describe for us how these
small, tiny pieces of matter, how they actually contribute to
asthma, how they contribute to chronic bronchitis, how they
contribute to heart disease or irregular heartbeat? How does it
actually happen?
Mr. Gray. That is one of the areas that we have learned
much more about in the last 10 years, partially through
research that has been funded by the Environmental Protection
Agency. We know that at certain levels of exposure, for
example, levels that are frankly higher than what we have in
the air today, it can increase the thickness of atherosclerotic
plaques.
Senator Carper. Say that again?
Mr. Gray. Atherosclerotic plaques, atherosclerosis, the
stuff you worry about with cholesterol building up and making
your arteries get small. That can happen in mice. They are
sensitive mice. They are bred to be sensitive, but that can
happen to them if they are exposed to high levels of
particulate air pollution. It is one example of a study that
has been done to help us understand how these particles may
have their adverse effects.
Senator Carper. My last question, Mr. Chairman, would be
this, we have our current standard, and I think you said the
annual standard is 15 micrograms. The CASAC had recommended I
think going down to anywhere from 12 to 14. EPA chose to stay
at 15. With respect to the daily standard, the current standard
is 65. The CASAC had suggested going anywhere from 25 to 40,
and you have come in at EPA at 35 micrograms.
You chose not to adopt or move toward the annual standard,
and you chose, or adjusted the daily standard. You went to the
high end of the CASAC recommendation. Could I ask why?
Mr. Gray. Again, these are judgments on behalf of the
Administrator and they are laid out very nicely in our proposed
rule with the scientific reasoning behind each of the choices
that were made. But the important thing, and I want to
emphasize this again, I have said it a lot of times, we asked
for comments about a range of other potential standards to get
information from the outside community, from scientists, from
the interested public about other potential levels, and that is
all going to play a role in setting the final standard, and it
will happen in September.
Senator Carper. All right. Thanks very much.
Mr. Stephenson, sorry I didn't get to you, but another day.
Thank you.
Senator Voinovich. I would like to thank both of you for
your testimony before the committee. Mr. Gray, I want to thank
you for your fine testimony here today, not an enviable
position if you are getting shot at from both sides on this
one. So maybe that is good. Thank you very much.
Our next panel will come forward: Dr. Roger McClellan who
is the Advisor for Toxicology and Human Health Risk Analysis,
and also is a CASAC, Clean Air Science Advisory Committee
member; Dr. George Thurston from New York University; Dr. Anne
Smith from CRA International; and Mr. Dan Greenbaum from the
Health Effects Institute, which is jointly funded by EPA and
the industry.
I would like to urge the witnesses to limit their testimony
to no more than 5 minutes. Your entire testimony has been
inserted into the record. We really appreciate your being here
today. We look forward to hearing what you have to say.
Dr. McClellan, we are going to start with you.
STATEMENT OF ROGER McCLELLAN, ADVISOR, TOXICOLOGY AND HUMAN
HEALTH RISK ANALYSIS
Dr. McClellan. Thank you very much. Good morning, Mr.
Chairman and members of the committee. I appreciate this
invitation to present my views on EPA's current review of the
national ambient air quality standard for particulate matter.
Since 1999, I have served as an advisor on issues related
to air quality, drawing on my more than 45 years of experience
in comparative medicine, toxicology, and aerosol science and
risk analysis. In particular, my testimony draws on my
experience serving on the Clean Air Scientific Advisory
Committee, which I chaired from 1988 to 1992, and service on
all of the CASAC PM panels from the late 1970's to the most
recent PM panel, as we moved from consideration of total
suspended particulates to PM10 to PM2.5,
and now consideration of a PM10-2.5 standard.
I would like to make several points. First, I want to
emphasize there is no scientific methodology which can
determine a specific indicator, precise averaging time,
numerical level, or statistical form that will be adequate to
protect public health. The available scientific information can
inform those decisions, however the Administrator must
ultimately use policy judgments in making decisions on each of
those four elements of the standard, drawing on an array of
scientifically acceptable options.
Two, I personally find acceptable the Administrator's
policy choices for the PM standard as were published in the
Federal Register. Specifically, I find acceptable a proposal to
reduce the 24-hour standard from 65 to 35 micrograms per cubic
meter, with a 98th percentile form. I found it acceptable in
terms of science for him to propose retention of the
PM2.5 annual standard at 15 micrograms per cubic
meter. With reluctance, I concurred scientifically with the
setting of a PM10-2.5 indicator for the 24 hour
averaging time concentration set at 70 micrograms per cubic
meter, within a 98th percentile form.
I say with reluctance because the science base for that is
extremely weak and uncertain. I would have preferred retention
of the PM10 standard.
Third, it is important to recognize that although the
criteria document is hundreds of pages in length and compiling
the results of what we know in terms of the world of science
about PM, at the end of the day the key information for setting
the standards are the results of the epidemiological studies.
That has been the basis for changes in standards from a total
suspended particulates set in 1971, to the PM10
standard set in 1987, PM2.5 in 1997, and now
consideration of this new PM10-2.5 standard.
So the current review focused on the PM2.5
indicator. Now, some might have said, well, why didn't we look
at sulfates or elemental carbon or some particular chemical
species? The fact of the matter is we are chained to the
``regulatory compliance lamppost'' in terms of monitoring data.
You cannot conduct epidemiological studies unless you have the
monitoring data. Our past obsession with monitoring that which
is regulated has precluded the development of the richer
science base that really need to consider these options in
terms of speciation.
I am not optimistic that that is going to change. I can
envision us sitting here 5 years, 7 years from now and
discussing, well, why not for specific standards; why didn't we
perhaps preclude or set outside of the standards on material,
because it is very innocuous; or why didn't we focus on
something in particular? In fact, we won't have the
epidemiological data because I don't think we are developing
adequate monitoring data.
Fourth, we can then turn to the question of toxicology. I
am a toxicologist. I am intellectually married to the subject,
if you will. We have exciting new methods at hand, but at the
end of the day, I have to say that our toxicology methods are
simply too blunt and yield results that can only be
qualitatively extrapolated to the human population. I know of
no scientific method for using the results of the toxicology
studies with PM, including those conducted with specific
chemical constituents, to develop quantitative numerical
standards that are the core of the PM NAAQS.
Fifth, we have heard some discussion about the issue of
uncertainty. I certainly, as a scientific colleague of Dr.
Gray, am an enthusiast for trying to determine how we can bring
our scientific information together and relate all of the
uncertainties, so we can have better informed policy decisions.
I want to comment on one aspect of that, and that is the
use of expert elicitation. That was covered in the GAO report.
I served as one of those five experts on the pilot expert
elicitation. I hesitate to say whether it is expert advice or
expert opinion. It is a challenge when you are asked to present
information. The answers that are given can be heavily
influenced not only by your knowledge of the science, but your
personal choices. All of us want quality life. Well, how do we
separate that from our judgment on the science?
I am concerned about the use of that expert elicitation
advice, and I certainly would urge the Administrator to use a
high degree of caution in using that in the regulatory impact
analysis and in making decisions on the standard.
Senator Voinovich. Dr. McClellan, could you finish up?
Dr. McClellan. I am going to just wrap up with that and
emphasize that one of the challenges we have today is
separating out people's views on the science versus their views
on the science wrapped up with the policy and some desired
outcome--the level of the standard. I think in some cases, as
individuals and professional groups have weighed in, they are
weighing in not just on the science, but the policy outcome
they want.
I do think it is important to have a distinction between
the science and policy choices. In my comments on the
acceptability of the Administrator's choices, I viewed the
science options laid out in the staff paper as acceptable, and
he used his judgment in selecting from among those options. I
think that was appropriate.
Thank you.
Senator Voinovich. Thank you very much for being here.
Dr. Thurston.
STATEMENT OF GEORGE THURSTON, ASSOCIATE PROFESSOR, NEW YORK
UNIVERSITY, SCHOOL OF MEDICINE, DEPARTMENT OF ENVIRONMENTAL
MEDICINE
Dr. Thurston. Good morning. I am George Thurston, a tenured
Associate Professor of Environmental Medicine at the New York
University School of Medicine, where my research involves the
investigation of human health effects of air pollution.
I first wish to present for inclusion in the record letters
from many major medical societies and public health groups,
including the American Medical Association, the American
Thoracic Society, and the American College of Chest Physicians,
the American Public Health Association, the American Lung
Association, the American Heart Association, the American
Cancer Society, and more. These letters show the unprecedented
support that exists for PM standards that are much tighter than
EPA has proposed.
Especially note that they all agree that the science
supports tightening annual standards to no more than 12
micrograms per meter cubed. If this were to be done, many
thousands of premature deaths could be avoided each year.
I also present a copy of Dr. Rogene Henderson's letter from
the EPA's Clean Air Science Committee, CASAC, urging that the
Administrator comply with their recommendations, with CASAC's
recommendations, to implement more stringent PM standards than
now proposed by the EPA.
Now, in my written testimony, I have addressed three
factors that need to be considered in the EPA's proposed
revisions to the particulate matter air quality standards.
First, I address the fact that we are now far more certain, as
discussed earlier, of the adverse impacts and biological
mechanisms of PM health effects. The uncertainties raised at
the time of the initial setting of the PM2.5
standard are now greatly reduced.
As outlined in figure one, if someone could put that up,
from my testimony, the PM research funded since the setting of
the last PM2.5 standard has collectively shown the
existence of numerous biological pathways capable of causing
damage to the human heart, the lung, the nervous system and the
circulatory system. This is consistent with the health impacts
found by the PM epidemiology studies upon which the
PM2.5 standard was set.
This has greatly reduced scientific uncertainty associated
with the mechanisms by which PM has such severe effects on
human health.
Second, I documented reducing ambient PM levels can and do
result in significant reductions in the mortality risk
associated with this pollutant. Since the setting of the
original PM2.5 standard, more recent follow-up
analyses of the landmark Harvard Six Cities and ACS studies
have now considered longer records of time and have confirmed
and expanded the conclusions from these two major studies.
As shown in figure two, an extended analysis of the Harvard
Six City study through 1990 has now shown that reductions in
long-term ambient PM pollution results in concomitant
reductions in health risks associated with PM. Large reductions
in PM at four of the Harvard cities have resulted in likewise
large reductions in the relative risk of mortality in those
cities. We see S, Steubenville, where the pollution levels have
come down, and so has the risk of mortality from that
pollution. In Harriman, TN, the pollution levels have come down
and the risk of mortality in that city has diminished as a
result of the lower pollution. St. Louis, the L, has come down
similarly, and Watertown, right, Boston, has similarly come
down.
So the places where they have had improvements, we see that
lower PM2.5 lowers mortality. But I think it is
important to say we still have a long way to go, and a lot of
improvement yet to be made and benefits to be reaped from
lowering the PM standard.
Finally, I show that the adverse health effects of PM air
pollution extend below the PM2.5 standard of 15. A
recent NIOSH-funded extension of the ACS study, of which I was
a principal investigator, strengthens the original conclusions
of the ACS study, and it importantly now links increased risk
of lung cancer to long-term exposure to PM, as shown here. As
the pollution level goes up, the risks of lung cancer rise, as
do all cause and cardiopulmonary.
As seen in this figure, the risks from PM2.5
extend well below 15 micrograms per meter-cubed.
In conclusion, since it was the level of uncertainty about
PM biological mechanisms and effects at concentrations lower
than 15 micrograms per meter-cubed that limited the standard to
that level in 1997, and, I point out, not some specific
acceptable level of health risk from PM, and since new sound
scientific studies have greatly reduced or resolved those
uncertainties, then concern about the health of the public
clearly indicates that the long-term PM2.5 standard
should now be reduced below 15, consistent with CASAC's advice.
And finally, I just want to point out that I was involved
with the actual expert elicitation that has just been finished.
It is finishing up under EPA. It is clear from the expert
elicitation; I was shown the results of all the experts at the
final meeting in New Orleans a few weeks ago. It is clear that
the experts are, there is a consensus that, the risk from fine
particle is much higher than previously thought, and that
expert elicitation gives us an estimate of somewhere on the
order of 1 percent decline in mortality per microgram per
meter-cubed of fine particles.
So we are talking about a very large reduction in health
risk, if you consider the fact that over two million people die
every year in the United States, a reduction of 1 percent in
that would be 20,000 deaths per year. So we are talking about
many thousands of premature deaths that can be avoided by
lowering this pollution.
I did want to respond to the question that you raised about
asthma, if I have a second. Do I have time to respond to that?
Senator Voinovich. Why don't you bring it up in the
question period.
Dr. Thurston. OK, we will talk about it then.
Well, thank you for this opportunity to testify.
Senator Voinovich. Thank you.
Dr. Smith.
STATEMENT OF ANNE SMITH, VICE PRESIDENT, CRA INTERNATIONAL
Dr. Smith. Mr. Chairman, thank you for inviting me to
participate in today's hearing. I am Dr. Anne Smith. I am a
Vice President, CRA International.
I have been analyzing the risks and policy options
associated with fine particle standards for over 10 years. The
opinions I will present today are my own and not those of my
company, CRA International.
When EPA set the first ever national ambient air quality
standards for PM2.5 back in 1997, in the face of
substantial knowledge gaps at that time, it presumed that fine
particles do have a causal relationship with public health. The
law then required that standards be set for that
PM2.5 at a level that would protect the public
health with an adequate margin of safety.
The courts ruled that the current standards they set did
indeed provide that adequate margin of safety. EPA and the
courts have also made it clear that that margin of safety does
not eliminate estimated risk. Today, EPA is deciding whether
new evidence accumulated since 1997 justifies tightening these
standards. In both 1997 and now, EPA prepared a quantitative
risk analysis using the available health studies. EPA says this
is to help decide whether to tighten the standard. As I will
show, it can help with this.
However, if you take the position that the PM2.5
standard has to be tightened just because the risk analysis
produces a body count at the prevailing standard, you will find
yourself having pre-decided to tighten the standard in every
future review cycle, even if the evidence never changes in the
interim. This cannot be what Congress intended.
In thinking about whether to tighten the standard, the more
appropriate question to ask is: Has anything changed in our
knowledge that would undermine the Administrator's 1997
judgment regarding the adequacy of the margin of safety? Having
a large number of new studies today, having successfully
reanalyzed the studies we had originally, and even having some
evidence of a still elusive biological mechanism, all fail to
provide a justification to tighten the standard. They merely
confirm that a standard is needed in the first place, which was
the judgment made in 1997.
However, one could justify tightening the standard if the
new evidence that I have just spoken of indicated that risks at
the attainment level of the current standard have increased
since 1997. To test this, I went back to my 1997 files to
construct a comparison of EPA's estimates of mortality risks on
a then and now sort of basis. I found that the estimates of
mortality risk at the attainment level have actually fallen
since they were first estimated in 1997.
Risks due to long-term mortality have fallen in every
location. The risks due to daily exposure, which vary by city,
have fallen in six of the eight cities that are in EPA's risk
analysis. This new information suggests that the margin of
safety provided by the current standards is actually greater
than we originally thought.
What I have told you so far is based entirely on EPA's own
point estimates of risk. However, it is worth looking more
closely at those estimates and the basis for them. Every one of
the new epidemiological studies actually contains multiple
alternative risk estimates. Sometimes there are dozens of them
within a single paper. Which one is best? Which one should be
used in the risk analysis? The Health Effects Institute has
pointed out in the face of this issue that there is no gold
standard for deciding this.
When I reviewed the full body of evidence in each paper
that EPA used for its risk analysis, I went back to the
original papers, I noticed that EPA had consistently selected
the highest risk estimate or nearly the highest one from each
of the papers. Therefore, EPA's risk estimates overstate what
the full body of evidence supports today.
The evidence on fine particle risks has weakened in a
number of other ways. Several of EPA's own point estimates
today of the risk are not statistically significant. What this
means is that behind the purported body count that comes out
the risk analysis there is also actually a rather large
probability of no health impact at all associated with that
very same estimate. I found that none of the risk estimates for
any of the eight cities in EPA's risk analysis remains
statistically significant across all of the reasonable
estimates in those papers.
This body of new evidence has also substantiated some of
the concerns that were raised in 1997 that fine particles may
be a scapegoat, while the real villain hides behind a veil of
statistical uncertainty. The only clear trend in the new
evidence has been that the estimated levels of risk that are
being attributed to fine particles are lower than originally
thought.
I would like to conclude by saying that we still have no
idea of which constituent of PM2.5, if any, is the
culprit underlying this complex body of evidence. Even if the
risk estimates are good, correct risk estimates, by regulating
this generic mass of many compounds that is PM2.5,
we may gain little if any public health benefit.
Thank you for this opportunity to share my views on this
important topic.
Senator Voinovich. Thank you, Dr. Smith.
Dr. Greenbaum.
STATEMENT OF DANIEL GREENBAUM, PRESIDENT, HEALTH EFFECTS
INSTITUTE
Mr. Greenbaum. Thank you, Mr. Chairman. It is a pleasure to
be back in front of the committee. I thank you for the
opportunity to testify.
I am the President of the Health Effects Institute, which
as you know is an independent research institute funded jointly
by the EPA and industry to provide high quality, impartial
science on the health effects of air pollution.
I am going to focus today on the science progress we have
made since 1997, the most recent findings on the relationships
between different ambient concentrations of PM and health, and
key science needs going forward.
Since Congress identified the need for substantial enhanced
research in 1997, much progress has been made. We know much
more about the sources of fine particles and about personal
exposure. We have conducted the first multi-city epidemiology
studies and analyzed and reanalyzed many of these studies,
finding that in general the earlier studies were well done.
At the same time, there has been some evidence that the
health effects we have seen in those earlier studies may in
some cases be smaller. Unlike in 1997, we now have a number of
toxicology studies that have begun to indicate potential
biological mechanisms by which PM may cause health effects,
although as always there is still much more to learn.
Among the most important questions addressed in recent
years is whether exposure to PM has health effects at all
levels, or whether there is a threshold below which no effects
are expected. This question is central to setting the level of
the NAAQS. This has been looked at in both short-term and long-
term studies. Perhaps the most rigorous short-term study
conducted since 1997 is the National Morbidity, Mortality and
Air Pollution Study, which was funded by HEI and led by Johns
Hopkins. That study examined daily changes in air pollution and
health in the 90 largest U.S. cities.
As shown in figure one in my testimony, the study found
that there appeared to be a linear relationship between
mortality and air pollution for all causes of mortality and for
deaths from heart and lung disease without an apparent
threshold. Our review committee, which intensively peer-reviews
all HEI research, advised caution in drawing conclusions from
the apparent absence of a threshold, but noted that the
reported associations in the study are at ambient
concentrations well below the current U.S. daily standard.
There were two principal long-term studies in 1997, the
Harvard Six Cities and the American Cancer Society study, and
there have been some additional studies since, but attention
has primarily focused on HEI's reanalysis of these two and on
extended analysis in the ACS population.
For the reanalysis, we gained access to all underlying data
in the studies and chose independent investigators who tested
the original studies against a wide variety of alternative
explanations. The investigators also conducted an analysis of
the concentration response and figure two in my testimony
presents the results, summarizing for each community the annual
air pollution level and the risk of death due to heart and lung
disease. As you can see, there is some scatter of effects. It
is not all in a straight line, especially at the highest and
lowest PM levels studied, but also an overall trend of
increasing mortality risk with increases in pollution levels,
starting at relatively low levels.
Following the reanalysis, the original investigators for
the ACS study conducted an extended analysis of the data, which
Dr. Thurston has mentioned and participated in, and found
similar results.
While we have made much progress, there continue to be, as
there always are in science, important questions. Two key areas
are: First, we need continuous improvement in the statistics
used in epidemiology to better test the sensitivity of the
results, to quantify the uncertainty, and to communicate both
the results and the uncertainties clearly.
Second, no other question will have as much impact on
future regulations than determining whether some components of
the complex mixture of PM are more toxic than other components.
Ultimately, this data will be essential to ensuring that
regulations are targeted at reducing those emissions which will
have the most public health benefit at the least cost.
This has also become important in light of the current
proposal for a PM NAAQS for coarse particles, which is proposed
to exclude certain particles from regulation. There have been
some individual city studies of this question, but no
systematic national effort to compare results from both
epidemiology and toxicology. To fill that gap in time to inform
a next round of PM NAAQS review, HEI has launched, with support
from both EPA and multiple industries, a set of systematic
multidisciplinary national studies.
As indicated in both the NRC review of PM research
priorities and in today's GAO report, these PM component
studies will be central to ensuring that future PM actions are
the most effective possible.
I might add that while I obviously have the utmost respect
for Dr. McClellan and take his skepticism that we can answer
this question as a true challenge, as a long-time Boston Red
Sox fan and a hopeless optimist, I enter this challenge with
the hope that we can, if we all put our minds to it, answer
this question of which components much better 5 or 6 years from
now than we can today.
Thank you for this opportunity to testify. I would be
pleased to answer any questions you might have.
Senator Voinovich. I want to thank all the witnesses for
your testimony this morning.
Before I start the questioning, Chairman Inhofe asked that
I insert testimony for the record from Dr. Borack on the issue
of coarse particulate matter. Without objection, we will do
that.
[The referenced Testimony can be found on pages 171-187.]
Senator Voinovich. Mr. Greenbaum, as you know, there is an
ongoing debate about whether EPA must lower the annual
standard. We talked about it last week, and this week we are
trying to get some of you that are smarter than we are in the
scientific area to give us your best thoughts.
Your opinion is invaluable in this debate as an entity that
sits squarely in the middle. You are funded by both industry
and the EPA. Clearly, the statute says the standard shall be
set, ``in the judgment of the Administrator.'' In your opinion,
is it reasonable from a scientific and health perspective for
the Administrator to retain the annual standard at 15 and not
lower it?
Mr. Greenbaum. Well, Mr. Chairman, I appreciate that that
is the $64 million question at this hearing. I tried to present
the best view of the science that we at HEI can give you on
that question. While the dose response or the relationships
going down to the lowest levels are very clear for why we are
tightening the daily standard, they are strong, but not as
clear, for the long-term studies, as shown in the data I
presented.
There is evidence with cities showing reduced effects as
you go down below 15 micrograms per cubic meter. I think it is
exactly in that area where there is some certainty about these
results, but where the Administrator has to make a policy
judgment about whether that is certain enough to require
additional actions. As Dr. McClellan said, there is really no
scientific method even that HEI has that could set that.
I should add that we at HEI, since our inception, have
never taken a position on a specific standard because of the
clear concern that such a position would be viewed as therefore
biasing any of our future science. So we don't take positions
on this. We don't advocate. The science is definitely stronger
on the concentration.
Senator Voinovich. The question is, is it reasonable?
Mr. Greenbaum. I think I am going to have to leave that to
the Administrator to make that judgment. I think that there is
evidence that we didn't have in 1997 of effects still
continuing below 15. But whether that is enough evidence to
make a call for a reduction in the standard is a public health
policy judgment, not a scientific judgment.
Senator Voinovich. Dr. Smith, what is your opinion?
Dr. Smith. The way to think about setting the annual
standard is to look at the chronic health studies. The chronic
health studies have been the ones that have been found to have
FE.
Senator Voinovich. What studies again?
Dr. Smith. I am sorry. Chronic health studies, which are
studies of long-term exposure to PM2.5 so they are
more relevant to measures like the annual standard and annual
average, rather than to a day to day peak, which is the daily
standard. So the studies that look at the long-term exposure of
PM2.5 with mortality are the ones that are in
question for whether we need to tighten the annual standard.
Those studies are some of the most difficult ones to properly
control for in statistical methods.
What we found in the reanalysis that Dr. Greenbaum's
organization performed is that in fact when reanalyzed with
alternative methods of control, the estimate gets more and more
insignificant as additional factors are brought into play such
as SO2. SO2 when added into the analysis
causes the PM2.5 health effect to become
insignificant, issues that are complex statistical issues, and
I will use a technical term, spatial autocorrelation, are
problems if you have those in the analysis. They were found in
the health effects studies for long-term standards and when
controlled for, the effect becomes statistically insignificant,
again meaning a very large probability that there may be no
effect at all.
So the evidence has weakened since the standard was set at
15 micrograms per cubic meter based on that data set. I will
also say, when that standard was set, there was evidence of
effects below 15 in that data set. That hasn't changed. That is
not new information. So while the analyses have held up to
reanalysis and been confirmed that you can find this sort of
correlation in the long-term studies, they are much weaker in
terms of how well you can make a causal interpretation of them.
Another very important finding that came out of the Health
Effects Institute's reanalysis is that in fact if you break the
population that the data, if you break the cohort that is being
studied into three groups, different educational levels, and
you look at the group that has even 1 year of education beyond
high school, which is what good deal of our population has,
there is no effect at all. It is gone. It is not in that
analysis.
So the effects are very dependent, the effects that are
being purported and attributed to PM2.5 out of that
study are now known to be very dependent on whether a person
has had any education beyond high school or not. That
relationship is found in all the studies.
Senator Voinovich. What does that mean?
Dr. Smith. It suggests that there is a missing variable in
the equation, basically. We know that education does not in
itself make somebody more or less susceptible to exposures to
fine particles, but there may be something.
Senator Voinovich. Does this mean that the people that are
1 year in high school live out in the suburbs and others live
in the urban area and are more impacted by this?
Dr. Smith. It may have something to do with the poor
evidence on what they are being exposed to. It may be something
about lifestyle that actually creates risk for them that the
PM2.5 is picking up on, but isn't really
exacerbating. It is just serving as another thing, a scapegoat.
Senator Voinovich. Let me ask you something. If I am a
smoker, and please forgive me if these are elementary question,
but if I am a smoker, am I more subject to fine particle PM
than if I am a nonsmoker?
Dr. Smith. I actually am not quite sure how that
relationship works out in the study, but they have controlled
for smoking in the relationship. Certainly, smoking is related
to educational level, but smoking was controlled for when they
did this analysis of whether there was an effect at different
educational levels. So the correlation between smoking and
educational level is not the education for what is going on
there. It means there is something totally different at play.
Until you find out what that problem is, through more
study, more research, and it can be done, until you find out
what that problem is that explains this unexpected and bizarre
result, you don't have an unbiased estimate of what the risk
is. It could be larger or it could vanish, the PM2.5
risk, that is.
Senator Voinovich. OK. In terms of the particulate matter,
again, we are trying to reduce NOX and
SOX, mercury. When they do the analysis of the
particulate matter, they just segregate that out and you don't
take into consideration the impact that NOX or
SOX. Or does particulate matter have something to do
with NOX and SOX?
Dr. Smith. The results that are being used in the risk
analysis do not account for SOX or NOX or
any of the other pollutants and whether they may have an
explanatory role, too. As I said a minute ago.
Senator Voinovich. In other words, when you are doing the,
I always have a tough time with that word, epidemiologist do
the studies, OK, the fact is that they are just looking, they
are trying to figure out just what the particulate matter is.
They don't take into consideration that I am exposed as a
person to both NOX, SOX, particulate
matter, mercury and the rest of it, and somehow they are able
to pull out and say these particulate matters are the things
that are really contributing to morbidity.
Dr. Smith. As I was saying, the mortality, as I was saying,
the estimates that are being used right now to estimate risk
from PM2.5 only consider PM2.5. However,
the studies in the many, many other results that are in these
papers look at other formulations, other methods of doing the
estimate that put the other pollutants in to the model as well
and try to understand which pollutant is it. Does one of them
explain the effect that we are seeing for PM2.5 when
that is the only one in the model? Is there another pollutant
that could explain it better, when we put them both in the
model?
What is found in study after study after study when this is
done is that the PM2.5 effect goes away, becomes
statistically insignificant is what I mean. The other pollutant
that is in the model that is being explained by the analysis at
the same time retains its statistical significance. I looked at
all the PM2.5 short-term studies that I could find
that did this analysis, this type of two pollutant analysis. I
found that the vast majority of them had the PM2.5
effect that was statistically significant if it was the only
one in the model, become statistically insignificant when one
of the gaseous pollutants, and these included ozone,
SO2, NOX, and carbon monoxide. The same
thing is going on in the chronic studies for SO2.
Senator Voinovich. To get back to the original question. If
the Administrator, is it reasonable from a scientific and
health perspective for him to retain the annual standard at 15
and not lower it? Is it reasonable?
Dr. Smith. Given the basis for the standard of 15, which
was the same chronic studies that I have just described to you,
but before we studied them and discovered some of the
underlying uncertainties in them, and given that the risks are
lower now based on the new evidence using their point
estimates, it seems quite reasonable to say that we haven't
learned anything that says that that standard has a smaller
margin of safety than it had when the standard was set.
Senator Voinovich. So you would say that if he did it, it
would be reasonable.
Dr. Smith. That could be reasonable.
Senator Voinovich. May I ask another question? In terms of
sulfur and NOX, in terms of public health, and
particulate matter, which is the worst?
Dr. Smith. Between sulfur and NOX SO2
and NOX
Senator Voinovich. Sulfur dioxide or NOX, if you
were able to reduce those substantially, in terms of public
health, which one of those is the worst?
Dr. Smith. I really cannot say.
Senator Voinovich. I would be interested in all of you.
Dr. Smith. I just don't know. I have not actually done a
relative analysis of SO2 and NOX.
Senator Voinovich. Well, has anybody done an analysis at
the table? Dr. Thurston, you have been around a long time. You
have testified a lot of times before this committee. Do you
have an opinion?
Dr. Thurston. Well, yes, I would like to respond to a few
things that were just said, including that.
Mr. Greenbaum. Senator, I realize you are going to go
through everybody, but I would like, since it was HEI who
conducted the reanalysis I would like to have the opportunity
to clarify for the record what we did conclude.
Senator Voinovich. Get your mike there, please. Go ahead.
Mr. Greenbaum. I'm sorry.
Senator Voinovich. OK, go ahead.
Mr. Greenbaum. I just want to, and I don't mean to take
away from Dr. Thurston, but I did want to clarify what the HEI
reanalysis of the long-term studies in effect did and didn't
find, because there has been some degree of confusion just
thrown into it unfortunately by Dr. Smith's comments.
First of all, I think everybody can in this process pick
their favorite study and their favorite result from the study
and try to emphasize that, so they can emphasize the one that
has multiple pollutants in it and say there is no effect, or
they might emphasize the one that has no other pollutants but
PM in it and say there is a big effect of PM.
Our Review Committee, in our review of the reanalysis--this
very intensive look these studies--had to look across all of
the analyses to figure out which ones were the right ones,
which ones were the wrong ones, which ones were statistically
stronger or weaker. While, of course, that group of scientists
identified continuing questions for those studies, we did find
as strength of the evidence that the effects of PM were robust
to a number of analyses that were done.
We also found for the first time a strong effect of sulfur
dioxide in that, although our committee, and this is in partial
answer to your question, did not think that sulfur dioxide
itself could be attributed to doing that, based on toxicology
information, but that it might be a marker for something else
in the air that would travel with it, but we are just not
measuring that, and therefore we don't have information about
that other substance.
The reanalysis definitely found, after test after test
after test, stronger results from those studies, with some
continuing questions. We certainly had that.
Second, on the question of education, and how education
modifies the results, certainly there was a stronger result
shown for those with the least education in terms of increased
mortality. We do largely view educational levels as one marker
for poverty or socioeconomic status. We know that poverty and
socioeconomic status themselves can have effects on health.
People can have less good medical care, less good nutrition,
and a number of other things that shorten their lives. We know
that and our reviewers certainly understood that as well.
In that analysis, in the reanalysis, however, when you
looked at heart and lung disease deaths, which is not what Dr.
Smith put in her testimony, you still see some effect in
everybody else in the population. That effect was marginally
statistically significant, but still there. Our investigators,
our reviewers felt two things, that: One, that there was a
continued effect at these different levels, with the strongest
effect being that those with the least education and probably
least socioeconomic status, and that might be because of higher
exposures, less access to air conditioning, and a number of
other things that would increase the effect.
Two, they also did note that no study is perfect, so there
probably is still, somewhere in there, some continued
confounding where the results are showing an effect of being
poor, but they didn't suggest that that explained the whole
effect. I just wanted to clarify that.
And then they did test all of the pollutants in the
reanalysis, including nitrogen oxides and sulfur dioxide and
ozone. Sulfur dioxide was the only one that showed effects and
had any effect on the PM results. As I said earlier, they
thought that was important to understand. They understood that
we need to do more work on that. They weren't sure that it was
the sulfur dioxide. In fact, the sulfur dioxide might have been
a marker for other types of particles that are coming along
with the sulfur dioxide, not the gases themselves. There could,
for example, be metals that come out of the same sources.
So I thank you for the opportunity. I just wanted to
clarify what our reanalysis did say, and obviously the full
report was actually presented to you at an earlier hearing.
Thanks.
Senator Voinovich. Dr. Thurston.
Dr. Thurston. Yes, thank you.
Before we start out, you were noting that Dr. Greenbaum is
between Government and industry. I just want to point out that
I represent, I think, a disinterested party. I am a tenured
professor at NYU. My salary is paid by NYU. I don't do
consulting for the vested interests involved here. So I feel
that my testimony should be viewed as a disinterested party,
not an uninterested party, but a disinterested party.
I have to agree with Dr. Greenbaum's discussion of the
SO2 and the HEI reanalysis. SO2, it is
clearly stated. You know, I think that unfortunately what Dr.
Smith has done here is exactly what she accuses EPA of having
done. She has cherry-picked certain results that support her
position and unfortunately hasn't looked in a balanced unbiased
way at this question.
Senator Voinovich. Can I ask you something?
Dr. Thurston. Yes.
Senator Voinovich. What is it that contributes to the
particles? You have coarse particles, you have fine particles.
We are becoming more concerned about fine particles. I know I
have been at the University of Cincinnati, the Children's
Hospital down there. They are doing a study of diesel fumes on
urban kids----
Dr. Thurston. Yes, I have done that myself.
Senator Voinovich [continuing]. The development and so
forth. But we are talking about the things that cause the
particles. Is it basically emissions from automobiles? What is
it? What is the cause of this?
Dr. Thurston. On that question, you made a good point, you
know, that these are oftentimes markers. I think the
SO2 being a marker of largely power plant pollution
in the United States; NOX being largely a marker of
traffic. It is very likely, actually, that that SO2
in the ACS study, because I have done some of this analysis, of
course, as being PI of the NIH-funded portion of that, that the
SO2 is a marker for power plant pollution, and that
contribution to mortality.
It depends on the health outcome you are looking at and
where you are.
Senator Voinovich. The particulate matter we are talking
about, you know, what is the biggest source of particulate
matter?
Dr. Thurston. Well, there are different types, and I think
what we are wrestling with is: Which effects are caused by
which types? But I think it is safe to say if you lower the
fine particle levels, you are going to better protect health.
So the standard is necessary. It is appropriate. As we learn
more, we may be able to refine that further.
Senator Voinovich. Just for example, we are trying to make
things better, and we are going to set a standard. Well,
somebody is going to have to meet a standard, but if you were
going to meet the standard, do you do that by reducing diesel
emissions? Or what is it that is going to have the biggest
impact on, if you assume what you are saying, that fine
particle matters are bad and if you set it at 12, everybody's
going to live longer than if you have it at 15, what is it that
you have to do in order to make sure that you don't expose
people to these things?
Dr. Thurston. Well, I think that is going to vary depending
on where you are. If you are in Los Angeles, I think traffic
would be a good thing to try, the pollution from traffic. If
you are in the Northeastern United States, power plants are a
big problem there, but not so much in Los Angeles.
So I think that the States will implement controls on the
pollution problems that are the largest. I think as we go on, I
mean, this is going to take us 20 years before we are finished
with this process, in reality, of meeting the standard that is
set. So as we go through it in that time, we should be able to
provide the kind of research that HEI is funding, that EPA is
funding. In 5 years, I think we will be able to give you a
better answer. As the States move to the implementation phase,
they will have that information about the sources.
Senator Voinovich. But the answer that they come up with is
that at 12 it would have a major impact on the economy of this
country and the environment in various places. There is just no
question about it.
Dr. Thurston. But it would have major benefits, health
benefits of lowering it to 12. I agree with that. The past has
shown that environmental controls are not a big problem
economically. We have been able to. We have other problems with
economics, as you are well aware.
Senator Voinovich. Dr. Thurston, let me tell you something.
I am a former Governor of Ohio, OK. The first thing I did was
try to get my counties into compliance with the current ambient
air standards because my businesses told me they weren't going
to expand, and we knew from an economic development point of
view that those counties that weren't in attainment wouldn't
even get smelled by businesses that we were trying to bring to
our State.
Today, with the new rule, we have a number of counties
throughout the United States that are not compliant with
particulate matter. They have this stamp on them: don't expand,
cost you more, don't go there, cost you more. If we reduce the
standard from 15 to 12, and there are going to be a whole lot
more of those throughout the country, it will have an impact on
the economy of those respective communities.
Now, that is not supposed to even be paid attention to, so
they can't do it, but it seems to me that we have a lot of
speculation going on whether it should be 15 or 12 or 14, that
we ought to be pretty sure, pretty certain about some of the
conclusions that we come to.
I have a problem, that I am just wondering about, you know,
the budget, the monitoring budget, the $17 million. Senator
Carper and I have been trying to get them to have an Assistant
Administrator for Science in the EPA. Does the EPA have the
brains and the people to get the job done?
Dr. Thurston. I can't answer that one. But you know, I
think the last time we spoke, we both agreed that we as a
country need to set a standard for the rest of the world, and
we need to develop the technology to meet these standards and
then sell it to the rest of the world.
Look at China. The pollution problems there are huge. That
is a huge market for the United States. If we develop the
control technology here first, if we confront this problem
first, we can then turn lemons into lemonade, if you want to
view it that way. We can be first to clean the air, and sell
that to the rest of the world. We agreed on that, as I recall,
the last time. I think it is time we got on with that process.
Now, to get back to the issues, this education question.
Being one of the authors of this, we have looked into this. I
agree with Mr. Greenbaum on this. I just would like to point
out, one of the things we have looked at is the migration
question, and that is that people of lower income tend to not
move out of their MSA. They are not highly mobile. We looked at
the data and they tend to stay exactly where they are year
after year after year.
Whereas higher educated people are much more mobile, and
that makes it so that the exposure estimates that we are using
in our studies are less reliable and therefore, some of them
move to higher exposure areas, some move to lower, so that we
are going to have greater difficulty finding an association in
those places.
That leads us to believe that the estimates of the lower
income people are probably more accurate and yes, the higher
income people, the higher educated people are having the
effects, but they don't show up in a study because the studies'
estimates of their exposures are not as good as for the less
educated people.
So I think what that says is that what we have reported is
the average of all three. Probably the lower income people are
giving us the most accurate estimates, which means they are
much higher than we thought, and that is part of the equation
why, in the expert elicitation, the various experts, a dozen
experts, have come up with estimates much higher than the ACS
study because they have adjusted for this factor and others
that have come.
So when you resolve uncertainties, we are not going to
lower estimates of the pollution. We are resolving
uncertainties and figuring out what is going on, and actually,
the impacts are going up in terms of the long-term exposure
which is, you know, that is what most of the deaths are
associated with, the long-term exposure to this pollutant,
which is pervasive year after year, day after day, for every
American.
Senator Voinovich. Thank you.
Senator Carper, I didn't think you were coming back. I was
taking everybody's questioning time.
I think in fairness to Dr. McClellan, we have less than a
minute, because Senator Carper is here.
Dr. McClellan. I would like to respond to your question in
terms of lowering the annual standard.
Senator Voinovich. The question was reasonable, would it be
reasonable?
Dr. McClellan. Yes. I am on record as stating that I think
that the staff paper laid out the science. It had a range of 12
to 15 mg/m3. I felt that was the appropriate
expression of the science, and, thus, the Administrator's
policy judgment to go with 15 mg/m3 I thought was
appropriate.
I did not think, as a member of CASAC, that it was
appropriate to narrow the range of 12 to 15 down to 13 to 14.
In fact, I thought that was just one step away from saying,
this is the standard that should be set. I thought that
overstepped the bounds.
I would like to use one of Dr. Thurston's graphs, if I
could, to point out some of the difficulties here. This is the
Harvard Six Cities study, started back in 1979 which I view as
a platnium study. The late Dr. Ben Ferris was the guiding light
that pulled together people in teams of six different
communities. One of those happens to be in your State,
Steubenville. OH. It started as an extensive study I think of
about 8,000 people over time.
Now, here are the results from this. I think it is
important to look at the graph. Across the bottom we have
PM2.5 micrograms per cubic meter, zero to 30. Now,
we have a single number there on that. Now, you will note that
each of the cities has changed over time. Steubenville was
originally up close to 30 and now it is down just above 20. You
can't express in this graph what it was in 1950? What was it in
1940? Most of the deaths that are expressed in here are of the
elderly. We know that. In the United States, people live long
lives. So in any expression of death, they are going to be the
individuals being counted.
What we are not able to show on this graph is what these
people were exposed to in those earlier years before we had the
monitoring? We know that that was substantially higher, so we
haven't captured that value in this data. What I am willing to
say, looking at long-term data, is those values must have been
substantially higher.
Now, we look over here at the mortality relative risk of
mortality related to PM exposive. I want to emphasize, we
sometimes talk about excess deaths. It is important to
recognize that when we conduct epidemiological studies we do
not create extra deaths. What we do is try to take the deaths
that have occurred and tease out what they may have been
associated with.
In terms of cardiopulmonary deaths, we know the biggest
risk factor is cigarette smoking. If you go to the data of
Engstrom, out of LA, his relative risk factors for the all
cause mortality for smokers, about a pack a day, was 2.0. That
would be way the heck up there on this graph. A value of 1.0
means there are no deaths assignable to PM.
So one of the things we do know if we look in terms of the
baseline for this relative risk, death rates for
cardiopulmonary deaths have been going down. For lung cancer,
we have seen for the first time a decrease in lung cancer
mortality. Why, because people quit smoking?
So the data that Dr. Smith related to, I suspect, if I were
able to examine her calculations, we would find that in all of
the eight cities, there were improvements, and so the base for
that relative risk factor is going down resulting in fewer
calculated deaths related to PM in the air.
What I am saying is when you look at all of this data, I
think the Administrator was fully appropriate in his policy
judgment to continue at 15 micrograms per cubic meter, and I
think when that is ultimately met across the country, we will
have I think some small improvement in total health.
Senator Voinovich. I thank you very much for that. I will
turn it over to Senator Carper. I just couldn't help, and I
think you will understand this, I will never forget campaigning
for Governor in 1990 in Steubenville, OH. I ran into a woman
who was an immigrant, and we were talking about the air, and
she said, ``Mr. Voinovich, I put the sheets out in the old days
and there would be black stuff on it. Now I put it out, no
black stuff. The sheets are clean, but nobody is working.''
And that was it. We didn't get into the health or anything
else, but the fact of the matter was that things were better
for her environment, but the thing that she was concerned about
is that nobody was working and wanted me to do something about
bringing jobs to Steubenville, OH.
Senator Carper. That was actually a pretty good segue,
because I have just come from a hearing in the Banking
Committee which is going on downstairs. Every 6 months, the
Chairman of the Federal Reserve comes and testifies to brief
the House and the Senate on monetary policy. He was just
sharing with us how strong our GDP growth is, 5.6 percent I
guess for the last quarter, how many new jobs are created thus
far this year, and large growth in Federal revenues and that
sort of thing.
But the progress, the economic growth is uneven across the
country. Frankly, the growth in incomes is uneven across our
citizenry, as we know.
I apologize for missing your testimony. Whenever the
Chairman of the Federal Reserve comes and testifies, he doesn't
do it often, I try to be there. I like to kid and say he won't
go forward until I am there, but that is not really the case.
But I wanted to be there to be able to offer a statement or two
and maybe ask a question of him.
I appreciate the fact that you are here. We appreciate your
testimony, your input, and your willingness to respond to our
questions. I have a couple of specific questions of Dr.
Thurston. Before I do, I have sort of an unwritten rule that
whenever one witness uses the charts of another witness to make
a point, we always give the owner of the charts the opportunity
to have the last word. Do you have any comment or any response
you want to make to Dr. McClellan's comments?
Dr. Thurston. Yes, I do. I thank you for that opportunity.
Well, you know, I think that his reference to, well, first
of all, his references to cigarette smoking, these studies have
been controlled for cigarette smoking.
Senator Carper. Say that again?
Dr. Thurston. Have been controlled for, this study that is
represented here has been controlled for cigarette smoking.
Senator Carper. OK.
Dr. Thurston. OK. So these reductions in relative risk have
nothing at all to do with cigarette smoking and the fact that
people are smoking less. This has to do with the reductions in
air pollution after controlling for all these other factors
that they studied. This is a cohort study where they had
individual information about each of the people in the study.
The other thing he is talking about, well, exposures long
ago that could be responsible for these effects. There is
recent research that indicates that it is really the exposure
in the last 5, at most 10 years of your life, or up until the
time of death that are the most important. Even if you look at
cigarette smoking, when someone quits cigarette smoking, the
Surgeon General's report points out, within 10 years their
mortality risk is the same as the general public. It is similar
with air pollution. The benefits are yielded very quickly,
actually, after the exposure is reduced. We see that in this
study and we see it in others.
We are looking at the ACS study, looking at these windows
of exposures, and we will be publishing on that as well. But I
know there are papers out there, like Kunsley's recent paper
pointing this out. So that is really my response to his
comments.
Senator Carper. Good. Well, thank for sharing that.
Perhaps a couple of specific questions of Dr. Thurston. Dr.
Smith and Mr. Greenbaum, I missed your testimony, as you know.
What I am going to ask you to do is, take no more than a
minute, but just give me my takeaways. It is impossible for us
to remember everything you said or say, but just, if we
remember nothing else, of a couple of key points that you made,
what should we take away from here?
Dr. Smith. Just a minute.
I think the key takeaway is that if you want to understand
whether to tighten the standard, rather than whether we need
the standard that we have, then you need to look to the
question of what is happening to the margin of safety that was
deemed acceptable when that standard was set, and using the new
evidence that we have today is that the margin of safety that
is associated with the current standard has, if anything,
grown, not narrowed.
So while there is better information across the board,
there does seem to be confirmation that there is some subtle
effect going on with air pollution, that was assumed at the
time the standard was set, and the standard at that time was
set so that it provided a margin of safety, and that margin of
safety remains today even in the face of all of this new
evidence.
Senator Carper. Thank you.
Mr. Greenbaum.
Mr. Greenbaum. Sure. Just three things. First, we have made
tremendous progress since 1997 in understanding a whole series
of questions we had then. In fact, they are not even on the
table now in the same way because of research that was
chartered by Congress and then put into place over the last 6
to 7 years to answer those questions.
Second, one part of that research has been to look at the
so-called concentration response, what happens at higher and
lower levels of pollution? Do we see more effects, less
effects? Is there a point below which we don't see any effects?
And there, we have seen, 0 with some uncertainty at both ends
of the range, generally increasing effects with increasing
levels of pollution, a generally linear relationship between
those two.
At the same time, and the third point, looking forward,
there are still, as science always has, important continuing
questions, probably most importantly, the one around
understanding whether there is a different toxicity of
different components of PM, really different sources of PM, to
guide future regulatory decisions.
Senator Carper. Good, thanks. My thanks to both of you.
Dr. Thurston, a couple of questions, if I could. Do you
believe that there is certain, I would underline the word
certain, scientific evidence to justify lowering the annual
standard?
Dr. Thurston. Yes, I certainly do. The bulk of the
evidence, as EPA has presented it, shows that we need to
tighten these standards, that there are health effects below
the present standards and that they need to be tightened.
Senator Carper. OK. I think, and I didn't hear this issue,
but Dr. Smith I believe you may have suggested that the risk of
fine particles has actually decreased in your testimony. I
would ask of Dr. Thurston if you want to comment on that. I
don't want to mischaracterize what she said, but if I have that
right, would you just comment on it?
Dr. Thurston. Well, as I said in my testimony, I believe
that, and I think the evidence supports, that the uncertainties
associated with our estimates have actually decreased. We are
more certain of the effects, the biological mechanism, and the
size of the effects. So uncertainty has been diminished.
I want to clarify the distinction between, I guess,
uncertainty and doubt. There is no doubt that air pollution is
causing these effects, and there is no doubt that lowering
these pollution levels will reap huge health benefits. There is
uncertainty exactly how much benefit for how much pollution. So
that is uncertainty of the estimate. It could be higher. It
could be lower. In the case of the long term, the scientific
consensus is moving that the health effects of PM2.5
long-term exposures are actually higher than we previously
thought.
So there is uncertainty around the estimates, but there is
really no doubt here.
Senator Carper. All right. My thanks to each of you. I am
glad you were still here when I got back. Again, I appreciate
very much your testimony and responding to our questions. Thank
you.
Senator Voinovich. Thank you much. I really believe that I
have gotten some more clarity here, at least in terms of
people's differences in terms of this standard. Hopefully, this
will help the Administrator make the right decision.
Thank you very, very much. The hearing is adjourned.
[Whereupon, at 11:30 a.m. the committee was adjourned, to
reconvene at the call of the Chair.]
[Additional statements submitted for the record follows.]
Statement of Hon. Joseph Lieberman, U.S. Senator from
the State of Connecticut
Thank you, Mr. Chairman.
In my view, EPA's most fundamental responsibility under the Clean
Air Act is to tell Americans truthfully whether the concentrations of
pollution in the air they breathe are at levels that endanger their
health. If EPA knows that particulate matter hurts people at lower
concentrations than those reflected in the agency's existing air-
quality standards, then I believe the agency has a legal and a moral
responsibility to tighten the standards.
Following an exhaustive review of peer-reviewed studies on the
subject, the scientists, doctors, and public servants on the
congressionally chartered Clean Air Science Advisory Committee have
told EPA that air-borne particulate matter is triggering large numbers
of asthma attacks, heart attacks, and premature deaths in many areas of
the country that meet EPA's existing air quality standards for
particulate matter. That leads me directly to the conclusion that EPA
must make those standards more stringent. An alternative risk analysis
requested by the White House Office of Information and Regulatory
Affairs has just come to the same conclusion.
Unfortunately, EPA has proposed to not lower the annual particulate
standard at all, and to lower the daily standard to a level that
remains above the limit that the Science Advisory Committee has
identified as necessary to protect public health.
I do not believe that EPA can justify this disregard for the
Science Advisory Committee's recommendations. Skeptics enjoyed
representation on the Committee and ample opportunity to press their
views. To me, it makes no sense to disregard the Committee's
conclusions based on complaints that were not sufficiently compelling
to convince that expert body. Here I note that only two of the twenty-
two members of the Committee's panel on particulate matter dissented
from the panel's conclusions, and that all seven members of the
committee agreed with the panel's majority.
Having reviewed the statements and testimony delivered at last
week's hearing, it appears to me that the impetus behind the calls for
EPA to disregard the Science Advisory Committee's recommendations is
not dissatisfaction with the scientific work of the Committee, but
rather concern over the cost of bringing air quality into line with
more stringent standards.
Clearly, achieving further reductions in particulate-forming
emissions will cost money. That is why the Clean Air Act's system for
implementing the health-based air-quality standards includes, at nearly
every turn, generous regard for what is practicable and what is not.
There is no need, then, to flout the Act by infecting the standard-
setting process with considerations of implementation costs. I would
point out, moreover, that by any reasonable measure, the economic
benefit of the lives saved and illnesses averted by bringing
particulate levels down to the levels recommended by the Science
Advisory Committee would vastly outweigh the economic costs of the
added pollution controls.
The concentrations of particulate matter persisting in many parts
of the country cause more than 45,000 premature deaths every year. The
problem is too grave and too large to be concealed. We can solve this
problem, and the first step is to level with the American people. That
is why I urge EPA to set the revised particulate matter standards at
the levels that the Science Advisory Committee has determined necessary
to protect public health. Thank you, Mr. Chairman.
__________
Statement of George Gray, Assistant Administrator for Research and
Development, U.S. Environmental Protection Agency
Good morning, my name is George Gray, and I am the Assistant
Administrator for the Office of Research and Development in EPA. I wish
to thank you for the opportunity to discuss the EPA's evaluation of the
scientific evidence for potential health effects of airborne
particulate matter (PM). Last week the Subcommittee on Clean Air,
Climate Change and Nuclear Safety heard from my colleague, William L.
Wehrum, the Acting Assistant Administrator for Air and Radiation, on
the EPA's review of the national ambient air quality standards (NAAQS)
for PM. During that testimony he explained the crucial role of science
in helping to inform decisions about the National Ambient Air Quality
Standards.
Today, I would like to talk with you in more detail about this
science: how it is prioritized and developed, and how it is synthesized
and integrated. As Assistant Administrator for EPA's Office of Research
and Development (ORD), I am responsible both for the development of new
scientific information targeted to address critical Agency research
needs and for the evaluation, synthesis, and integration of the world's
peer-reviewed science literature into a document that informs EPA
decision-makers. As we characterize the current state of our collective
scientific knowledge, we are careful to point out the strengths and
weaknesses of this large body of information, so that informed
decisions can be made. It is clear that the scientists and staff of ORD
play a crucial role in the development and evaluation of the world's
scientific information to inform the review of National Ambient Air
Quality Standards.
We all agree that environmental protection efforts must be based on
high quality science. High quality science includes both the conduct of
research--in the laboratory and in the field--and the careful
evaluation of that body of research to inform policy making. High
quality research is focused appropriately on generating new knowledge
that addresses complex scientific issues and helps reduce important
scientific uncertainties. It is carefully planned, well conducted, and
thoroughly peer reviewed by independent scientific experts. The careful
and balanced characterization of the body of knowledge created by high
quality science requires an open process, interaction with appropriate
subject matter experts, and serious consideration of the ways in which
the results are communicated to decision makers. To me, an important
component of high quality science is the characterization of the
uncertainties related to individual studies and, more generally, the
characterization of the weight of the scientific evidence.
First, let me discuss EPA's efforts to develop new and relevant
science on particulate matter. The Agency has a longstanding and strong
program to develop and use new scientific knowledge on the health
effects of airborne PM. After the last review of the PM NAAQS in 1997,
EPA embarked on a very ambitious research effort to advance our
knowledge and address important uncertainties in the science related to
PM. Congress requested that we sponsor the National Academies of
Science (NAS) to provide us advice. The NAS Committee on Research
Priorities for Airborne Particulate Matter in the National Research
Council completed four reports, published between 1998 and 2004, which
provided the scientific basis EPA used to target its resources to
address the highest priority PM research needs. These needs are being
addressed by the Agency's particulate matter research program, with
more than $500 million during the past 10 years committed by EPA in
support of the highest priority research topics identified by the NAS.
These funds have supported numerous research efforts by EPA's
intramural laboratories, as well as extramural researchers funded
through our competitively awarded Science to Achieve Results (STAR)
program, our PM Research Centers, and interagency agreements with other
federal agencies. EPA also coordinates closely with other federal
agencies on PM research through the Committee on Environment and
Natural Resources (CENR) Air Quality Research Subcommittee and its
Interagency Working Group on Particulate Matter.
We learn about the potential health effects of PM through several
different types of research, especially epidemiology and toxicology.
Guided by the NAS and other advisors, the Agency has funded research in
all of these areas. Epidemiologic studies supported by EPA and others
provide key information in our evaluation of PM. This research includes
population-based studies that evaluate potential associations between
human exposure to PM and health outcomes, including death,
hospitalization, illness, and potential precursors to illness. We have
sponsored research on populations of tens to hundreds of thousands of
individuals in the United States that evaluates the effects of long-
term exposure to PM on illness and death. These include both cohort
studies and panel studies. Other research uses a different design--
called time-series studies--in which air pollution levels are tracked
on a day-to-day basis and compared with daily variations in health
statistics to evaluate the effects of short-term exposures to PM on
health. These time-series studies included hundreds of communities and
databases that describe millions of residents. Other epidemiologic
studies attempt to identify factors affecting people's susceptibility
and the role of co-pollutant exposures.
Toxicology studies, sponsored by EPA and others, provide both
information to evaluate the strength and plausibility of the
associations identified through epidemiology and hypotheses that form
the basis of new epidemiological studies. Important studies include
those that evaluate the components of PM that may be producing
toxicity, and the mechanisms by which such toxicity might occur.
These research efforts have resulted in literally thousands of
published studies in the peer-reviewed literature over the past several
years. In 2005, EPA prepared a report, Particulate Matter Research
Program: Five Years of Progress, which highlighted the early results of
EPA's substantial investment in PM. When it came time to prepare the
science basis for the next evaluation of the PM standards (the 2004 Air
Quality Criteria Document), more than 4000 articles from the peer-
reviewed literature were reviewed--many of which came from research EPA
had done in our laboratories or had funded through our STAR grants.
A second, and equally important, function of EPA efforts is the
synthesis and integration of these thousands of individual ``acts of
science'' to provide a clear characterization of our knowledge and the
degree to which we still are uncertain about aspects of PM health and
environmental effects. We have a scientifically rigorous process by
which we evaluate and interpret this important body of knowledge and
ensure that our interpretation of them is complete, transparent,
unbiased, and consistent with the array of views in the scientific
community. A fundamental step in the review of the National Ambient Air
Quality Standards is the evaluation of scientific evidence and the
preparation of scientific assessments, by the National Center for
Environmental Assessment of the Office of Research and Development,
known as ``criteria documents.'' The development of criteria documents
involves the review of thousands of peer-reviewed research
publications, evaluation of those studies most relevant to the review
of the air quality standards, and integration of the scientific
evidence across disciplines. The body of evidence must be reviewed,
evaluated, weighed and then accurately and objectively described to
inform our decisions about National Ambient Air Quality Standards.
For the current PM review, EPA evaluated research studies that
addressed a wide range of issues including PM toxicology, epidemiology,
atmospheric chemistry, human exposure, and other areas such as
environmental effects. Thousands of studies were reviewed and over 2000
studies were referenced in the criteria document, many of which were
conducted or funded by EPA's Office of Research and Development.
Considered together, these new studies significantly advanced our
understanding of PM's potential effects on public health and welfare
and reduced the uncertainty associated with some important aspects of
the science. Drawing on the evaluation of studies reviewed in the PM
criteria document about health effects and dose-response, as well as
information about exposures to PM, EPA also completed a risk assessment
to estimate the degree to which various approaches for revising the
standards would potentially affect the public health risks posed by PM.
Further, the Agency prepared a document known as a ``staff paper'' that
utilized the evaluation and characterization of scientific evidence in
the criteria document together with the results of the risk assessment
to help inform the policy judgments required in making decisions on the
NAAQS.
Extensive independent external peer review was conducted on the
criteria document, risk assessment, and staff paper by the Clean Air
Scientific Advisory Committee (CASAC). CASAC, statutorily-mandated
under the Clean Air Act, is a group of independent scientific and
technical experts appointed by the Administrator to review EPA's
evaluation and use of scientific and technical information related to
air quality and make recommendations as appropriate. CASAC is made up
of nationally-recognized scientists from a variety of relevant
disciplines. For PM, CASAC was extensively involved in reviewing and
commenting on several drafts of the PM criteria document, staff paper,
and risk assessment. Their efforts, and those by EPA staff to address
CASAC's comments, resulted in a PM science assessment that provides
comprehensive, relevant information suitable to serve as the scientific
basis for Administrator Johnson's decisions on the PM NAAQS.
Let me briefly highlight some scientific information available on
particulate matter. First, as a scientist, I know that all scientific
research includes aspects of uncertainty. For example, we often do not
understand the mechanisms by which pollutants such as particulate
matter produce health effects in the population. We know our
measurements of environmental conditions and biological response
contain some uncertainty due both to our understanding and
technological limits. To have uncertainty is normal. Uncertainty is a
factor to be characterized and considered in the evaluation of studies
and other data. We always consider the strengths and limitations of the
available evidence when drawing conclusions about what that evidence
means for decision making.
For example, we highlighted the uncertainty in the evidence linking
chronic exposure to PM2.5 with premature mortality in the
1997 review of the PM NAAQS. In the next few years, EPA responded by
funding a major reanalysis by independent investigators of two
important long-term studies that used data from a Harvard Six Cities
cohort and an American Cancer Society cohort. The quality of the data
was evaluated, and an extensive series of sensitivity analyses were
performed using various statistical models to test for the influence of
many potential co-variables. The results duplicated the association
between levels of chronic exposure to PM2.5 and premature
mortality. These analyses were important in reducing our uncertainty
about the consideration of these data in the standard-setting process.
In addition, the analyses identified other avenues of research. For
example, one study indicated that the estimated effects of fine
particles appeared to vary with education level.
In another example of our efforts to tackle uncertainty, EPA
sponsored a number of multi-city epidemiologic studies designed to
address the limitations inherent in single-city studies. Multi-city
studies allow the assessment of risks of mortality or hospitalization
across cities, thus reducing uncertainty regarding the effects of local
features, such as differing mixes of pollutants and climates, on the
interpretation of study findings. The results of these multi-city
studies provide additional evidence that levels of exposure to
PM2.5 are likely to be linked with serious health effects.
Another major area of uncertainty remaining from the previous
review was the lack of demonstrated biological mechanisms or pathways
by which PM exposure could result in the effects observed in
population-based studies. An important factor in evaluating the
associations uncovered in epidemiologic investigations is biological
plausibility, i.e., whether there is a coherent way in which the
reported association could be expected to occur in the body. As noted
in our 2005 report, EPA-funded research has provided crucial insights
into numerous hypothesized mechanisms; including evidence that exposure
to particles may contribute to atherosclerosis development and affect
cardiac rhythm, thus linking the findings of mortality in the
epidemiologic studies to plausible biological mechanisms of toxicity.
Looking across the large landscape of study findings, our
assessment of the research results for particulate matter finds
evidence of a coherence of health effects associated with
PM2.5 across many types of study designs, biological
endpoints and time frames. The body of evidence--the thousands of
studies from a wide variety of disciplines we have evaluated with the
help of CASAC--demonstrates that PM2.5 exposure is likely
causally associated with outcomes such as cardiovascular and
respiratory morbidity and premature mortality from both epidemiologic
and toxicology studies. Toxicology studies help us understand the
mechanisms that provide some evidence of biological plausibility in the
observations from epidemiological studies. We recognize that
uncertainty exists, but uncertainty is not a barrier to decision-
making; rather it is critical information to be factored into informed
decisions.
We also recognize that science is not static. New studies on PM are
being published in the peer-reviewed literature all the time. As a
continuation of the scientific review process, EPA recently conducted a
survey of the evidence reported in the scientific literature since
completion of the literature review reflected in the 2004 criteria
document. This new survey includes some 700 additional studies and has
emphasized the studies most relevant to the PM NAAQS decision. The
provisional assessment of these new studies has only just been
completed. To provide the public with an opportunity to review the
survey results, we will provide notice of the completion of this survey
and post the results on our Web sites. In brief, the provisional
assessment concluded that taken in context, the new information and
findings provide additional support regarding the health effects of PM
exposure made in the 2004 PM Air Quality Criteria Document but do not
materially change any of the broad scientific conclusions.
In summary, the Bush Administration is committed to the development
and use of the highest quality science to inform environmental decision
making. The mission of the Office of Research and Development is to
develop, evaluate, and communicate relevant scientific information to
the Administrator, and to assure that the Administrator is well
informed of the nature, strengths, and limitations of this information.
EPA has sponsored a targeted and effective research program on
particulate matter and I am pleased to convey to you and others the
value of this investment. We have made a great effort to evaluate and
characterize the existing and new scientific results available on
particulate matter, and I am personally pleased to share with you my
views on this work. I look forward to addressing any questions you may
have.
Thank you.
______
Responses by George Gray to an Additional Question from
Senator Inhofe
Question 1. The General Accountability Office found that EPA has
fully implemented only 8 of the National Academy of Sciences' 34
recommendations in conducting its PM air quality review. One of its
recommendations was to include a cal culation of the total costs and
benefits, including indirect benefits, in its regulatory impact
analysis. It is well established that economic growth is associated
with health, and conversely, economic costs and associated job losses
is associated with increased mortality. In light of this, it is only
possible to determine the net health benefit if these indirect costs
are calculated. Does EPA intend to in the final RIA calculate the net
benefits, including indirect disbenefits?
Response. EPA believes there are indirect effects that should be
considered when the measured costs and benefits are sufficiently large.
However, quantifying these issues is very difficult. Until better
methods are developed to assess these second-order effects, EPA has
focused on assessing the primary benefits and costs that result from a
partial equilibrium analysis.
EPA uses a partial equilibrium analysis when assessing the benefits
and costs of regulatory options. Partial equilibrium analysis, by
definition, only looks at the direct effects of the regulation. It
ignores the general equilibrium effects that may occur throughout the
economy. Note that the partial equilibrium models ignore both the
secondary benefits and the secondary costs that may occur. Partial
equilibrium models are generally used when the costs and benefit
impacts are small so that secondary impacts may be safely ignored.
However, if the regulation has sufficient impacts on benefits and
costs, these secondary costs may be significant. There are at least two
effects that must be measured to do this correctly:
The health-wealth tradeoff with costs: This literature suggests
that costs of environmental regulation may be higher than the direct
expenditures of a regulation. Specifically, raising costs also raises
prices and reduces real wage income. This lower real wage reduces net
income and may, therefore, reduce the overall health of workers. (The
lower our income, the less health we can afford.) This is the effect
mentioned in the question.
The health-wealth tradeoff with benefits: When our environment
improves, we reduce the sick loss days of workers, and improve the
overall health of the population. These changes increase labor
productivity, raising the real wage. As the real wage increases, the
populations' health increases. (As our income increases, we can afford
to buy better health care, etc.)
In general, EPA assumes that these two effects roughly cancel. We
have taken some exploratory looks at this issue and found this to be
the case, though generalized conclusions cannot be made. These issues
are also discussed in the context of EPA's analysis done under section
812 of the Clean Air Act amendments.
In addition to the above issues, it is our understanding that the
NAS recommendation concerning indirect benefits focuses on whether any
particular EPA analysis can and should try to capture potential
indirect effects pertaining to changes in exposures. For example, the
NAS gave as its first illustrative example of the indirect effects
concept the possibility that air pollution regulations may change how
fuels are made or how combustion devices are operated, which could in
turn affect human health through other pathways.
______
Responses by George Gray to Additional Questions from
Senator Voinovich&
Question 1. Dr. Gray, please put yourself in our position as the
oversight Committee or the publics position whether a regulated
industry, environmental group, or just a concerned citizen. Last week,
we discussed at length the fact that EPA has put together an incomplete
Regulatory Impact Analysis. Basically, we do not know the impact of
revising the standards. Let me remind EPA that this information is
required by Section 108 of the Clean Air Act: Administrator shall. . .
issue to the States. . . information on air pollution control
techniques, which information shall include data relating to the cost
of installation and operation, energy requirements, emission reduction
benefits, and environmental impact of the emission control technology.
We are told that a more complete analysis will be addressed in the
final rule.
Today, we learn from GAO that EPA has implemented less than 25
percent of the National Academies recommendations. We are told more of
these recommendations will be addressed in the final rule.
Chairman Inhofe has pushed EPA to consider the numerous new studies
that have been completed since the criteria document was finished. We
are told that this new science will be addressed in the final rule.
Don't we have an oversight responsibility and doesn't the public
have a right to know what EPA is doing behind closed doors on this
rule?
Response. EPA is committed to working with Members and committee
staff to ensure the public is well-informed with regard to Agency
rulemakings. EPA has been, and will continue to be, in regular contact
with the committee regarding the PM2.5 NAAQS and the
associated RIA, and we believe our final analysis will be
comprehensive, useful, and robust.
The Clean Air Act (Act) requires EPA to set a primary standard for
each criteria pollutant that protects public health with an adequate
margin of safety. As interpreted by the Agency and the courts, the Act
requires EPA to base this decision on health considerations and forbids
consideration of economic factors. Thus, while the final RIA can
provide important insights into the costs and human health benefits
associated with attaining a revised PM2.5 NAAQS, EPA cannot
use estimates of attainment cost to inform the NAAQS decision.
This prohibition against the consideration of cost in the setting
of the primary air quality standard, however, does not mean that costs
or other economic considerations are unimportant or should be ignored.
The Agency believes that consideration of cost is an essential
decision-making tool for the cost-effective implementation of these
standards. Under the Clean Air Act, the impact of cost and efficiency
are considered by the States during this process, when States are
making decisions regarding what timelines, strategies, and policies
make the most sense. The PM NAAQS final RIA is intended to inform the
public about the potential costs and benefits that may result when any
revised PM standards are implemented.
With regard to the new studies cited in your question, EPA's Office
of Research and Development has conducted a survey and provisional
assessment of relevant scientific information that became available
since the completion of the 2004 Criteria Document, as discussed in
EPA's notice of proposed rulemaking on the PM NAAQS. In summary, our
provisional assessment concluded that the new studies do not materially
change the broad conclusions of the 2004 document.
Question 2. Dr. Gray, Congress established the Clean Air Act to
protect human health. Last week:
Georgia Commissioner Heiskell stated: As an elected official, I
fear the lost tax revenues and increased stresses on local health
services that lay-offs associated with our non-attainment status bring.
National Black Chamber of Commerce President Harry Alford stated: .
. . the biggest health risk to African Americans anywhere. . . is
poverty.
Now, I understand that EPA cannot consider costs. However,
according to the GAO report:
Another recommendation (from NAS) that EPA. . . did not apply to
the draft regulatory impact analysis concerns whether the proposed
revisions to the particulate matter standards would have important
indirect impacts on human health and the environment.
This is a health based standard and there are indirect, negative
impacts on health from nonattainment designations that reduce jobs and
increase energy prices. Why isn't EPA following the NAS recommendations
and considering these impacts?
Response. Please see our answer to Question 1.
Question 3. Dr. Gray, does science dictate a particular number for
the air quality standards?
Response. Uncertainty is inherent in science. As new data become
available, our understanding of the health protectiveness of a given
air quality standard grows, as does our understanding of the extent of
uncertainties. That is why rarely, if ever, it is the case that a body
of scientific evidence has the potential to dictate a particular number
for an air quality standard. Science does provide the basic information
used in setting air quality standards, and our analysis of the science
includes consideration of the strengths and uncertainties in the
available evidence. We know that our measurements of environmental
conditions and biological responses contain some uncertainty due both
to the limits of our understanding of the underlying process and
available technology. EPA recognizes the importance of characterizing
the implications of this scientific uncertainty.
In the review of the PM air quality standards, the scientific
evidence has provided strong support that PM exposures can have adverse
effects on public health and the environment. The science summarized in
EPA's Criteria Document and used in EPA's risk assessment provided the
scientific basis for the range of options recommended in EPA's staff
paper. There is uncertainty even within this range of options. The
strength of the scientific evidence as well as the uncertainties are
necessarily considered by the EPA Administrator in proposing and
promulgating final PM NAAQS.
Question 4. Dr. Gray, as I stated in my opening statement, I do not
understand how EPA can revise the particulate matter standards when the
public health benefit is not fully understood. For example, there are
areas of the country, namely Southern California, that will never be
able to meet the Agency's proposal, not to mention even tighter
standards. When EPA estimates health benefits, how do you take into
account that this area will never attain the standards? Basically, does
the Agency estimate real benefits?
Response. This Nation has an excellent record of rising to the
challenge of reducing air pollution in a cost effective fashion. All
areas have made significant progress in reducing air pollution even if
some areas are still not yet in attainment of the current standards.
The regulatory impact analysis that will accompany EPA's decision
reflects the need to develop new ways and ideas for some areas to
attain the proposed standard. To the extent that known controls are not
sufficient to bring an area like Southern California into attainment by
2020, EPA has relied on information from a variety of sources on the
effectiveness of new control strategies under development to project
what the costs and benefits of attainment might be. While these
estimates have additional uncertainty associated with them, they give
the public an idea of the likely magnitude of the potential costs and
benefits of attainment.
Because of its severe air pollution problems, Southern California
has been one of the leaders in the development of creative ways to
reduce air pollution by harnessing the free market. For example, the
RECLAIM (Regional Clean Air Incentives Market) program has been in
place since 1994 to harness market forces to reduce air pollution. See
http://www.aqmd.gov/reclaim/reclaim.html for more information. Also,
the California Air Resource Board (CARB) is now aggressively pursuing
options for continuing to reduce air pollution. On April 20, 2006, CARB
announced a new program for reducing air pollution from its ports which
includes economic incentive elements. See http://www.arb.ca.gov/gmp/
gmp.htm for more information.
Question 5. Dr. Gray, the GAO report states that the Agency has
implemented less than 25 percent of the National Academies'
recommendations for estimating health benefits. Does the Agency have a
plan to implement all of the recommendations?
Response. EPA is committed to showing further progress in the final
RIA, which will be published when the NAAQS is finalized this
September. As noted in the GAO report ``Particulate Matter: EPA Has
Started to Address the National Academies' Recommendations on
Estimating Health Benefits, But More Progress is Needed,'' EPA
expressed to GAO that the Agency is committed ``to further enhancing
the transparency of the analysis by presenting clear and accurate
references to the supporting technical documents, which detail the
analytical assumptions and describe the data supporting the estimates
[of our Regulatory Impact Analyses.]'' GAO found that of the 34
detailed recommendations in the NAS report, the draft RIA implemented
(in full or in part) 22 of them. Some of the remaining 12
recommendations will be addressed in the final RIA to accompany the
final rule in September; others require further research or
development. For example, over the past few years, EPA ran an expert
elicitation designed to characterize uncertainty in the estimation of
PM-related mortality resulting from both short-term and long-term
exposure. This work has recently been completed and peer reviewed. The
results from this expert elicitation are being applied to the benefits
analysis in the RIA. However, it has not been possible to address all
the NAS recommendations since the September 2002 publication of the
report. In its report and its recommendations, the NAS recognized that
additional research and resources were going to be needed to address
some of its recommendations. And, in some cases, the recommendations
were in fact conditioned on the availability of improved scientific
data. EPA is committed to addressing these recommendations, including
those on PM speciation, and improving its analyses as new research
becomes available.
In its July 19, 2006, report, GAO acknowledged the progress EPA has
made on implementing the recommendations while also noting that EPA
needs to make more progress in addressing the National Academies'
recommendations on estimating health benefits.
Question 6. Dr. Gray, how does EPA decide which studies to rely on
and which ones to ignore? Do CASAC or EPA have any written criteria for
determining which studies are more important than others? Any intention
of developing this so that the NAAQS review process is more objective
and transparent?
Response. We believe that the consideration of scientific evidence
in the NAAQS review process is comprehensive, objective, and
transparent. EPA conducts an extensive literature search to identify
potential studies for inclusion in its science assessments. For
example, EPA uses standard searches of systems such as MEDLINE in the
National Library of Medicine, independent table of content searches by
EPA staff, and input from the public and independent scientists outside
EPA. These search methods are summarized in the Criteria Document. The
selection of studies for consideration in a NAAQS review, the strengths
and uncertainties of individual studies, and the subsequent
determination of studies to be given greatest weight are outlined and
discussed in the science assessment. These selections and
determinations are evaluated by the public and by CASAC at public
meetings. The inputs from the public and CASAC are documented,
considered, and incorporated as needed.
Question 7. What role did the White House and other Federal
Agencies play in setting the NAAQS proposal?
Response. Under Executive Order No. 12866, ``Regulatory Planning
and Review,'' signed by President Clinton in September 1993, the
Administrator of the Office of Information and Regulatory Affairs
(OIRA) carries out a regulatory review process on behalf of the
President. The objectives of this Executive order are to enhance
planning and coordination with respect to both new and existing
regulations; to reaffirm the primacy of Federal agencies in the
regulatory decision-making process; to restore the integrity and
legitimacy of regulatory review and oversight; and to make the process
more accessible and open to the public. In general, involvement by
other Federal agencies or the White House offices occurs during the
routine interagency review process, when the Agency is developing its
proposal and before making a final regulatory decision. Following this
procedure, the proposal package for the PM NAAQS went through the
interagency review process, which was managed by the OIRA and included
input from various agencies and offices throughout the executive
branch.
Once EPA has issued a proposal to the NAAQS, the Agency evaluates
public comments received on the proposal and develops a draft final
rule. The draft will then be subject to the interagency review process,
per Executive Order 12866, which includes input from across the
executive branch.
__________
Statement of Daniel Greenbaum, President, Health Effects Institute
Mr. Chairman and members of the committee, thank you for the
opportunity to testify before you today on the health effects of
particulate matter (PM). I come before you as the President of the
Health Effects Institute, a non-profit, independent research institute
funded jointly and equally by the U.S. EPA and industry to provide
high-quality, impartial science on the health effects of air pollution.
For over 25 years we have conducted targeted research on a variety of
pollutants and health effects, and I am pleased to summarize our
understanding concerning PM and health for you today.
I also had the privilege to serve from 1998 until 2004, as a member
of the National-Research Council Committee on Research Priorities for
Airborne Particulate Matter, a congressionally chartered panel that
both set priorities for national PM research, and monitored the
progress in implementing those priorities by U.S. EPA and other public
and private agencies.
I would like today to briefly highlight three topics of direct
relevance to the current discussion of U.S. EPA's proposal for
revisions to the National Ambient Air Quality Standards for PM (the
``PMNAAQS''):
Science progress we have made since 1997,
The most recent findings on the relationships between different
levels of ambient PM and health effects (so-called ``concentration-
response''), and
Keyscience needs going forward.
progress since 1997
Since Congress identified the need for substantial enhanced
research on PM in the wake of the 1997 PM2.5 NAAQS decision,
established the NRC Committee, and appropriated substantial new funds
for PM research, much progress has been made in answering key questions
for the current NAAQS review process, and for future ones.
Specifically:
We know much more about the sources and transport of fine
particles, and about personal exposure to those particles, especially
for sensitive groups like the elderly and children.
We have conducted the first multi-city epidemiology studies of
effects, and analyzed and reanalyzed many of the major studies of human
effects, finding that in general the earlier studies were well done and
could be confirmed. At the same time there has been some evidence that
the population health effects we had seen in those earlier studies may
in some cases be smaller than previously thought.
Unlike in 1997, we now have numerous laboratory, animal, and
human toxicology studies that have begun to indicate potential
biological mechanisms by which PM may cause health effects, especially
new findings of effects on the heart and circulatory system. Although
we have made progress, most science observers would agree that there is
still much to learn about the mechanisms by which PM may cause these
effects.
Although there continue to be, as there always are, important
questions about PM that need further research, I think Congress, the
Federal Government, and the scientific community can take tremendous
pride in the substantial progress that has been made.
The ``Concentration-Response'' Relationship: Ambient PM Levels and
Health Effects
Among the most important questions addressed over the past few
years is the question of whether exposure to PM has been shown to have
health effects at.all levels of pollution--ie. down to zero--or whether
there is a ``threshold'' below which no effects are expected. This
question is, of course, central to deciding at what level to set a
NAAQS. There are two major types of epidemiologic studies that have
been done--of short term effects and long term effects--and I would
like to briefly review what these studies have shown us.
Short Term Effects
In 1997, there were studies of daily changes in air pollution and
health effects in a number of individual cities (so-called ``daily time
series studies''). Since then, scientists have conducted much more
rigorous multi-city studies of daily air pollution and health, most
notably the National Morbidity Mortality and Air Pollution Study
(NMMAPS) funded by HEI and led by investigators at the Johns Hopkins
Bloomberg School of Public Health. That study examined daily changes in
air pollution and health in the 90 largest U.S. cities. To answer the
question of whether there is a threshold for effects, the investigators
analyzed mortality and pollution levels across the 20 largest cities
and the 90 largest.
In brief, as shown in Figure 1, they found that there appeared to
be a linear relationship between mortality and air pollution down to
the lowest measured levels for all causes of mortality, and for deaths
from heart and lung disease, without an apparent threshold. There did
appear, however, to be a threshold for the effect of PM on ``other''
causes of mortality (e.g. non-respiratory cancer, liver disease). The
HEI Review Committee, which intensively peer reviews all HEI research,
advised ``caution in drawing conclusions from the apparent absence of a
threshold'' for all-cause and cardiopulmonary mortality, for a number
of statistical and analytic reasons. They noted however that ``the
reported associations are at ambient concentrations well below the
current U.S. daily standard . . . thus the ambient concentration level
at which any threshold might exist is likely to be correspondingly
low.''
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Long-Term Effects
In 1997, there were two principle studies of the effects of longer
term exposure on mortality, the Harvard Six Cities Study and the
American Cancer Society Study (ACS), which examined death rates among
thousands of individuals living in cities with varying levels of
pollution. Since that time, although there have been other long-term
studies published there are still very few, and much attention has
focused on HEI's Reanalysis of these two studies and on extended
analyses in the American Cancer Society Study population (which still
covers the broadest national population). In both of these efforts,
initial analyses have examined the ``concentration-response''
relationship between levels of pollution in each community and levels
of health effects.
HEIReanalysis--At the request of Congress (in the FY 1998
Appropriations Bill), U.S. EPA, and others, HEI gained access to all
underlying data in the two studies and selected an independent
investigator--Dr. Daniel Krewski--and his team to conduct a detailed
audit and reanalysis. Their work, which was also intensively peer
reviewed by the HEI Review Committee, tested the original studies
against a wide variety of alternative explanations about why people in
the most polluted cities would have higher rates of premature
mortality. In the end, the investigators and HEI's Review Committee
agreed that these alternative analyses did not change the original
findings of associations between PM and premature mortality, although
there were new findings as well about an association of mortality with
sulfur dioxide.
Dr. Krewski and his team also conducted an initial analysis of the
``concentrationresponse'' relationship between PM levels in each of the
cities and relative risks of mortality. Figure 2 presents the results,
summarizing for each community (signified by a point on the graph) the
annual air pollution level and the risk of death due to heart and, lung
disease. They then attempted to estimate the ``average'' relationship
across all of the communities (the solid line) and the range of
uncertainty around that average (the ``95 percent confidence
intervals'' indicated by the dashed lines). As you can see there is
some ``scatter'' in the data, especially at the highest and lowest PM
levels studied, but also an overall trend of increasing mortality risk
with increases in pollution levels starting at relatively low levels.
In reviewing this initial analysis, the HEI Review Committee found that
``for all-cause and cardiopulmonary mortality, the results show an
increasing effect across the entire range of fine particles or sulfate
but no clear evidence either for or against overall linearity.''
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Extended Analyses in the American Cancer society Cohort Following
the reanalysis, the original investigators for the ACS study led a
broad team of experts in an extended analysis of the data, including
additional follow-up of more recent deaths among the study population,
and using new PM2.5 data from monitors installed since 1997.
That study found results similar to those found in the Reanalysis and
also conducted analyses of the ``concentration-response'' relationship
(shown in Figure 3). This also shows a general, though less steep,
upward trend in mortality with increasing pollution levels, with the
largest uncertainty being found at the very lowest and very highest
levels where there are fewer cities. The Investigators concluded that:
``Within the range of pollution observed in this analysis, the
concentration response function appears to be monotonic and nearly
linear. This does not preclude a leveling off (or even steepening) at
much higher levels of air pollution.''
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Summary: PM-Mortality Concentration-Response
In sum, recent analyses of the relationship between ambient levels
of pollution and mortality have found a generally increasing trend of
mortality with increases in pollution across a wide range of locations.
The strongest evidence that there is not a threshold for these effects
comes from studies of short-term effects, where any threshold is likely
to be well below the current ambient standards. The initial analyses of
these relationships in long term studies also shows this general
pattern, albeit with somewhat greater uncertainty at the lowest and
highest levels.
key research needs looking forward
While we have made much progress in understanding PM exposure and
health effects over the past decade, there continue to be, as there
always are in science, important questions to be answered to help
inform future decisions about ambient air quality standards and
protecting public health. Two key areas needing continued attention
are:
Continuous Improvement in the Statistics Used in
Epidemiology
The analysis and reanalysis of studies on population health, air
pollution and weather over the last decade have enhanced our ability to
determine whether health effects can be tied to certain air pollutants.
However, those same analyses have shown that the results can be
significantly affected by the choices of statistical techniques and the
assumptions made in each analysis. Looking forward, we need to pay
continued attention to understanding the sensitivity of the results to
different assumptions, quantifying the uncertainty of the results, and
communicating clearly for each analysis both the results and the
continuing uncertainties around those results.
Systematic Analyses of which PM Components and Sources May
Contribute the Most to Toxicity
Perhaps no other question will need as much attention, and will
have as much implication for future regulations, than determining
whether some components of the complex mixture of PM are more toxic
than other components. Ultimately, this data will be essential to
ensuring that regulations and control strategies are targeted at
reducing those emissions which will have the most public health benefit
at the least cost. This has also become important in light of the
current proposal for a PMNAAQS for ``coarse particles'' which has
proposed to exclude certain particles from consideration even before
the needed studies are complete.
To date, there have been some individual city analyses of toxicity
of different components supported by U.S. EPA and others; but no
systematic national effort to compare results from across the country
and from epidemiology and toxicology studies. To fill that gap in time
to inform a next round of PMNAAQS review, HEI has launched, with
support from EPA and multiple industries, a sysatematic, multi-
disciplinary effort which will:
Conduct comprehensive, multi-city analyses of PM components and
health
Combine and compare epidemiology and toxicology across the
country, and
Provide the first-ever analyses of long-term effects of different
PM components (all studies to date have looked only at daily changes in
air pollution and health)
As indicated in both the NRC review of priorities for future PM and
health research (NRC 2004), and in today's report of the Government
Accountability Office concerning data needed for future PM benefits
analysis, these studies will be central to ensuring that future PM
actions are the most effective possible.
summary: progress made and more to be done
In sum, we have made much progress since 1997 in answering key
questions about whether and how PM can affect public health. Initial
analyses have also helped us better understand the ``concentration-
response'' relationship between levels of ambient air pollution and
health effects and the generally increasing effects with increasing
levels of pollution. At the same, looking forward there continue to be
important issues to be addressed to inform future NAAQS and regulatory
decisions, especially around the toxicity of different component and
sources of PM.
Thank you for this opportunity to testify; I would be pleased to
answer any questions you might have.
______
Responses by Daniel Greenbaum to Additional Questions from
Senator Voinovich
Question 1. Mr. Greenbaum, as you know, there is an ongoing debate
about whether EPA must lower the annual standard. Your opinion is
invaluable in this debate as an entity that sits squarely in the middle
-funded by both industry and EPA. Clearly, the statute says that
standards shall be set " in the judgment of the Administrator." In your
opinion, is it reasonable from a scientific and health perspective for
the Administrator to retain the annual standard at 15 and not lower it,
does science dictate a particular number for the air quality standards?
Response. This is of course one of the major questions facing the
Administrator as he makes final decisions on the NAAQS. As I noted in
my testimony, the Health Effects Institute has not, since its inception
over 25 years ago, taken a position advocating either changing or
retaining a specific level of a standard. We do this so that no one can
ever question our science as having been created solely for the purpose
of supporting an advocacy position. We do, however, attempt to provide
the best possible impartial interpretation of the science to help
inform such decisions. In that spirit, I would like to re-iterate and
strengthen several points I made on this question in the hearing:
(a) First, the science on the relation between different levels of
air pollution--and whether there are health effects at all levels or a
"threshold" level below which there are no effects--has improved since
1997.
(b) Second, as I noted in my testimony, these so-called
"concentration-response" relationships have shown for short-term
studies (that inform the 24-hour standard) a fairly consistent
relationship between levels of air pollution and increases in premature
mortality at levels below the current standard.
(c) Third, for longer-term studies such as the American Cancer
Society Study (that inform decisions on the annual standard), there is
also evidence of a relationship between air pollution and premature
mortality that extends below 15, but the level of uncertainty on that
relationship is somewhat greater than for the short term studies.
(d) Finally, there is no widely accepted scientific method for
setting the level of a standard. Rather, as the Clean Air Act
envisions, decisions on levels of the standard are a public health
policy judgment by the Administrator, in which he has to weigh evidence
that suggests that there are effects below the current standard against
the level of certainty or uncertainty surrounding that evidence (i.e.
the degree to which he can be certain from the science that a lower
standard would result in greater public health benefits than the
current standard.)
Question 2. Mr. Greenbaum, please explain what you mean by this
statement in your testimony: ``there has been some evidence that the
population health effects we had seen in those earlier studies may in
some cases be smaller than previously thought." Since 1997 there have
been extended reanalyses of a number of daily and long term studies of
relationships between air pollution and health.
In the case of short term (daily) studies, after HEI investigators
at Johns Hopkins discovered an issue with the software to conduct these
studies in 2002, those investigators and many others revised their
analyses using better techniques, and HEI was asked to intensively peer
review those results and draw conclusions on the new findings. In that
report HEl's Review Committee found, among other things, that for the
major multi-city daily time series studies done in the United States
and Europe--some of the most systematic and rigorous of these studies--
the estimates of risk went down between 30 percent and 50 percent
(although they were still statistically significant). For example, for
the National Morbidity, Mortality, and Air Pollution Study (NMMAPS)
conducted by HEI investigators, the estimates of effects went from 0.4
percent increase in premature mortality per 10 u/m3 of
paniculate matter to 0.2 percent per 10 u/m3. (Health
Effects Institute. 2003. Revised Analyses of Time-Series Studies of Air
Pollution and Health. Special Report. Health Effects Institute, Boston
MA.) In the case of long-term studies (of people living for many years
in more-and less-polluted cities), the effects have been smaller in
some analyses and larger in others. HEI's reanalysis of the American
Cancer Society (ACS), which audited and in general validated the
results of the original study, found lower estimates of health risk
when one includes other pollutants in the analysis. However, some
recent analyses of the ACS data by HEI investigators using improved
estimates of personal exposure have actually found higher estimates of
effect (see: Jerrett, M; Burnett, R. T.; Ma, R.; Pope, C. A., Ill;
Krewski, D.; Newbold, K. B.; Thurston, G.; Shi, Y.; Finkelstein, N.;
Calle, E. E.; Thun, M. J. (2005) Spatial analysis of air pollution and
mortality in Los Angeles. Epidemiology 16: 727-736. )
Question 3. Mr. Greenbaum, please comment on the peculiarity that
no effect is found for people with higher education?
Response. As I mentioned in response to a similar question in the
hearing, a careful reading of the full results of recent studies does
not necessarily suggest that the studies have found ``no effect'' for
people with a higher education. The HEI Reanalysis of the Harvard Six
Cities and American Cancer Society studies was the first analysis to
attempt to look at whether there were differential effects in people
with different levels of education (which is generally an indicator of
different levels of socioeconomic status). That analysis, and a
subsequent extended analysis in the ACS study, found a distinctly
higher effect on mortality for those with lower education (and likely
lower socioeconomic status). The HEI Reanalysis also found, for those
with more than a high school education, a positive but not
statistically significant association with all- cause mortality (a
relatively crude measure of mortality since it includes causes of death
which we expect could not be related to air pollution). When one looks
at more specific causes of death, one finds that even for those with
higher education, there is an association of mortality for
cardiopulmonary deaths, and especially cardiovascular deaths. (See
Table 52 in Health Effects Institute. 2000. Reanalysis of the Harvard
Six Cities Study and the American Cancer Society Study of Paniculate
Air Pollution and Mortality: A Special Report of the Institute's
Particle Epidemiology Reanalysis Project.) There are several reasons
why these differences in effects among people of different levels of
socioeconomic status might be seen. First, it is plausible that people
of lower socioeconomic status have higher exposures to air pollution
due to: living in more heavily polluted areas; the likelihood that they
have less access to air conditioning; and the jobs they have which may
involve more outdoor exposure. Second, it is also well known that
poorer people have worse levels of nutrition and poorer access to
quality medical care, which could reduce their underlying health and
make them more susceptible to the effects of air pollution. Third, it
is also possible that, although these studies have made extensive
efforts to separate the effects of ``being poor'' from the effects of
``being exposed to air pollution'', there are some results in these
studies that continue to reflect the impacts of poverty on health.
__________
Statement of Roger McClellan, Advisor, Toxicology and Human Health Risk
Analysis
executive summary
The scientific basis for policy decisions on setting the
PM2.5 National Ambient Air Quality Standard remain highly
uncertain.
The continued use of the PM2.5 indicator is a default
decision driven by EPA's past emphasis on regulatory compliance
monitoring--``monitor that which is regulated.'' As a result, there is
no database for considering alternative PM indicators that might target
specific PM constituents or exclude certain constituents.
The scientific database provides a basis for the Administrator
making policy choices for a PM2.5 NAAQS with 24-hour
averaging time concentration in the range of 25 to 35 m3,
with a 98th percentile form, and an annual standard in the range of 12
to 15 m3.
The scientific database for policy decisions on setting a
PM10-2.5 NAAQS is very weak and highly uncertain. A science-
based decision, as contrasted with a judicial decision, would be to
continue with a PM10 NAAQS.
There are major uncertainties in risks associated with exposure
to ambient PM2.5 at current levels and the benefits of
reducing PM2.5. These uncertainties need to be clearly
documented and conveyed in numerical calculations used for policy
decisions and in the Agency's final Regulatory Impact Analysis.
Expert elicitations of opinions on PM2.5 risks are
very likely flawed with a blurring of the distinction between
scientific evaluation and policy choices. Scientists, as do all
citizens, have values that influence choices of standard setting
options. However, scientific evaluations should be as free as possible
of concern for the ultimate policy decisions.
Good Morning, Mr. Chairman and members of the subcommittee. Thank
you for the invitation to present my views on the U.S. Environmental
Protection Agency's current review on the National Ambient Air Quality
Standards (NAAQS) for Particulate Matter (PM).
my background
My biography is attached to this statement. Since 1999, I have
served as an Advisor to public and private organizations on issues
related to air quality in the ambient environment and workplace drawing
on more than 45 years of experience in comparative medicine,
toxicology, aerosol science, and risk analysis. I served as President
of the Chemical Industry Institute of Toxicology in Research Triangle
Park, NC from 1988 to 1999, providing leadership for a research program
directed to understanding the mechanisms of action of chemicals in
producing either beneficial or harmful effects on humans. I was with
the Lovelace organization in Albuquerque, NM from 1966 to 1988,
providing leadership for one of the World's major research programs
directed toward understanding the potential human health effects of
inhaled materials.
The testimony I offer today draws on my experience serving on
numerous scientific advisory committees. This has included service on
many EPA advisory committees from the origin of the Agency to date,
including the Clean Air Scientific Advisory Committee (CASAC), which I
chaired from 1988 to 1992, all of the CASAC PM Panels as well as CASAC
Panels that considered other criteria pollutants. My involvement in
advising EPA on the setting of NAAQS for criteria pollutants began with
my chairing in 1977 and 1978 an ad hoc committee to review the first
lead criteria document, a committee that was required since the
Congress had not yet authorized creation of CASAC. I also served on the
National Academy of Sciences/National Research Council (NAS/NRC) on
Research Priorities for Airborne Particulate Matter and the earlier
NAS/NRC Committee that produced the report ``Science and Judgment in
Risk Assessment.'' It is important to note that the testimony I offer
today reflects my own views and is not being offered on behalf of any
of the Committees I have served on for the EPA, the NAS/NRC nor for any
other agencies or firms.
setting national ambient quality standards
Each NAAQS consists of four elements: (a) an indicator (such as
PM2.5), (b) an averaging time (such as 24 hours or annual),
(c) a numerical level (such as 65 m3 for PM2.5
averaged over 24 hours), and (d) a statistical form (such as a 98th
percentile). The indicators for five of the criteria pollutants are for
measurement of the mass concentration of specific chemicals such as
O3, SO2, NO2, CO and Lead. Only in the
case of particulate matter is the indicator based on the mass
concentration of airborne particulate matter in a specific size range,
irrespective of the chemical composition of the PM.
Under the Clean Air Act, the EPA Administrator is required to
review the NAAQS for the criteria pollutants at 5-year intervals to
evaluate whether or not the four elements of the NAAQS are still deemed
to be acceptable based on current scientific knowledge as it applies to
the assessment of public health risks. In practice the interval between
reviews has been longer. The process for review and promulgation of a
NAAQS, either continuation of the existing standard or establishing a
new NAAQS, consists of multiple phases. The initial phase, which is
obviously on-going, consists of conduct of research on the various
criteria pollutants. This includes a broad spectrum of activities;
understanding emissions of pollutants, transport and transformation of
pollutants in the atmosphere, ambient measurements of pollutants,
estimation of personal exposures to pollutants, assessment of toxic
effects and mechanisms of action in cells, tissues and animals, conduct
of controlled exposure studies to pollutants in human volunteers and
epidemiological investigations of human populations. Most of the
research is funded by the EPA, some in the Agency's own laboratories
and some in academic and other laboratories, the National Institutes of
Health and, to a modest extent, private industry. The dominance of
Federal Government support of research on criteria pollutants relates
to their effects being of broad societal concerns with the pollutants,
by and large, having no unique industrial emission source.
The findings of this research are used by the EPA's Office of
Research and Development to prepare a criteria document (CD). Each CD
traditionally has been essentially an encyclopedia of everything known
about a given criteria pollutant and is used as a basis of information
for the preparation of a Staff Paper (SP) by the EPA's Office of Air
Quality Planning and Standards. This is a Policy Assessment of
Scientific and Technical Information; in short, an integration and
synthesis of the information in the CD that is most relevant to setting
the four elements of a NAAQS. In recent years, the Staff Papers have
made substantial use of risk assessments for the criteria pollutant
being considered. These risk assessments have been conducted by a
single EPA Contractor organization. The various versions of the CD and
SP are released to the public with an invitation to provide comments as
a basis for improving the documents.
Throughout this process, a Clean Air Scientific Advisory Committee
Panel, operating as an element of the EPA's Science Advisory Board, is
involved in reviewing and advising on the scientific content of both
the CD and the SP, including the related risk assessment. This has
typically involved several revisions. Prior to the current cycle of PM
review, the CASAC Panel sent a closure letter to the EPA Administrator
when the CASAC was of the opinion that the revised documents were
suitable for use by the Administrator in promulgating a NAAQS. In the
current review, the ``closure letter'' process was abandoned.
At the next step, the Administrator proposes, via a Federal
Register Notice, a NAAQS including specific proposals for each of the
four elements of the NAAQS; the indicator, averaging times, numerical
levels and statistical forms. Comments are solicited from the Public
with the opportunity to submit written comments to a specific Docket.
In the current PM review, the CASAC PM Panel offered written comments
on the Administrator's proposal.
The next step is for the Administrator to promulgate a NAAQS
consisting of the four elements discussed previously. I purposefully do
not use the phrase ``final step,'' because the Courts may have a role
in deciding whether the Administrator's proposed NAAQS will stand. The
NAAQS are to be based on the available scientific information reviewed
in the CD and SP and summarized in the notice of proposed standards.
The primary, health-based NAAQS are to be set at a level that will
protect public health, including sensitive populations, with an
adequate margin of safety. The Administrator is precluded from
considering cost in the setting of the NAAQS.
At this point, I would like to emphasize that there exists no
absolute and unambiguous scientific methodology that can determine
which specific indicator, the precise averaging time, numerical level
or statistical form that will be adequate to protect public health. The
available scientific information can inform the NAAQS decisions,
however, the Administrator must ultimately use policy judgment in
making decisions on each of the four elements from among an array of
scientifically acceptable options including consideration of their
attendant scientific uncertainties.
Once the NAAQS are finalized, individual states have responsibility
for planning and taking actions to meet the NAAQS. This includes the
formal step of preparing ``State Implementation Plans (SIPs). In
developing strategies for meeting the NAAQS, the States can give
consideration to costs in setting the pace for achieving the NAAQS.
However, attainment of the NAAQS cannot be postponed indefinitely.
epa administrator made policy choices consistent with the science
At this juncture, I note that I personally find acceptable the
Administrator's policy choices for the PM NAAQS, as published in the
Federal Register (January 17, 2006) from among an array of science-
based options, to be acceptable. Specifically, I find scientifically
acceptable his proposal to use (a) a PM2.5 indicator with a
24-hour averaging time and a reduction in the concentration level from
65 m3 to 35 m3 with a 98th percentile form, (b)
retention of the PM2.5 annual standard at 15 m3
with additional constraints on the use of spatial averaging, and (c)
use of a PM10-2.5 indicator with a 24-hour averaging time
concentration level set at 70 m3 with a 98th percentile
form. I support the exclusion of any ambient mix of PM10-2.5
where the majority of coarse particles are rural windblown dust and
soils and PM generated by agricultural and mining sources.
Of these several policy choices, I have the greatest reservation
concerning the proposal for a PM10-2.5 indicator with a 24-
hour averaging time concentration level set a 70 m3 with a
98th percentile form. The scientific basis for the proposed
PM10-2.5 standard is very weak and uncertain. I would have
personally preferred to see the PM10 standard continued to
provide public health protection from particulate matter mass in the
PM10-2.5 range. However, EPA personnel have related that
this option has been precluded by Court decisions.
selection of a pm indicator--chained to the regulatory compliance
monitoring lamp post
The primary scientific data used to select indicators for PM NAAQS
has been derived from epidemiological investigations. Prior to 1970,
there was limited regulation of particulate matter in air pollution.
Limited monitoring, relative to that being carried out today, was
conducted using relative crude metrics of Black Smoke and Total
Suspended Particulates (TSP). TSP was the mass of particulate matter,
not identified as to chemical form, collected on a filter in a high
volume air sampler. This included material up to about 40 m in size.
Scientists studied the relationship between the air concentration of
these TSP measurements and increases in health effects. This
epidemiological data provided the basis for setting the 1971 PM NAAQS
with TSP as an indicator. The 24-hour averaging time standard set at
260 m3, not to be exceeded more than once a year, and an
annual standard set at 75 m3, annual geometric mean. The TSP
indicator then became the ``law of the land'' and TSP began to be
routinely monitored to determine regulatory compliance.
During the 1970s and early 1980s, an increased awareness emerged on
the role of particle size in determining the fraction of inhaled
particles that would be deposited and where they would be deposited in
the respiratory tract. This led to some groups making measurements of
ambient air particulate matter mass in different size fractions; less
than 15 m, less than 10 m, less than 2.5 m and less than 1 m. However,
the primary epidemiological data in the 1980s that could be used for
standard setting was TSP--remember TSP was required to be measured for
regulatory compliance.
In 1987, the PM NAAQS indicator was changed from TSP to
PM10. The choice of PM10 was heavily influenced
by a decision in the international community to use a PM10
metric rather than a PM15 metric. The United States followed
suit. Much of the epidemiological evidence for setting a
PM10 NAAQS was based on extrapolations from epidemiological
studies using the TSP monitoring data. The PM10 primary
standards were set at 50 m3, expected annual arithmetic mean
over 3 years, and 150 m3, 24-hour average, with no more than
one expected exceedance per year. With the promulgation of the
PM10 indicator the regulatory compliance monitoring shifted
from TSP to PM10. Unfortunately, ambient air monitoring of
PM15, PM2.5 and PM10 was essentially
discontinued. Obviously, it would have been expensive to continue, and,
after all, it was not required for regulatory compliance.
In the early 1990s, epidemiological data began to be published on
the association between elevated PM2.5 levels and their
association with increased health effects. The data came principally
from the Harvard Six Cities study that fortunately had included in its
early years measurements of PM10 and PM2.5. Other
analyses were published based on an American Cancer Society cohort
taking advantage of fragmentary PM2.5 ambient monitoring
data. Other investigations conducted using the PM10 ambient
monitoring data were extrapolated to a PM2.5 indicator.
These data provided the basis for promulgating a PM2.5 NAAQS
in 1997. The PM2.5 NAAQS were set at 15 m3,
annual arithmetic mean, and 65 m3, 24-hour averaging time
with a 98th percentile of concentration at each population-oriented
monitor. Associated with this was a change in the regulatory compliance
monitoring network to emphasize PM2.5 mass measurements
without regard to chemical composition. Because a PM10 mass
NAAQS was still in place measurements of PM10 mass, not
characterized as to chemical composition, continued. Using the
difference between the PM10 mass measurements and
PM2.5 mass measurements, it was possible to estimate
PM10-2.5 mass concentrations.
At various times there has been an interest in measuring PM sulfate
mass concentration, a secondary pollutant arising in the atmosphere
from conversion of SO2 gas. There have also been some short-
term monitoring campaigns in which extensive chemical characterization
of a number of particulate matter constituents have been measured.
However, the extent of this monitoring data is limited in comparison
with that developed for regulatory compliance purposes on
PM10 mass and PM2.5 mass, not characterized as to
chemical composition. Indeed, to date the database on specific PM
constituents has been insufficient to set a NAAQS for a specific PM
component. Obviously, Lead is an exception. Likewise, the data on
specific PM constituents were not viewed as to exclude any constituent
from regulation.
The most recent CD and SP focuses on the PM2.5
indicator. The focus on PM2.5 was not based on any careful
scientific analysis that led to the conclusion that PM2.5
mass, not identified as to chemical composition, as the most
appropriate metric to relate to an increase in health effects. The
simple fact is that because of the EPA's emphasis on regulatory
compliance monitoring, the only PM air quality metrics that could be
evaluated epidemiologically were PM10 mass, PM2.5
mass and to a lesser extent PM sulfate and to an even lesser extent,
PM10-2.5. I will be so bold as to say the focus on
PM2.5 mass, irrespective of chemical composition, was a
default decision, not a science-based decision.
My discussion so far has focused on epidemiological evidence
without considering the results of toxicological studies using cells,
tissues or laboratory animals. As a toxicologist, I wish I could give
more emphasis to the conduct and interpretation of toxicological
studies. However, such studies have a very limited role in the PM NAAQS
setting process. Although such studies can use new tools of modern
molecular and cellular biology and genomics, the results are not
necessarily relevant to setting the NAAQS. The challenges of
extrapolating from laboratory animals to humans, from high to low
levels of exposure, from studies of a few days or even a few months to
human lifetimes and from studies of a few normal healthy young animals
to large human populations including individuals with cardiopulmonary
disease, principally from smoking, are substantial. At best, the
toxicological investigations can help provide some guidance to the
design and conduct of epidemiological investigations. The toxicological
methods are simply to blunt and yield results that at best can be
extrapolated qualitatively to human populations. I know of no
scientific methods for using the results of toxicological studies with
PM, not characterized as to chemical composition, or those conducted
with specific PM constituents to develop quantitative numerical
standards that are at the core of PM NAAQS.
a shift in monitoring strategy to facilitate epidemiological
evaluations
What are the prospects for the next PM NAAQS review in 5 years
including a rigorous evaluation of specific PM constituents? Without a
major revolutionary change in the EPA's approach to ambient air
monitoring, I think it will be more of the same. In short, because of
the past focus on measuring PM2.5 and PM10-2.5,
these metrics will continue to be evaluated in future epidemiological
studies. Because of the substantial and continuing improvements in air
quality, including PM2.5, PM10 and
PM10-2.5, it will be even more difficult to detect
associations between these PM mass metrics and health effects. Future
epidemiological studies will also be challenged due to continuing
reductions in cardiopulmonary disease related to reductions in the
primary risk factor for these diseases--Cigarette Smoking.
How can the prospects for improved epidemiological investigations
be changed? If the EPA, in cooperation with States and Municipalities,
radically modifies its ambient air monitoring network over the next 2
years, it may be possible to have the results of improved
epidemiological studies in 8 to 10 years. The development of an
improved ambient air monitoring network will require some tough
decisions. It is obvious that the expense of an altered monitoring
network will require that only a modest number of PM constituents be
measured in multiple cities in different regions across the United
States. Some clear candidates would be sulfates, nitrates, organic
carbon, elemental carbon, silica and some specific metals for which
concern may exist as to their potential hazard. It is essential that
all of the criteria pollutant gases, ozone, SO2,
NO2 and CO, continue to be measured. With a richer array of
monitoring data available it may be possible to test hypotheses as to
the relative potency of the various PM constituents as well as the
gaseous pollutants. In any long-term studies, it will be crucial to
have accurate smoking history data if the very small potential effects
of air pollution are to be separated from the large cardiopulmonary
impacts of cigarette smoking. In addition, because of the relationship
between PM-associated hydrocarbons and volatile and semi-volatile
hydrocarbons these should be measured. In my opinion, it will be futile
to measure dozens of individual chemical species with the view that
these measurements could be useful in future epidemiological studies.
The current highly uncertain signal of air pollution associated health
effects is so small that ``teasing out'' effects related to any single
PM chemical constituent will be extraordinarily challenging.
selection of averaging times, numerical levels and statistical forms
Having selected an indicator, it is necessary to proceed to
decisions on the averaging times, numerical levels and statistical
forms for the NAAQS. These three elements are inter-related and are set
based on the epidemiological database. The averaging times are driven
by the temporal characteristics of the monitoring data, 24-hour
measurements that can be aggregated to yield annual values which, in
turn, are used in the epidemiological investigations. Hence, it is
reasonable to use 24-hour and annual averaging times.
The selection of specific numerical levels for the 24-hour standard
has been guided primarily by considering the results of epidemiological
studies of the association between daily changes in the PM indicator
and changes in mortality (all cause, cardiovascular and respiratory
mortality). The power of these studies is directly related to the size
of the population being studied and the number of days being monitored.
Thus, results can only be developed for quite large cities. This
approach would not be feasible for small communities and rural areas.
The primary input for establishing the PM2.5 annual
standard comes from long-term follow-up of cohort populations, the
Harvard Six City Study of about 8,000 individuals initiated in 1979 and
the American Cancer Society cohort assembled starting in 1979. In these
studies, sophisticated statistical techniques have been used to attempt
to tease out an association between differences in PM2.5
ambient concentrations in different communities and the risk of death
from various diseases. The analyses are very complicated because of the
numerous factors that can influence the death rate including age,
cigarette smoking, work history, education, socio-economic status,
exposure to other pollutants as well as other factors.
The results of the cohort epidemiological studies are typically
reported as a linear coefficient of increase in relative risk per 10
m3 of PM2.5 using whatever PM2.5
monitoring data are available for the specific cohort. Thus, for the
studies initiated in 1979, this may be PM2.5 measurements
made in 1979-1983. Recall that in the 1980s, there was a move to
regulate PM10 measurements of PM2.5 were
discontinued and not re-instituted until after the PM2.5
NAAQS was promulgated in 1997. The PM2.5 exposure of
individuals in the cohort prior to 1979 is unknown although it is well
recognized that in most areas air quality has substantially improved
since 1970.
A major challenge in analyzing and interpreting the results of the
cohort studies relates to the uncertain role of pollution exposures for
the individual populations prior to initiation of the studies and the
uncertainty in the statistical models used to attribute relative risk
to the various risk factors including PM2.5. The small size
of the PM2.5 relative risk poses a special challenge. This
includes the difficulty of determining the shape of the exposure-health
response relationship extending from past high levels down to current
levels. Especially vexing is the issue of whether a threshold does or
does not exist in the exposure-health response relationship. In my
view, the exposure-response relationships are highly uncertain in the
range of typical ambient PM2.5 concentrations in the United
States. The substantial uncertainty in the applicability of the
PM2.5 exposure-health response coefficients at current
ambient concentrations requires caution in calculating either
PM2.5 associated risks or the benefits of any reductions in
PM2.5 concentrations.
expert advice elicitation
In an attempt to better characterize the uncertainties in
PM2.5-associated health risks and, conversely, the benefits
in reductions in PM2.5, some individuals have suggested the
use of an ``expert advice elicitation'' approach. I am familiar with
this approach having served as one of the five experts in EPA's pilot
project to elicit opinions on the relationship between PM2.5
exposure and death. I have also participated in such approaches in the
initial stages of planning and interpreting safety assessment studies.
I think the expert opinion elicitation process may have merit in
obtaining a qualitative assessment of the impact of exposure to
hazardous materials. However, I have serious reservations as to its use
in eliciting quantitative characterizations of risk for various levels
of PM2.5 exposure for different populations in different
parts of the United States.
The interviewers eliciting the expert opinions play a major role in
determining the outcome of the process. In the session I participated
in, I found the interviewer focusing on eliciting quantitative linear
exposure-response coefficients. Since it is my professional opinion
that it is very unlikely that a linear relationship exists between
PM2.5 exposure and health responses down to and including
current ambient levels, the interview and the follow-up discussions
proved frustrating for both me and the interviewer. In short, the
sponsor (in this case, the EPA) can influence the interviewer to frame
a series of questions that will yield a pre-determined answer. In my
case, I felt the desired answer was what linear risk coefficient
(exposure health-response) would I prefer.
I am also concerned about the process used to select experts for
participation. In my opinion, the process should be very transparent
with regard to the criteria used to include or exclude individual
experts from a Panel. My concerns extend to the inclusion of
individuals who may have conducted and reported on the key studies
being used in the expert opinion elicitation process. It is human
nature to want to have one's own data and analyses used in the same
manner as originally reported.
Any additional concern with the process is the approach of using
secondary interviews in an attempt to gain consensus from the experts
as to the outcome. I understand that was done with the full-scale
expert elicitation panel whose input is to be used in the final
regulatory impact analysis. A major challenge in any elicitation of
expert advice is separating the individuals science-based input from
their personal sources with regard to a policy outcome. In my opinion,
the results of that expert elicitation are likely to be seriously
flawed. I would urge the Administrator to not use the results of the
expert opinion elicitation as input for quantitative estimates of
risks/benefits associated with PM2.5 exposure. Such an
approach is not a substitute for more rigorous uncertainty analysis
that attempts to characterize all the factors that impact on estimating
risks of PM2.5 exposure and the benefits of reductions in
PM2.5 exposure.
As an alternative to expert opinion elicitation, I urge CASAC to
document the scientific views of each of the CASAC PM Panel members
with regard to quantitative aspects of the PM NAAQS. This approach was
used in the previous PM review that concluded with promulgation of the
1997 PM2.5 NAAQS. A copy of the table included in the CASAC
PM Panel's ``closure letter'' is attached. As may be noted, individual
Panel members had a wide range of views with regard to setting the
PM2.5 NAAQS. I would personally prefer to see each of my
scientific colleagues express their individual science-based opinions
rather than have CASAC Panel participants cajoled to reach a consensus.
scientific evaluations versus policy decisions
A major challenge I see for all scientists, and especially for
CASAC PM Panel members participating in the NAAQS review process, is to
recognize the distinction between scientific evaluations and policy
judgments. In my comments to Mr. Bill Wehrum and Dr. George Gray on
improving the NAAQS review process, I noted--``It would be helpful if,
at each step in the NAAQS process including each meeting of the
scientists preparing the Criteria Documents and the Staff Paper and
their review by CASAC, if each participant were reminded. ``Every
individual should recognize the distinction between scientific
evaluation and policy decisions and recognize that the matters being
dealt with are at the interface of science and policy. Each individual
participant is asked to leave their individual ideologies and thoughts
on policy decision outcomes at the door before deliberating on the
science.'' This is not a matter of an individual's employment, i.e.,
academic, government, industry, etc. or political affiliation. It
applies to all participants. This is an especially vexing issue for
scientists involved in evaluating their own research results or that of
close colleagues. In today's resource constrained world everyone wants
to have their work used in the public arena, moreover, they would like
to see the door left open or opened wider for them to do more work on
the topic under consideration. Indeed, some individuals, including
CASAC Panel Members, desire a ``sense of accomplishment''--some
individuals interpret that as--did we participate in lowering the
NAAQSs' Some have suggested that there would be a ``limited sense of
accomplishment'' if only the 24-hour PM2.5 standard were
lowered and the Annual PM2.5 standard was left unchanged.
Yes, scientific evaluations and policy decisions do get intertwined by
individual scientists in expressing their own personal preferences on
life science issues.''
______
Response by Roger McClellan to Additional Question from
Senator Voinovich
Question 1. Dr. McClellan, as you heard Dr. Gray testify, he said
characterization of uncertainties is extremely important. EPA is using
a process called expert elicitation to respond to the NAS
recommendation on this important issue. We will see the elicitation's
results when the rule is finalized. Please elaborate on your concerns
with this process.
Response. I am very familiar with the ''expert elicitation'' being
used by the EPA to obtain expert opinions on the quantitative
relationship between changes in PM.5 concentrations in
ambient air and changes in indices of adverse health effects. As I
noted in my written testimony, I participated in a pilot expert
elicitation study conducted by an EPA contractor as a prelude to the
larger study recently conducted for the EPA.
In my professional opinion, the expert opinion elicitation process
being used by the EPA is not a scientifically adequate approach to
characterizing either the central estimate nor associated range of
potential values for changes in adverse health outcomes associated with
changes in ambient PM2.5 concentrations for contemporary
PM2.5 levels found across the United States. The process is
at best an elicitation of the opinions of a selected group of experts.
It is not a substitute for rigorous scientific characterization of
uncertainty of the kind I understand was recommended by the National
Academy of Science/National Research Council in its 2002 report,
``Estimating the Public Health Benefits of Proposed Air Regulations.''
The outcome of the expert opinion elicitation process is
influenced by five factors:
(1) the nature of the request from the sponsor, in this case the
EPA, to the
organization or individual(s) eliciting the opinions;
(2) the nature of the questions posed by the opinion elicitators;
(3) the manner in which the experts are identified and selected;
(4) the scientific background and personal biases of the experts;
and
(5) how the individual opinions are compiled and reported.
It is apparent that a requested organization can influence the
results by how they frame the question(s) presented for use in the
elicitation process. In similar fashion the manner in which the
questions are posed to the experts is critical. I understand that the
central question asked, which was similar to that posed to me in the
pilot study, was ``What is the percent reduction in excess health risks
associated with a 1 m3 decrease in the ambient
PM2.5 concentration'' This question assumes that there is a
linear relationship between changes in PM2.5 concentrations
and excess health risks. This is an assumption, not a scientific fact.
Indeed, I would argue that the results of the recent papers by Enstrom
(2005) and Laden et al. (2006) clearly indicate a lack of a
statistically
significant excess in mortality associated with contemporary
ambient PM2.5 levels and, thus, an absence of a linear
exposure-response relationship at low levels of PM2.5.
Interestingly, this includes results from SouthernCalifornia and
Stuebenville, OH, areas in which PM2.5 levels were quite
high in the past and have been substantially reduced.
The scientific background of the experts, their personal biases and
how they are selected can influence the outcome of the elicitation
process. Individuals whose research findings are under consideration in
the expert opinion elicitation process are placed in an awkward
position. Are they willing to set aside their own vested interest in
seeing their results used in favor of giving a broader opinion? In
giving an opinion, can they avoid being concerned with how the results
will be used? Will a tighter standard result in a generally more
favorable view of their research?
The manner in which results are ultimately compiled and presented
is of critical importance. In the most recent process, the experts were
polled individually and then brought together as a group to re-affirm
the outcome. I argue that this consensus building approach suppresses
uncertainty rather than contributing to a full expression of
uncertainty. I would urge the EPA Administrator to not give significant
weight to the results of the expert opinion elicitation process in
setting the annual standard for PM2.5. Moreover, I
definitely feel it would be inappropriate to use the results of the
expert elicitation process as input to the calculation of benefits in
the Regulatory Impact Analysis.
Let me again emphasize that it is my professional opinion that it
is imperative that the uncertainties associated with estimation of the
excess adverse health effects of exposure to PM2.5 at
contemporary levels be rigorously characterized. In my opinion, the EPA
at each step in the PM review process, failed to adequately
characterize uncertainties. This was true of the Criteria Document, the
Staff Paper and the Risk Assessment that under-girded the Staff Paper.
A rigorous uncertainty analyses would have started with critically
examining the base data from critical studies and the very
sophisticated statistical methods used. When the original published
reports did not contain sufficient detail, the EPA should have obtained
more detailed data and, if necessary, conduct additional analyses. This
was done in a laudatory manner for some studies under the auspices of
the Health Effects Institute, however, even more critical analysis
would have been useful in better characterizing uncertainty in
estimating PM2.5 risks.
For example, very little was done to examine the validity of the
Cox proportional hazard model for characterizing the very small
estimated excess risks for PM2.5 exposure and other
confounders. Moolgavkar (2005 and 2006) has noted the limitations in
this widely used model for estimating small excess risks. In none of
the key studies being used to set the annual PM2.5 standard
were results presented for cigarette smoking, the major risk factor for
cardiorespiratory disease. These results would have given an indication
of how well the Cox model was working for the major risk factor,
cigarette smoking, and, thus, gave some greater confidence in the use
of the model for characterizing much lower levels of risk for
PM2.5. If historical information were not available on
cigarette smoking this should have been identified as an uncertainty
and, indeed, quantified.
The impact of cigarette smoking is apparent from the analysis of
Enstrom (2005) using data for Southern California. Beyond presenting
the analysis of PM2.5 risks, the Enstrom paper also included
the results of an analysis of the relative risk of death from all
causes by cigarette smoking status. These results are of special value
because they illustrate the substantial magnitude of the cigarette
smoking effects relative to other risk factors such as air pollution.
The baseline was never smokers (as of 1959 and 1972) for deaths 1973 to
2002 set at 1.000. Former smokers (as of 1959 and 1972) were Relative
Risk (RR)-1.054 and increased to 1.253 for former smokers (as of 1972
only). Current smokers (as of 1972) had relative risks that increased
with smoking intensity 1-9 cigarettes per day (cpd)RR-1.239; 10-19
cpdRR-1.97 cpdRR-1.871, 21-39 cpdRR-2.068 and 40+ cpdRR-2.543. The
large relative risks related to cigarette smoking level provide
perspective for the small relative risks reported for long-term
PM2.5 exposure. Indeed, in part because he had smoking
histories available on the subjects in the Southern California cohort
he was able to conclude These epidemiologic results do not support a
current relationship between fine particulate pollution and total
mortality in elderly Californians, but they do not rule out a small
effect, particularly before 1988. For 1983-2002, the RR was 1.00. This
included a substantial number of individuals exposed to
PM2.5 at concentrations above the current Annual Standard of
15 m3. Moreover, the substantial effects of cigarette
smoking emphasize the importance of accurate inclusion of cigarette
smoking history in any long-term cohort study of the effects of PM and
dictate that consideration of smoking be included in any quantitative
characterization of uncertainty in estimating PM2.5 risks.
Question 2. Dr. McClellan, you have been on these panels for a long
time. Does the science ever point specifically to a number or are these
standards really set in the judgment of the Administrator?
Response. I have served on each of EPA Clean Air Scientific
Advisory Committees that have provided advice to the Administrator of
the setting on revision of National Ambient Air Quality Standards for
Particulate Matter. There has been a general acknowledgement by the
Advisory Panel members that the science should inform the decisions on
the four elements of the standard; (a) indicator, such as
PM2.5, (b) averaging time, such as annual, (c) numerical
level, such as 15 m3, and (d) the statistical form. There
has also been general recognition that the Administrator has the
ultimate responsibility for setting the standard using his/her
judgment. The Advisory Panels have regularly reviewed and commented on
the range of numerical levels presented in the Staff Paper, thereby
acknowledging that the science can inform a broad range for setting the
standard rather than the science yielding a specific numerical level.
Each successive review of the Particulate Matter standard has become
more contentious. In my opinion, some of the contentious nature has
been driven by early concern by special interest groups and some
members of the Panel as to the outcome of the process. In short, a
premature view has developed that the PM2.5 standard should
be tightened, a view advanced long before the scientific evaluation had
been completed. One way for CASAC to enhance the potential for the
standard being tightened is to endorse a range of numerical values
below that of the current standard. If this is done the judgment of the
Administrator is constrained. This is exactly what was done when the
CASAC PM Panel truncated the proposed range for setting an Annual
PM2.5 standard from 12 to 15 m3 to 13-14
m3. The CASAC PM Panel advanced an argument for changing the
lower limit of the range from 12 to 13 m3, i.e. uncertainty
increased below 13 m3. In none of the letters from the CASAC
Chair to the Administrator was a rigorous rationale provided for
reducing the upper bound of the range from 15 to 14 m3. In
my view, by endorsing an exceptionally narrow range of 13 to 14
m3 the CASAC PM Panel ignored the uncertainty in the science
and attempted, inappropriately in my view, to exercise some of the
judgment that is reserved for the Administrator by the Clean Air Act.
The CAA wisely calls for CASAC to advise the Administrator and for the
Administrator to use judgment in setting the standard. The CASAC PM
Panel, in narrowing the range to 13 to 14 m3, stopped one
small step short of attempting to set the PM2.5 Annual
standard.
Question 3. As a member of the CASAC panel was there complete
agreement on the recommendations.
Response. There was not complete agreement among CASAC PM Panel
Members on the narrowing of the range for the Annual PM2.5
standard from 12 to 15 G53 to 13 to 14 m3. George Wolff and
I, who had both previously served as Chair of the Clean Air Scientific
Advisory Committee, disagreed with the proposal. It is of interest that
the decision to narrow the range was reached during short conference
calls of the Panel and by electronic exchanges among small groups of
members. The decision was not one reached on the basis of a typical
face-to-face public meeting of the Panel. There was also intense
pressure to obtain and present a consensus view and to provide a letter
from CASAC to the Administrator that was devoid of attached individual
views as customary for CASAC letters to the Administrator, especially
on important matters. I exercised my rights as a private citizen to
prepare a letter to the docket expressing my views on the setting of
the PM standard. In my professional opinion, the available scientific
information is consistent with setting an Annual PM.5
Standard in the range of 12 to 15 m3 as articulated in the
EPA Staff Paper with the specific numerical level to be selected by the
Administrator based on judgment as specified in the Clean Air Act.
__________
Prepared statement of Anne Smith, Vice President, CRA International
Mr. Chairman and members of the committee, thank you for your
invitation to participate in today's hearing. I am Anne Smith, and I am
a Vice President of CRA International. I am a specialist in
environmental risk assessment and integrated assessment to support
environmental policy decisions, which was a core element of my Ph.D.
thesis at Stanford University in economics and decision sciences. I
have performed work in the area of risk assessment over the past 30
years, including as an economist in the USEPA's Office of Policy,
Planning, and Evaluation, as a consultant to the USEPA Air Office, and
in many consulting engagements since then for Government and private
sector clients globally while employed first at Decision Focus
Incorporated and then CRA International. I have also served as a member
of several committees of the National Academy of Sciences focusing on
risk assessment and risk-based decision making.
I have been deeply involved in assessment of the evidence on risks
from ambient fine particulate matter (PM2.5) since EPA first
turned to the task of identifying an appropriate National Ambient Air
Quality Standard (NAAQS) for PM2.5 over 10 years ago. I
testified to this same committee in 1997 on the nature of the
scientific evidence underlying the PM2.5 NAAQS proposed at
that time. I thank you for the opportunity to share my perspective
today on the current scientific evidence and associated risk assessment
for PM2.5 and how it has evolved since 1997. My written and
oral testimony today provide a statement of my own research and
opinions, and does not represent a position of my company, CRA
International.
I would like to start by summarizing what I think are the most
important and overarching considerations that should be accounted for
when considering whether to alter the current PM2.5 NAAQS,
which include an annual average limit of 15 m3 and a 24-hour
average limit where the 98th percentile of observations over all days
must be below 65 m3. I will then summarize results of
analyses I have done to synthesize the recent PM2.5 health
studies into an assessment of risks. Complete details and documentation
of my analyses are in my written comments on the current Proposed Rule
for a revised PM2.5 NAAQS, which were submitted into the
PM2.5 docket in April, 2006. I am attaching a copy of my
written comments to EPA to further substantiate the points that I make
in my testimony today.
The key points that I wish to make about the scientific evidence on
risks of PM2.5 that are relevant for making a decision on
the standard are:
EPA and the courts recognize that the PM2.5 NAAQS must
be set at a level that still has some positive level of risk, because
the science has yet to advance far enough to identify any threshold
exposure level for effects, below which risk would be indistinguishable
from zero. This was true in 1997 and it remains true today.
EPA's own quantitative estimates of mortality risk at attainment
of the current NAAQS are lower today than they were when EPA set that
standard in 1997 ``with an adequate margin of safety,'' after
accounting for the many uncertainties. This is true for both long-term
(``chronic'') exposures to PM2.5 (which are addressed by the
annual average limit) and short-term (``acute'') exposures (which are
addressed by the 24-hour average limit).
The reduction in the quantitative estimates of risk is apparent
even in EPA's own risk analysis, but most of the reasonable alternative
results reported in the same studies that EPA has relied on imply even
lower quantitative risk estimates for PM2.5.
Looking more broadly beyond quantitative risk estimates, the many
additional studies of PM2.5 mortality risks since 1997 have
demonstrated that many of the risk estimates become ``statistically
insignificant'' when re-estimated in reasonable alternative ways. A
``statistically insignificant'' result directly implies a positive
probability that there is no effect at all. Thus, when we look at all
of the data in the new studies as a group, we find more statistical
evidence now than was available in 1997 that PM2.5 may not
be the culprit pollutant, and that there may be no causal relationship
at all between PM2.5 and mortality.
In thinking about whether to tighten either the annual or daily
standard, one might ask, what has changed in our knowledge since 1997
that would undermine the Administrator's 1997 judgment that the current
PM2.5 NAAQS are neither more nor less stringent than
necessary to protect the public health with an adequate margin of
safety? A thorough review of the new evidence suggests that the margin
of safety that the Administrator selected in 1997 is likely to be
larger than was thought at the time.
quantitative estimates of risk remaining at the current standard have
fallen
EPA has acknowledged that the PM2.5 NAAQS cannot be set
at a level that corresponds to zero risk.\1\ However, EPA has also
argued that its quantitative risk estimates cannot be used to identify
a specific point where it should set a standard:
---------------------------------------------------------------------------
\1\ See, for example, the Proposed Rule, p. 2622 (i.e., 71 FR
2622).
``[I]n the Administrator's view, a risk assessment based on studies
that do not resolve the issue of a threshold is inherently limited as a
basis for standard setting, since it will necessarily predict that ever
lower standards result in ever lower risks, which has the effect of
masking the increasing uncertainty inherent as lower levels are
considered. As a result, while the Administrator views the risk
assessment as providing supporting evidence for the conclusion that
there is a need to revise the current suite of PM2.5
standards, he judges that it does not provide a reliable basis to
determine what specific quantitative revisions are appropriate.'' \2\
---------------------------------------------------------------------------
\2\ 71 FR 2648.
I concur that a risk assessment that makes no attempt to
incorporate the uncertainty on where a threshold may exist will indeed
only serve to promote ever lower standards without a sound basis. Since
EPA has not incorporated such uncertainty into its risk assessment,
that risk assessment is indeed incapable of helping to identify where
to set the standard. However, since EPA views the risk assessment as
supporting a conclusion on whether there is a need to revise the
standard, it is appropriate and relevant to compare EPA's current
quantitative risk estimates and the associated statistical measures of
a PM2.5 effect to those estimates that were available in
1997. In the Proposed Rule, EPA partially acknowledges that risk
estimates are lower today than in 1997 for the two cities that were
included in both its 1997 and current risk analyses. With respect to
short-term exposure risk estimates, EPA states that ``the magnitude of
the estimates associated with just meeting the current annual standard.
. . is similar in one of the locations. . . and the current estimate
is lower in the other location.'' \3\ With respect to the long-term
exposure risks, EPA states that the risk estimates ``are very similar
for the two specific locations included in both the prior and current
assessments.'' \4\
---------------------------------------------------------------------------
\3\ 71 FR 2640.
\4\ 71 FR 2640.
---------------------------------------------------------------------------
EPA does not provide the actual numerical estimates for these two
cities. They are:
For acute risks in Los Angeles, in 1997 EPA estimated that 1.7
percent of mortality would continue to be attributable to
PM2.5 once Los Angeles would be in attainment with the
current NAAQS. Today EPA's risk estimate has fallen to 0.5 percent and
this current estimate is statistically insignificant (which means that
there is a fairly large chance that this particular estimate suggests
that there is really no PM2.5 effect at all).
For acute risks in Philadelphia, in 1997 EPA estimated that 1.5
percent of mortality would continue to be attributable to
PM2.5 at attainment of the current NAAQS. The risk estimate
that EPA now uses for Philadelphia is 2.2 percent. Although this is
higher than in 1997, EPA has selected a single estimate out of a very
large number of estimates reported in the epidemiological study it is
relying on for Philadelphia. In fact, that study actually concluded
that PM2.5 did not appear to explain the mortality risk as
well as ozone, and the residual risk for PM2.5 after
simultaneously accounting for the role of ozone would have produced a
lower estimate--about 0.8 percent--which is lower than in 1997. This
more thoroughly-controlled estimate also is not statistically
significant.
Chronic risk estimates do not vary from city to city, because the
statistical method to estimate relative chronic risks produces a single
value that applies to all cities. I will therefore only relate the
results for Los Angeles here. For chronic risks, in 1997 EPA estimated
that 2.0 percent of mortality would continue to be attributable to
PM2.5 at attainment of the current NAAQS. Today, EPA's risk
estimate for the same attainment status is 1.8 percent--in other words,
the chronic risk estimate also is lower now, even though the quote from
the Proposed Rule above suggests that the estimate has not changed.
The Proposed Rule only referred to a comparison of risks for these
two cities. However, it is actually possible to make the same
comparison for the other six cities that EPA has included in its
current risk analysis. This is because there was only one
PM2.5 acute mortality study it could have used for each of
those cities back in 1997 the same one that it used for Los Angeles and
Philadelphia.\5\ For five of the other six cities in the current risk
analysis, EPA's acute risk estimates today are lower than they would
have been estimated to be in 1997, and all the cities have lower
chronic risk estimates. Table 1 summarizes the cities and the results
of my comparison of their risk estimates.
---------------------------------------------------------------------------
\5\ This was the paper by Schwartz, Dockery, and Neas (1996) on
acute risks in six U.S. cities.
---------------------------------------------------------------------------
When I reviewed the original papers that EPA is relying on, I also
found that EPA's risk analysis has selectively used the highest or
near-highest risk estimates supported by each paper. This means that
risks estimates that more fully reflect the body of evidence are likely
lower still than EPA's risk analysis suggests. Additionally, as for
Philadelphia, I found that San Jose would have had a much lower risk
estimate than in 1997--literally zero now--if EPA had chosen to use the
one reasonable alternative result for PM2.5 reported in the
San Jose study.\6\ Thus, the full body of evidence can support risk
estimates that would be lower now than in 1997 for every one of the
eight cities in EPA's current risk analysis.
---------------------------------------------------------------------------
\6\ That is, EPA's risk estimate for San Jose is based on a 1-
pollutant regression that associated mortality with PM2.5 on
the same day as death. The study also reported results of a comparable
1-pollutant regression that was identical in all ways except that it
associated mortality with PM2.5 from the day before death.
The latter regression produced a negative risk estimate, which I
interpret to be evidence of no effect at all (rather than evidence of a
beneficial effect of PM2.5).
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Table 1 also reports that the PM2.5 findings are not
statistically significant across all of the alternative reasonable risk
estimates in each underlying study. This was not the case in 1997. At
that time, there was a much more limited set of studies and estimates
within each study--but for some cities, all the estimates available at
the time were statistically significant. Today, the opposite it true.
Every single study that EPA has relied on for its current risk analysis
contains alternative estimates that indicate that PM2.5 does
not have a statistically significant association with mortality, yet
EPA chose not to use this part of the new information.
In conclusion, EPA has stated that the risk assessment's role is to
provide ``supporting evidence'' on whether there is a need to revise
the PM2.5 standard. In this role, EPA's own risk analysis
provides no evidence supporting a decision to tighten the standard now.
The risks are lower now than they were when the standard was set in
1997. The higher estimates of risks were determined to be ``requisite
to protect the public health with an adequate margin of safety'' in
1997, and the quantitative risk analysis suggests that that margin of
safety has grown, not narrowed, as a result of the many more recent
PM2.5 health effects studies.
The question then remains whether other aspects of the new evidence
provide an overriding reason for tightening the standard. The other
part of EPA's reasoning for how to set the standard relies on what EPA
calls an ``evidence-based approach.'' Simply put, EPA looks at all of
the studies that estimate the statistical relationship of
PM2.5 with health effects, and seeks to identify a level of
PM2.5 above which statistically significant effects are
found, and below which statistically significant effects are not found.
In applying the evidence-based approach, EPA states that the large
quantity of new studies of acute effects justifies the use of acute
studies to set the 24-hour standard, and that chronic studies should be
used to determine where to set the annual standard:
``Given the extensive body of new evidence based specifically on
PM2.5 that is now available, and the resulting broader
approach presented in the Staff Paper, the Administrator considers it
appropriate to use a different approach from that used in the last
review to select appropriate standard levels. More specifically, the
Administrator's proposal relies on an evidence-based approach that
considers the much expanded body of evidence from short-term exposure
PM2.5 studies as the principal basis for selecting the level
of the 24-hour standard and the stronger and more robust body of
evidence from the long-term exposure PM2.5 studies as the
principal basis for selecting the level of the annual standard.'' \7\
---------------------------------------------------------------------------
\7\ 71 FR 2648.
I will next discuss how the evidence in the long-term exposure
studies of PM2.5 has weakened since 1997, thus removing any
necessity to tighten the annual standard under EPA's evidence-based
approach. I will then discuss how the evidence in the short-term
exposure studies of PM2.5 that are the basis for the 24-hour
standard also has weakened.
the evidence in long-term exposure studies has weakened since 1997
In 1997, the two prominent long-term exposure studies (one based on
a sample population, or ``cohort'' in 154 U.S. cities that was tracked
by the American Cancer Society, and one based on a sample population in
just six US cities that was tracked by Harvard School of Public Health)
both had published findings of a statistically significant relationship
between long-term exposure to PM2.5 and life expectancy.
These studies were subjected to an extensive process of reanalysis
under the auspices of the Health Effects Institute (HEI) that was
released in 2000.\8\ This reanalysis is widely reputed to have
confirmed the original studies' results; however, a complete reading of
the actual report shows that some major statistical concerns underlying
those results were unearthed. Although a positive PM2.5
effect was still found in those data sets, the ability to interpret
those results as clearly causal in nature was weakened.
---------------------------------------------------------------------------
\8\ The report on findings of these reanalyses is Krewski et al.
(2000).
---------------------------------------------------------------------------
EPA acknowledges that the concerns identified in the HEI reanalyses
of the long-term exposure studies for PM2.5 remain
unresolved to the present time:
The Administrator also recognizes a contrasting view as to the
interpretation of and weight to be accorded to the results from the
ACS-based studies (Pope et al., 1995; Krewski et al., 2000; Pope et
al., 2002). In this view, the ACS-based studies are not sufficiently
robust to support a policy response that would tighten the annual
PM2.5 standard based on the evidence. This view emphasizes
the sensitivity of the results of these studies to plausible changes in
model specification with regard to accounting for the geographical
proximity of cities and the correlation of air pollutant concentrations
within a region, effect modification by education level, and inclusion
of SO2 in the model. In this view, these sensitivities
suggest potential confounding or effect modification that has not been
taken into account. For example, concern has been raised about the
sensitivity of results in the reanalysis of data from the ACS cohort
study (Krewski et al., 2000) to inclusion of SO2 in the
models. [T]he reanalysis found that PM2.5, sulfates, and
SO2 were each associated with mortality in single-pollutant
models. However, in two-pollutant models with SO2 and
PM2.5, the relative risk for PM2.5 was
substantially smaller and no longer statistically significant, whereas
the effect estimates for SO2 were not sensitive to inclusion
of PM2.5 or sulfates in two-pollutant models. In this view,
the ACS-based risk estimates are more robust for SO2 than
for PM2.5 or sulfates. In further extended analyses, Pope et
al. (2002) reported that effect estimates were not highly sensitive to
spatial smoothing approaches intended to address spatial
autocorrelation, while findings of effect modification by education
level were reaffirmed. Results of multi-pollutant models were not
reported by Pope et al. (2002). Because the correlation coefficient
between PM2.5 and SO2 was 0.50 in the ACS data,
in this view it is plausible to believe that the independent effects of
the two pollutants could be disentangled with additional study.'' \9\
---------------------------------------------------------------------------
\9\ 71 FR 2652.
---------------------------------------------------------------------------
The quote above is lengthy, which highlights that the concerns
identified in the long-term exposure studies are many. The quote above
also indicates that the new set of results using the American
Cancer Society cohort that was published after the HEI reanalyses
(i.e., Pope et al., 2002) did not help resolve these issues.
Specifically, the 2002 paper ignores concerns that the purported
PM2.5 effect instead might be attributable to the gaseous
pollutant SO2, and re-affirms a troubling finding that
PM2.5 only seems to create mortality risk only for
individuals who have not continued their education beyond the high
school level. (The latter finding is discussed further below.)
Additionally, the 2002 paper still finds that the PM2.5
effect is diminished and rendered insignificant when applying
statistical methods to correct a clear statistical error that the HEI
report found in the original results. Nevertheless, the Pope et al.
(2002) paper continues to use the estimation method that is subject to
error except in a sensitivity analysis; and EPA continues to rely on
the uncorrected estimates in its risk analysis. Even with these dubious
selections from the full body of literature, EPA's estimates of long-
term exposure risk are lower than in 1997.
Figure 1 illustrates the degree to which the evidence on long-term
exposure risk has fallen, both in the overall magnitude of the risk
estimate, and also in terms of a greater degree of uncertainty in the
estimate. Figure 1 uses the case of Los Angeles at attainment of the
current standard, yet the relative patterns evident in this figure are
the same for all cities in the U.S. All of the risk estimates in Figure
1 labeled ``1'' through ``7'' are based on the American Cancer Society
cohort, which has received the majority of attention. The estimate on
the far left of the figure, labeled ``1'' is the estimate from 1997
(note that the estimate is 1.5 percent, as reported for Los Angeles in
the preceding section), and the estimates to the right are other key
results from the HEI reanalyses and from the more recent Pope et al.
(2002) paper. The estimate labeled ``2'' is the single result from the
many new estimates that is used for the current EPA estimates of long-
term risk (which is 1.3 percent, as I stated in the previous section).
It is quite apparent from the figure that the current risk estimate
is among the highest that could be found among the more recent results.
If any of the others (labeled ``3'' through ``7'') had been used for
EPA's risk analysis, the current risk estimate of 1.3 percent for Los
Angeles would instead be in the range of 0.3 percent to 1.0 percent--
much lower than the original 1997 risk estimate that was available when
the current standard was first set.
The set of results on the far right of the figure (labeled ``8''
through ``12'') reflect the findings based on a new study of a third
sample population that had not been identified or studied as of 1997.
It is known as the ``Veterans' Cohort.'' I believe this study to be of
some policy relevance regarding whether or not the annual standard
needs to be tightened, given that this study finds no effect at all of
PM2.5 on life expectancy in this particular cohort. EPA has
chosen to give ``greatest weight'' to results from the American Cancer
Society and the Six Cities cohorts because they have been reanalyzed
and scrutinized so thoroughly. While this may be a reasonable judgment,
EPA has actually gone further than that, and accorded the Veterans'
Cohort results zero weight. Its findings should be acknowledged with
somewhat more than zero weight. When one does so, the overall evidence
regarding long-term PM2.5 risks is further weakened.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
EPA also refers to the perplexing finding that level of education
determines whether or not there is a PM2.5 association.
Figure 2 illustrates this finding, which was first identified in the
HEI reanalyses, and which remains in the more recent Pope et al. (2002)
study. Clearly education per se is not believed to be the cause of
sensitivity to exposure to PM2.5, yet the important (and
still unanswered) question is: what is educational level indicating
about risks that these sample populations face? What could possibly
explain the complete lack of a PM2.5 effect among those with
higher educations? When such a pattern appears in epidemiological study
results, it indicates that there is still an important explanatory
factor that is missing from the statistical estimation method something
correlated with education. Until that factor is identified and included
in the estimation of PM2.5 risks, estimates of the effect of
PM2.5 are biased. The PM2.5 estimate could be
higher, or it could vanish altogether. Thus, the unexplained pattern
related to education in all of these studies remains a very important
warning about the pitfalls of making a causal interpretation regarding
long-term exposure risks of PM2.5.
In summary, the evidence against a need to tighten the annual
standard is not just founded on the fact that the numerical long-term
risk estimates are now lower than when the current standard was set.
The more important point is that the basis for interpreting the long-
term studies as unbiased evidence of a causal relationship between
PM2.5 and chronic mortality risk has weakened. This was a
concern in 1997, and the reanalyses and new studies since then have
done more to amplify these concerns than to allay them. In the face of
this evidence of greater uncertainty, combined with the reduced
quantitative risk estimates, there is no justification for tightening
the annual standard on the basis of the long-term exposure studies.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
A Summary CASAC's Case for Tightening the Annual Standard
EPA's Clean Air Scientific Advisory Committee (CASAC) has made the
case to tighten the annual standard with two lines of reasoning,
neither of which is founded on the long-term risk studies.\10\
---------------------------------------------------------------------------
\10\ CASAC's reasoning is stated in a letter to the Administrator
on March, 21 2006, pp. 3-4.
---------------------------------------------------------------------------
CASAC's first line of reasoning is that EPA reports substantial
risk would remain at the current standard. As EPA and the courts have
long established, the PM2.5 NAAQS cannot be a zero-risk
standard. CASAC was concerned by the estimates of remaining risk, but
never deliberated the question of whether this risk estimate had risen
or fallen since the standard was deemed ``requisite to protect the
public health with an adequate margin of safety.'' As I have shown
above, the risk estimates fell, both for chronic and acute risks, but
EPA never reported this fact to CASAC during CASAC's review of the
Staff Paper and associated risk assessment. In the face of this fact,
the only other argument to tighten the annual standard might be if
there were stronger reason to believe that the effects found in these
studies are causal in nature. However, EPA set the current standards
with a presumption (precautionary in nature) that the estimated
PM2.5 risks were causal. This cannot therefore be the
rationale to tighten the standards.
Hence, to argue that the standard should be tightened because there
is evidence that risk remains at the current standard is a logic that
would force a tightening of the standard in every future review cycle,
even if no new evidence were to have become available at all since the
previous review. There is nothing in the law or in precedent that
dictates that the standard has to be tightened as the result of a NAAQS
review.
CASAC's second line of reasoning comes closer to the heart of how
EPA first set the standard. CASAC notes that there are three new acute
studies that find PM2.5 associations with mortality at
annual averages below the current annual standard (all with reported
annual averages in the range of 13 to 14 m3). These studies
are: Burnett and Goldberg (2003) for 8 Canadian cities combined; Mar et
al. (2003) for Phoenix, and Fairley (2003) for Santa Clara County, CA
(referred to in the risk analysis as San Jose).\11\
---------------------------------------------------------------------------
\11\ CASAC actually cites a paper by Lipsett et al. (1997) in its
letter, but that paper has nothing to do with PM2.5
mortality. I interpret CASAC to have wanted to cite Fairley (2003).
---------------------------------------------------------------------------
The first thing to realize about this part of CASAC's case for a
tightened annual standard is that it is using studies that consider
only how day-to-day changes in PM2.5 levels affect day-to-
day numbers of deaths relative to the number of deaths that might
otherwise be expected on each day (e.g., relative to numbers of deaths
that are expected to occur on each day based on established patterns
related to the time of year, time of week, weather, etc.). Such acute
effects of a pollutant are generally believed to be associated with
spikes in PM2.5, although studies to date rarely report
evidence of any threshold level below which the association disappears.
Nevertheless, there is no clear linkage between the annual average in a
city, and the extent to which day-to-day spikes in PM2.5
might be occurring. If an acute effect is found in a city that happens
to have a low annual average, there is no reason to believe that the
estimated association is not still due to sudden upward changes in
PM2.5 from one day to the next. The city may simply have a
large number of very clean days that pull the annual average
PM2.5 down, while not eliminating the presence of many days
of sudden increases in pollution that are logically likely to be the
cause of any acute risk that the study is finding.
Thus, it is not necessarily correct to assume that if acute risks
are found in a city with a low annual average pollution, then such
risks exist in all cities with low annual average pollution. In fact,
if one believes that there must be a threshold where the acute risk
from exposure to pollution drops off somewhere above zero, then the
linkage between the annual average of pollution and existence of acute
risk is not only unclear, but illogical. For reasons such as these, EPA
has decided to use the plethora of acute risk studies now available to
set the 24-hour standard that that type of study more meaningfully
informs. EPA has decided not to use acute studies to set an annual
standard.\12\
---------------------------------------------------------------------------
\12\ See the quote on p. 5 above, taken from 71 FR 2648.
---------------------------------------------------------------------------
Nevertheless, even if one were to use acute effects studies to
determine an ``adequate margin of safety'' for lower levels of long-
term exposure to pollution, there are good reasons to believe that the
annual average PM2.5 reported for each of the three studies
cited by CASAC may not be a good indicator of the long-term exposure
levels that account for the risk findings in these studies. I explain
why for each of the three:
Goldberg and Burnett (2003). This study reports a PM2.5
association for eight Canadian cities combined. The annual average of
13.3 m3 is an average over all of the eight cities, while
the annual averages in the individual cities vary from 9.5
m3 to 17.7 m3. There are no city-specific results
reported to help indicate whether the estimate of an acute effect is
due to effects in each of the eight cities, or only in a few.\13\
Evidence in the paper suggests that there may in fact be different
effects in each city.
---------------------------------------------------------------------------
\13\ The other multi-city PM2.5 mortality studies (based
on the Six Cities data set) report effects by individual city as well
as for the combined set. This was the data on which EPA set the current
standards, and in doing so, EPA used annual averages for only the
individual cities that did have significant effects within the set of
six. The lowest such city was Boston, with an annual average of 15.6
m3, which was the basis for the current annual average
standard of 15 m3.
---------------------------------------------------------------------------
Another concern with this study is that it is a reanalysis of a
more comprehensive study that included consideration of the role of
gaseous pollutants as well.\14\ The original study concluded that the
gaseous pollutants had a much greater ability to explain mortality
risks
---------------------------------------------------------------------------
\14\ The original paper was Burnett et al. (2000)
---------------------------------------------------------------------------
than both PM2.5 and PM10-2.5 combined.
However, when the paper had to be reanalyzed, the authors did not
reanalyze the portions that considered gaseous pollutants in
conjunction with particulate pollution, and so this finding is no
longer discussed.
Fairley (2003). This study used data from Santa Clara County, CA,
over a seven year period, and during that time pollution levels were
falling dramatically. Although the annual average PM2.5 that
is attributed to this study is 13.6 m3, the annual average
was as high as 18.4 m3 at the start, and fell progressively
to 9.5 m3 by the end of the 7 years studied.\15\ Peak levels
of PM2.5 were also falling, starting at a 98th percentile of
88 m3 for the first year and ending at 25 m3.
Such a wide range within this one city's data set begs the question:
are the reported acute effects relationship driven largely by the high
levels in the early years, or are they also evident in the later years?
This highly relevant question is never mentioned, let alone analyzed,
by the authors. Lacking any exploration of such an obviously relevant
issue, it would seem a dubious proposition to use the annual average
over the entire time period in this one study as the basis for a
national ambient standard.
---------------------------------------------------------------------------
\15\ To know this, one must go back and read the original study
that this is a reanalysis of, Fairley (1999).
---------------------------------------------------------------------------
Another concern with this study is that it reports PM2.5
risk estimates for two alternative methods of estimation, both of which
are reasonable. One method considers whether deaths tend to fluctuate
with the same day's PM2.5 levels and the other method
considers whether deaths tend to fluctuate with the previous day's
PM2.5 levels. The same-day estimate finds the positive
association that this study is known for, but the estimate based on
PM2.5 on just the previous day is actually in the negative
direction. Complete reversal of evidence of a PM2.5
mortality effect by considering PM2.5 levels only 24 hours
apart in time presents a concern for interpreting the study's same-day
estimate as a causal one. However, there is no discussion of what these
conflicting results might mean.
Mar et al. (2003). This study considered acute risks in Phoenix,
AZ, with annual average PM2.5 levels of 13.5 m3.
There are 10 estimates of PM2.5 risk in the paper, and only
3 of them are significant. More importantly, this is not the only paper
that studied the ability of this same set of PM2.5 data to
explain acute mortality risks in Phoenix. One of the other studies
found that PM2.5 did not have any explanatory power, and
found instead that the coarse fraction of PM had explanatory power.\16\
The third study found evidence that there is a threshold below which
PM2.5's apparent ability to explain changes in daily
mortality disappeared.\17\ That threshold appeared to be above 20
m3. If there is a threshold, then the rationale for a
linkage between annual average PM2.5 and acute risks simply
falls apart.
---------------------------------------------------------------------------
\16\ Clyde et al. (2000).
\17\ Smith et al., (2000)
---------------------------------------------------------------------------
A final concern with all three of the Phoenix studies is that none
of them considered whether the PM2.5 effect would remain if
pollutants such as CO, SO2, ozone, or NO2 were
also included in the analysis. This is a critical gap in many of the
current studies because the new body of papers on PM2.5
health effects reveals that PM2.5 effects usually disappear
when one of the gaseous pollutants is explored. This is addressed in
the next part of my testimony.
In summary, CASAC makes its case to tighten the annual standard on
the basis of acute, not chronic effects studies. There are logical
problems with this approach to setting an annual standard; these
logical problems become apparent when looking at each of the three
acute studies that CASAC cites as its basis for recommending an annual
standard that is tighter than the current one.
the statistical evidence on acute effects of pm2.5 has also
weakened since 1997.
In 1997, there existed only one study that had used actual
measurements of PM2.5 and estimated whether daily numbers of
deaths might be associated with day-to-day variations in the
PM2.5. This was a study using the data from the Harvard
study of six U.S. cities reported in Schwartz et al. (1996), and it was
used as the basis for the current standards. In that study,
statistically significant associations of PM2.5 and acute
mortality were found in three of the four cities with the highest 98th
percentile PM2.5 levels, which ranged from 42 m3
to 44 m3. The city with the highest PM2.5 98th
percentile (which was 82 mG53) did not produce a statistically
significant association, nor did the two cities with the lowest 98th
percentile levels of 32 and 34 m3. This was the best
available information at the time,. Other than the anomaly for the city
with the highest PM2.5 exposures, it did at least suggest
that there might be a range above which effects were more likely and
below which they were more unlikely.
While this study was used as the primary basis for the current
standards (including the annual standard), there were many concerns
expressed with uncertainties in the estimation methods. In particular,
there was concern that this study had not considered the explanatory
role of any of the other common pollutants like CO, SO2,
ozone, and NO2. (These are often called the ``gaseous
pollutants'' because that distinguishes them from various forms of
particulate pollutants that are regulated under the PM NAAQS.) It was
argued that PM2.5 might be simply playing a proxy role for a
gaseous pollutant also present in the air in these cities.
As new acute PM2.5 studies were performed after 1997, a
number of these studies did strive to explore the respective roles of
PM2.5 and gaseous pollutants in the observed statistical
associations. This was done by using ``2-pollutant'' or ``multi-
pollutant'' methods, as contrasted to the ``1-pollutant'' method that
only allows a single pollutant (e.g., PM2.5 in this case) to
have any opportunity to explain mortality risk. One of the little
recognized but important insights of this body of studies is that when
gaseous pollutants also have been considered in a study, the gaseous
pollutant has taken over the explanatory role from PM2.5 in
a majority of the cases.
I determined this in my review of the studies since 1997.
Specifically, I attempted to identify all of the PM2.5
health effects studies cited in the Criteria Document (including both
mortality and morbidity effects studies) that had reported results of
any estimates for PM2.5 using a 2-pollutant method of
estimation for at least one gaseous pollutant. I found 10 such papers
among all the new studies that did report a statistically significant
association for PM2.5. Of these 10, 8 saw PM2.5
lose its ability to explain mortality risk when studied using a 2-
pollutant method. (In the other two studies, both the PM2.5
and gaseous pollutant retained statistical explanatory power.)
Often it is suggested that 2-pollutant methods are not useful
because it is impossible to unravel the effects of two pollutants that
both move up and down together in near synchrony (i.e., they are highly
``correlated''). However, my review of these papers did not find
evidence that this was a problem. If it were a problem, then both the
PM2.5 and gaseous pollutant would lose their explanatory
power. What I found instead was that in seven of the eight studies
where PM2.5 lost its erstwhile explanatory power when it was
the only pollutant considered, the gaseous pollutant retained its
explanatory power. Otherwise stated, of the ten studies that I started
with, only one seemed to be affected by intractable statistical
problems making it impossible to unravel the separate effects of the
two pollutants.
These papers are summarized in Table 2, which is more fully
explained in my written comments to EPA of April 2006, which I am
submitting with this testimony.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
The above findings represent just one of many ways that the new
body of acute effects evidence has been found to vary depending on the
particular method of estimation. Other sources of variation in the
evidence include the methods for accounting for time and weather
considerations. The new studies have demonstrated that concerns with
variability of epidemiological estimates of risk, which were expressed
but not well explored in 1997, are real. Table 1 at the beginning of my
testimony shows that even the ``best'' PM2.5 health effects
studies that EPA could select for its risk analysis present a highly
uncertain picture of whether PM2.5 is playing a causal role
for acute effects. Even if there is a causal relationship, which is
what the current standard assumed when EPA set it in 1997, there
appears to be great difficulty in determining what the size of the
effect is. The one trend that is clear is that the size of the
PM2.5 acute mortality estimates found in the many new
studies since 1997 are generally lower than the estimates that were
available when the current standards were set.
As I have already noted, EPA has chosen to use an evidence-based
approach to set the 24-hour standard. EPA has proposed to tighten the
24-hour standard from the current level of 65 m3 to 35
m3. This decision was made even though the quantitative risk
estimates based on these studies are lower and statistically weaker
than they were when the standard was set. Nevertheless, it is true that
there are many more studies available now than at the time of the
standard, and it is relevant to ask if this new body of evidence might
provide a better indication of a 98th percentile PM2.5 level
where observed effects start to drop off. EPA has attempted to make
such a case for a cut-off point of 35 m3 in the Proposed
Rule.\18\ I have gone through that case very carefully, and I have
found it incomplete. I will state what I found in general terms
here.\19\
---------------------------------------------------------------------------
\18\ See 71 FR 2649.
\19\ My written comments to EPA that are being submitted into the
record with this testimony provide complete documentation of my review
of the literature and application of an evidence-based approach for the
24-hour standard. Although the full discussion includes both mortality
and morbidity studies, I only summarize the mortality findings here.
However, the patterns I describe are similar in studies of
hospitalizations, emergency room visits, and frequency of ``symptoms''
that are not severe enough to entail a hospital visit.20 Further,
although a majority of estimates are statistically significant because
the paper focused on the lag period that was found to be significant in
the simplest, 1-pollutant starting point of the analysis, the single
other alternative 1-pollutant result found a negative risk estimate,
implying no risk at all.
---------------------------------------------------------------------------
EPA's verbal summary of its evidence-based approach used a selected
subset of ten PM2.5 mortality studies. I found another eight
such acute studies of United States or Canadian mortality cited in the
Criteria Document, that used actual measurements of PM2.5,
and that did not appear to have any unreanalyzed statistical problems
associated with the GAM software. (If a single paper reports results
for more than one city, I treat each city as a separate ``study''.) Six
of the eighteen studies that I considered are the original ``Six
Cities'' used to set the current standard. All of the others are
studies published between 1997 and the cut-off time for consideration
in this review cycle.
I read each study, and determined whether all the PM2.5
estimates reported in a study were ``more often insignificant than
significant'', ``a near 50-50 mix'', or ``more often significant than
insignificant.'' After categorizing them in this way, I found that
there is no clear pattern where statistically significant results tend
to be found for studies with higher PM2.5 levels, and that
increasingly mixed evidence is found in studies with progressively
lower PM2.5 levels. Figure 3 graphically summarizes my
findings for the mortality studies. It shows that many of the data sets
with the highest 24-hour average PM2.5 levels demonstrate
the least likelihood of a statistically-significant association with
mortality. This is contrary to what EPA states in its discussion of the
evidence-based approach in the Proposed Rule. I attribute the
difference to the fact EPA considered only a selected set of the new
studies, and not the more complete set that I identified. (The ten
studies EPA considered are shown as blue diamonds in Figure 3, while
the additional eight studies that I also considered are shown as red
diamonds in Figure 3.)
Figure 3 also shows that I determined that only 3 of the 18 studies
found statistically significant PM2.5 effects for a majority
of the methods of estimation that they reported. Of these:
One is for eight Canadian cities combined by Goldberg and Burnett
(2003), which I described earlier in this testimony. Its 98th
percentile value is about 39 m3, but this value has the same
flaw that I described for its annual average--the actual peak exposures
faced by people in the eight separate cities ranged from 27 to 48
m3, and there is no information to indicate which of the
various city-specific 98th percentiles might be accounting for the
effects estimated when all cities are combined.
Another of these was Fairley (2003) for Santa Clara County, CA,
with a 98th percentile value of 59 m3, which I also
discussed above. This study had very high exposures at first, and we
have no idea whether the statistical significance is related to the
earlier high levels, or equally attributable to later, lower
PM2.5 peaks.\20\
---------------------------------------------------------------------------
\20\ Furthery, although a majority of estimates are statistically
significant because the paper focused on the lag period that was found
to be significant in the simplest, 1-pollutant starting point of the
analysis, the single other alternative 1-pollutant result found a
negative risk estimate, implying no risk at all.
---------------------------------------------------------------------------
The third is the Boston data from the Harvard Six Cities study
that served as the basis for the current PM2.5 NAAQS in
1997. The magnitude and statistical significance of the association
observed in this data set has been reduced in reanalyses since 1997,
and none of these estimates include consideration of the potential role
of any gaseous pollutants in explaining these associations.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
In conclusion, the evidence regarding a causal relationship between
short-term exposures to PM2.5 and health has not
strengthened since 1997. To draw this conclusion, one must consider
more than just the number of new studies that have reported at least
one statistically significant association; one must also explore the
extent to which the effects reported in these studies remain
statistically significant under a range of different plausible methods
for making such estimates. In particular, the new evidence strongly
suggests that many or most of these associations may actually be
attributable to a gaseous pollutant, not PM2.5.
But even setting aside the weaknesses in the statistical evidence,
EPA's evidence-based approach for where to set a 24-hour standard for
PM2.5 leads us right back to the very data set on which the
current PM2.5 NAAQS were based in the Boston ``Six Cities''
data set. Thus, the evidence-based approach that EPA is trying to apply
provides little additional insight beyond the simple point that I
started my testimony with: the quantitative estimates of risks
remaining at the current standards are lower now than when they were
determined to offer an ``adequate margin of safety.'' They therefore do
not support a tightening of the current NAAQS.
integrated analyses of alternative results may help inform
naaqs decisions better
It is easy to feel lost regarding how to effectively interpret a
plethora of alternative studies, and of alternative risk estimates
within each study. EPA's method in performing risk analyses has been to
rely on a single estimate that it selects from the large pool of
alternatives, and to base its summaries of quantitative risk estimates
on that single estimate. In many of these summaries, even the
statistical errors associated with that one estimate are often not
reported. Some, but not all, of the remaining alternative estimates are
studied through ``sensitivity analyses.'' However, these are usually
relegated to the back pages of a technical support document. The result
of this approach is that the degree of certainty about the risk
estimates becomes greatly overstated by the time summary results reach
the eyes of decision makers, advisors, and the public. Further, the
method of selecting the single risk estimate to rely on for the primary
analysis can lead to a substantial bias in the quantitative risk
estimates reported.
There are alternative methods for performing risk assessments that
integrate multiple alternative risk estimates, and even key
uncertainties that remain purely judgmental. These methods are
sometimes called probabilistic analysis, or integrated uncertainty
analysis. EPA has not used such methods in the documents supporting the
Administrator's decision on the PM2.5 NAAQS, such as the
Staff Paper. I believe that such methods could be very useful, and
would reveal better the true extent of uncertainty that I have tried to
characterize qualitatively in my testimony above.
In 2003, at an early stage of the drafting of the current risk
assessment, I prepared some illustrative examples of an integrated
uncertainty analysis to show how the reams and reams of sensitivity
results in the risk assessment document could be condensed to more
decision-relevant information. The results of that illustrative
analysis remain of some interest:
Using just the alternative long-term exposure studies in the
Criteria Document, I found that there could be about a 40 percent
probability that there would be no long-term mortality benefit from
tightening the current NAAQS. I also estimated that the probability
that actual longterm mortality would be less than the primary risk
estimate that EPA reports in its risk analysis is about 75 percent.
I did a less thorough example for the short-term mortality risk,
based only on Los Angeles. (Short-term risk estimates and their
uncertainty vary by city). For Los Angeles, I estimated a 42 percent
probability there would be no benefits from tightening the standard
from the current level when using only the risk estimates that EPA had
itself cited in its risk analysis, and a 64 percent probability that
acute risk reductions would be lower than EPA's primary risk
estimate.\21\
---------------------------------------------------------------------------
\21\ For documentation of these calculations, see Smith (2003).
These probability estimates were based solely on actual estimates
in the new body of literature on PM2.5 mortality, and do not
include any external judgments such as whether any of these estimates
can be interpreted as causal, whether some particles are more toxic
than others, or the hypothetical presence of a threshold.
(Consideration of these issues would raise the probabilities that I
calculated.) They are strictly based in the published evidence reviewed
in the Criteria Document. They are thus indicative of the degree of
uncertainty that the published studies themselves reveal.
I believe that the process of decision making leading up to the
point where a new rule is proposed would benefit greatly if such a
synthesis of statistical and modeling uncertainties were to be
developed as a part of that process. Controversies would remain
regarding the judgments that are necessary for such estimates, but if
they are conducted in an open manner, with ample opportunity for public
review and comment, more insight about the overall implications of the
body of scientific evidence would be created before a decision must be
made than we have at the present moment. I emphasize that this should
be done during the NAAQS review cycle, with opportunities for public
review and comment, before a rule is proposed.
______
Responses by Anne Smith to Additional Questions from
Senator Voinovich
Question 1. As you stated in your statement, EPA, Dr. Thurston, and
others seem to be ignoring your argument that the acceptable level of
risk identified in 1997 has gone down. This is a very intuitive and
convincing argument. If risk has gone down, then the standard should
not be revised. I would like to give you more time to explain this
argument and respond to the other side.
Response. The current standards for PM2.5 were set by
EPA in 1997 with full knowledge that these standards would not reduce
health risks to zero. The risk analysis produced by EPA at that time
reported positive estimates of risk remaining at the current standard.
EPA's risk analysis also explicitly estimated lower levels of risk for
alternative, tighter standards than the standard that EPA ultimately
selected. EPA had to defend the risk level that it adopted in court.
The DC Circuit Court of Appeals ruled that the risks associated with
the current PM2.5 standards were ``requisite'' to protect
the public health with an adequate margin of safety, as required by the
Clean Air Act.\1\ ( ``Requisite" means that the standards are ``not
lower or higher than is necessary'' to protect public health with an
adequate margin of safety.\2\) Thus, the level of risk estimated in
1997 for the current standards was associated with an adequate margin
of safety.
---------------------------------------------------------------------------
\1\ United States Court of Appeals for the District Of Columbia
Circuit in American Trucking Associations, Inc., et al., v.
Environmental Protection Agency, No. 97-1440, March 26, 2001.
\2\ Ibid, p. 10.
---------------------------------------------------------------------------
As I wrote in my statement: ``In thinking about whether to tighten
either the annual or daily standard, one might ask, what has changed in
our knowledge since 1997 that would undermine the Administrator's 1997
judgment that the current PM2.5 NAAQS are neither more nor
less stringent than necessary to protect the public health with an
adequate margin of safety? A thorough review of the new evidence
suggests that the margin of safety that the Administrator selected in
1997 is likely to be larger than was thought at the time."
As I documented in my written testimony to the Committee, EPA's own
estimates of health risks remaining at the current standards are lower
now than the earlier 1997 estimates. They are lower in all cities for
long-term exposure risks, and they are lower for short-term exposure
risks in 6 of the 8 cities that are in the current risk analysis.
I presented my comparison of risk estimates from 1997 to those
estimated today in person to EPA air office staff on several occasions,
and never heard any disagreement regarding its factual accuracy. I have
never heard secondhand of EPA staff disagreeing with these facts in any
subsequent meetings or conversations where I was not present. I also
submitted detailed documentation of the risk comparisons in my written
comments to EPA on the draft risk analysis documents and staff paper
for PM2.5. However, EPA did not provide any such comparative
analysis to its Clean Air Scientific Advisory Committee (CASAC) during
the course of CASAC's deliberations leading up to CASAC's
recommendation to EPA for tightening the PM2.5 standards.
Thus, CASAC members were probably unaware of these facts when they made
their recommendation.\3\
---------------------------------------------------------------------------
\3\ Although my risk comparisons are based on material that EPA
placed into the record during both the 1997 and current rulemakings,
CASAC could not be expected to have easily made such comparisons on
their own. For one, the relevant measures of risk are not readily
observed given the formats in which EPA has presented its risk
findings. Further, they require one to consult risk analysis documents
from 10 years ago.
---------------------------------------------------------------------------
EPA is clearly aware of these facts, however, and even partially
acknowledged them at the time that it published the Federal Register
notice for the proposed new PM2.5 standards:
With respect to short-term exposure mortality and morbidity . . .
[c]omparing the risk estimates for the only two specific locations that
were included in both the prior and current assessments, the magnitude
of the estimates associated with just meeting the current annual
standard, in terms of percentage of total incidence, is similar in one
of the locations (Philadelphia) and the current estimate is lower in
the other location (Los Angeles). . . . With respect to long-term
exposure mortality risk estimates, the estimates in terms of percentage
of total incidence are very similar for the two specific locations
included in both the prior and current assessments.\4\
---------------------------------------------------------------------------
\4\ PR, p. 2640.
---------------------------------------------------------------------------
This quote is only a partial acknowledgment of the risk comparison,
however, because it only mentions two cities' risk estimates, and for
those two cities, one risk estimate rose and the other fell. However,
there were actually 8 cities in the current risk analysis, and the risk
estimate has fallen in 6 of those 8. Further, the statement that risk
estimates are ``similar'' for long-term exposure does not reveal that
the current risk estimate is actually lower. Notably, this statement
was made only after EPA had finalized the Staff Paper, and after CASAC
had made its recommendations.
Dr. Thurston, in his testimony before the committee, attempted to
rebut my argument in the following way:
You know, I think that unfortunately what Dr. Smith has done here
is exactly what she accuses EPA of having done. She has cherry- picked
certain results that support her position and unfortunately hasn't
looked in a balanced unbiased way at this question.\5\
---------------------------------------------------------------------------
\5\ EPW Committee Hearing transcript, July 19, 2006.
---------------------------------------------------------------------------
This is a completely false statement, as anyone should be able to
tell from the record of the hearing: my comparison is based only on
EPA's own numbers. All that I have done is assemble the information
from EPA documents and repeat it in a format that allows a direct
comparison. I made this clear in my own oral statement:
Risks due to long-term mortality have fallen in every location. The
risks due to daily exposure, which vary by city, have fallen in six of
the eight cities that are in EPA's risk analysis. This new information
suggests that the margin of safety provided by the current standards is
actually greater than we originally thought. What I have told you so
far is based entirely on EPA's own point estimates of risk.\6\
---------------------------------------------------------------------------
\6\ Ibid, emphasis added.
---------------------------------------------------------------------------
Dr. Thurston's comment is an attempt to sidestep the inconvenient
truth that even EPA's analyses support the conclusion that the
estimates of mortality risk associated with the current ambient
PM2.5 standard have fallen since the time that they were
deemed acceptable by EPA and the Court. His sidestepping tactic may be
useful in a purely oral exchange, but the transcript makes it clear
that his comment has no substance or merit in refuting my argument.
I did go on to describe how the evidence on lower risk estimates is
even stronger when one looks at the full body of evidence in the
studies that have been released since 1997, rather than just using
EPA's own numbers. However, it is illogical to claim that I have
``cherry-picked'' the evidence when I show that the downward risk
trends would be more pronounced if I use almost any other risk estimate
than those used by EPA. This is the opposite of cherry-picking.
I would also like to directly address Dr. Thurston's own argument
that there is new evidence to support tightening the annual standard.
The essence of his argument is that further analyses of the American
Cancer Society study find that there is no evidence of a threshold in
the long-term exposure risk data below the level of the current annual
standard. He presents a figure (Figure 3 in his written testimony),
that indicates that the risks from PM2.5 exposure extend
below 15 m3, supporting a reduction in the annual
PM2.5 standard at this time.\7\ While it is true that this
specific figure had not been produced at the time that the current
annual standard was established, it does not provide a case for
tightening the standard to a level below 15 m3. Dr.
Thurston's implicit premise is that the PM2.5 standard
should be a zero-risk standard, which was not and still is not the
premise of EPA (and upheld by the Court). Moreover, EPA assumed that a
causal relationship between PM2.5 and long-term mortality
risk did exist down to 9 m3
---------------------------------------------------------------------------
\7\ Statement of Dr. George D. Thurston, Sc. D. to the Committee on
Environment and Public Works of the United States Senate RE: the
Science and Risk Assessment of Particulate Matter (PM) Air Pollution
Health Effects, July 19, 2006.
---------------------------------------------------------------------------
in its risk analyses when it chose to set the annual standard at 15
m3. Thus, the ``new'' information that Dr. Thurston says
supports reducing the standard to a lower level was already assumed to
exist in the 1997 risk analyses.
Question 2. Your testimony indicates that EPA has selectively used
study results in developing the proposal. EPA emphasized studies that
show associations between particulates and negative health effects and
ignored those that do not. You also point out that in many of the
studies relied upon, the associations become statistically
insignificant when another pollutant is included in the analysis.
Please elaborate.
Response. Each of the epidemiological studies of PM2.5
risk contains many different alternative estimates of risk. Different
estimates of risk are possible because the researchers have to specify
what set of risk factors in addition to PM2.5 will be
simultaneously accounted for when estimating the quantitative amount of
risk that is associated with PM2.5. For example, one can
estimate a PM2.5 risk level with or without also including
the humidity level. PM2.5 concentrations may tend to be
higher on days of higher humidity, yet humidity is known to be a
physiological stressor and there is an established increased risk of a
sick individual dying on the relatively more humid days of a year. If
humidity is left out of the analysis, then PM2.5 could serve
as a ``proxy'' for the missing humidity information, and this could
bias the level of risk estimated for PM2.5 upwards. All of
the current studies take care to address humidity, because its role as
a stressor is so well established. However, there is no objective
method for determining exactly how to specify humidity. Further, there
are many alternative ways to specify a wide range of other factors that
also need to be accounted for to avoid bias in the PM2.5
effect, and not all of the relevant health stressors are known. Thus,
researchers tend to try different methods for incorporating non-
PM2.5 determinants of health, and each different method
results in a different PM2.5 risk estimate.
Sometimes two seemingly comparable methods produce dramatically
different results. For example, the study of PM2.5 risk in
Santa Clara County, CA by Fairley (2003) reports results of two
estimates that differ in one single way: in one case, mortality risk
one a particular day was compared to PM2.5 levels on the
same day, and in the second, mortality risk on each day was compared to
PM2.5 levels that occurred the day before.\8\ The first of
these finds a positive and statistically significant association
between rises in PM2.5 and mortality risk. The second finds
the opposite: a negative association (i.e., mortality risk tended to be
lower if PM2.5 on the day before had been relatively high).
There is no objective statistical means for determining whether one of
these statistical findings is closer to the truth than the other, but
they produce extremely different estimates of health risk from ambient
PM2.5. In this case, EPA has selected the positive and
statistically significant result for use in its risk analysis, and EPA
makes no mention of the opposite alternative result that it could have
selected from the same paper.
---------------------------------------------------------------------------
\8\ Fairley, D. (2003) Mortality and Air Pollution for Santa Clara
County, CA, 1989-1996 Revised Analyses of Time-series Studies of Air
Pollution and Health, Special Report. Boston, MA: Health Effects
Institute; pp. 97-106.
---------------------------------------------------------------------------
One important class of researcher judgments where very different
risk estimates can result is that related to ``single-pollutant''
versus ``multi-pollutant'' estimation methods. Most commonly, the only
pollutant that is included in the statistical estimation procedure is
PM2.5 (at least, in studies that report PM2.5
risk estimates). If any other pollutant has an association with
mortality risks, then PM2.5 will be the only pollutant that
could adopt its explanatory power and it will do so if PM2.5
is statistically correlated with the missing pollutant. As a result,
the PM2.5 risk estimate will be biased, and may appear
falsely to be statistically significant. In a multi-pollutant
statistical estimation procedure, PM2.5 and one or more
other pollutants are all included in the analysis at the same time.
Then, each can account for its own role as a health stressor, and the
resulting estimates for each pollutant, including PM2.5, are
less likely to be biased.
I attempted to identify all of the PM2.5 health effects
studies cited in the Criteria Document (including both mortality and
morbidity effects studies) that had reported results of any estimates
for PM2.5 using a multi-pollutant method of estimation for
at least one gaseous pollutant. I found ten such papers among all the
new studies that did report a statistically significant association for
PM2.5. Of these 10, 8 found that PM2.5 lost its
ability to explain mortality risk when studied using a 2-pollutant
method. (In the other two studies, both the PM2.5 and
gaseous pollutant retained statistical explanatory power.) This
suggests a reasonable concern that PM2.5 may not be the
pollutant that is actually causing the reported health risks. However,
EPA's risk analysis is relying solely on the one-pollutant results from
each of these papers, and thus is overstating the degree of confidence
in these results. Notably, 2 of the 8 cities in EPA's risk analysis are
being estimated using single-pollutant estimates from papers that also
reported that PM2.5 risk estimates are not statistically
significant if estimated in a two-pollutant model: Los Angeles and
Philadelphia.\9\, \10\ In both cases, the authors concluded that a
different pollutant had more explanatory power than PM2.5,
yet EPA has used only the obviously biased estimate of PM2.5
risk that came from the single-pollutant methods that the original
researchers have repudiated. (Even with use of the one-pollutant
estimate, EPA's mortality risk estimate for Los Angeles is not
statistically significant.)
---------------------------------------------------------------------------
\9\ Moolgavkar, S. H. (2003) "Air Pollution and Daily Deaths and
Hospital Admissions in Los Angeles and Cook Counties" Revised Analyses
of Time-series Studies of Air Pollution and Health, Special Report
Boston, MA: Health Effects Institute; pp. 183-198.
\10\ Lipfert, F. W.; Morris, S. C.; Wyzga, R. E. (2000a) Daily
Mortality in the Philadelphia Metropolitan Area and Size-classified
Particulate Matter Journal of the Air & Waste Management Association
15011513.
---------------------------------------------------------------------------
In my response so far, I have explained how EPA has been selective
in the way it has made risk estimates for 3 of the 8 cities in its
mortality risk assessment (i.e., for Santa Clara County/San Jose,
Philadelphia, and Los Angeles). Problems with the remaining 5 cities
risk estimates also deserve mention:
For both Pittsburgh and Detroit, the risk estimate that EPA uses
is not statistically significant.\11\, \12\
---------------------------------------------------------------------------
\11\ Chock, D. P.; Winkler, S.; Chen, C. (2000) A Study of the
Association Between Daily Mortality and Ambient Air Pollutant
Concentrations in Pittsburgh, Pennsylvania Journal of the Air & Waste
Management Association 50: 1481-1500.
\12\ Ito, K. (2003) Associations of Particulate Matter Components
with Daily Mortality and Morbidity in Detroit, Michigan. Revised
Analyses of Time-series Studies of Air Pollution and Health. Special
report. Boston, MA: Health Effects Institute; pp. 143-156.
---------------------------------------------------------------------------
For Phoenix, there are 2 other papers by other authors that use
the same PM2.5 data set.\13\ One of the other papers finds
no PM2.5 association at all (but does find a coarse fraction
association).\14\ The third paper reports a PM2.5
association only above a threshold. EPA's risk assessment does not
include such a threshold.\15\
---------------------------------------------------------------------------
\13\ The paper EPA relies on is: Mar, T. F.; Norris, G. A.; Larson,
T. V.; Wilson, W. E.; Koenig, J. Q. (2003) ''Air Pollution and
Cardiovascular Mortality in Phoenix, 1995-1997'' Revised Analyses of
Time-series Studies of Air Pollution and Health, Special Report Boston,
MA: Health Effects Institute; pp. 177-182.
\14\ Clyde, M. A.; Guttorp, P.; Sullivan, E. (2000) ''Effects of
Ambient Fine and Coarse Particles on Mortality in Phoenix, Arizona''
Seattle, WA: University of Washington, National Research Center for
Statistics and the Environment; NRCSE technical report series, NRCSE-
TRS no. 040.
\15\ Smith, R. L.; Spitzner, D.; Kim, Y.; Fuentes, M. (2000)
''Threshold Dependence of Mortality Effects for Fine and Coarse
Particles in Phoenix, Arizona'' Journal of the Air & Waste Management
Association 50: 1367-1379.
---------------------------------------------------------------------------
Risk estimates for both St. Louis and Boston are based on single-
pollutant methods because the researchers do not report any multi-
pollutant methods at all.\16\, \17\ However, for both cities, the
authors do report a number of different PM2.5 risk estimates
using different ways of accounting for non-pollutant variables that
also account for changes in mortality risk and EPA has adopted the
estimate for each city that produces the highest level of
PM2.5 risk. Several of the alternative estimates are also
not statistically significant (particularly for St. Louis). There is no
reason to expect those higher risk estimates to be more valid than the
alternative estimates reported by the researchers. In summary, there is
a substantial amount of uncertainty associated with risk estimates for
all eight of the cities in EPA's current risk assessment. The way EPA
has selected a single risk estimate out of the many estimates provided
in the original epidemiological studies for each city has a clear
tendency to overstate the level of PM2.5 mortality risk and
to understate the uncertainties.
---------------------------------------------------------------------------
\16\ Klemm, R. J.; Mason, R. (2003) ``Replication of Reanalysis of
Harvard Six-City Mortality Study.'' Revised Analyses of Time-series
Studies of Air Pollution and Health. Special report. Boston, MA: Health
Effects Institute; pp. 165-172.
\17\ Schwartz, J. (2003) ``Daily Deaths Associated with Air
Pollution in Six US Cities and Short-term Mortality Displacement in
Boston'' Revised Analyses of Time-series Studies of Air Pollution and
Health. Special report. Boston, MA: Health Effects Institute; pp. 219-
226.
Question 3. Please summarize the concerns that you have with the
three studies that CASAC relied on in recommending a tightened annual
standard.
Response. CASAC notes that there are three new acute studies that
find PM2.5 associations with mortality at annual averages
below the current annual standard (all with reported annual averages in
the range of 13 to 14 g/m3). Most importantly, there are
logical flaws in setting an annual standard based on studies of daily
mortality risk. Acute risks are associated with increases in
PM2.5 on certain days and there is no logical link between
the the annual average conditions at a given location and the riskiness
of spikes in PM2.5 at that location.
However, even if one were to want to set an annual standard based
on short-term exposure risk studies, each of these papers presents
reasons to be concerned about the prospect of using them as the basis
for a national standard:
Goldberg and Burnett (2003).\18\ This study reports a
PM2.5 association for eight Canadian cities combined. The
annual average of 13.3 m3 is an average over all of the
eight cities, while the annual averages in the individual cities vary
from 9.5 m3 to 17.7 m3. There are no city-
specific results reported to help indicate whether the estimate of an
acute effect is due to effects in each of the eight cities, or only in
a few.\19\ Evidence in the paper suggests that there may in fact be
different effects in each city.
---------------------------------------------------------------------------
\18\ Burnett, R. T.; Goldberg, M. S. (2003) ``Size-fractionated
Particulate Mass and Daily Mortality in Eight Canadian Cities.''
Revised Analyses of Time-series Studies of Air Pollution and Health.
Special report. Boston, MA: Health Effects Institute; pp. 85-90.
\19\ The other multi-city PM.5 mortality studies (based
on the Six Cities data set) report effects by individual city as well
as for the combined set. This was the data on which EPA set the current
standards, and in doing so, EPA used annual averages for only the
individual cities that did have significant effects within the set of
six. The lowest such city was Boston, with an annual average of 15.6
um3, which was the basis for the current annual average
standard of 15 m3.
---------------------------------------------------------------------------
Another concern with this study is that it is a reanalysis of a
more comprehensive study that included consideration of the role of
gaseous pollutants as well.\20\ The original study concluded that the
gaseous pollutants had a much greater ability to explain mortality
risks than both PM2.5 and PM10-2.5 combined.
However, when the paper had to be reanalyzed, the authors did not
reanalyze the portions that considered gaseous pollutants in
conjunction with particulate pollution, and so this finding is no
longer discussed.
---------------------------------------------------------------------------
\20\ Burnett, R. T.; Brook, J.; Dann, T.; Delocla, C.; Philips, O.;
Cakmak, S.; Vincent, R.; Goldberg, M. S.; Krewski, D. (2000)
''Association Between Particulate- and Gas-phase Components of Urban
Air Pollution and Daily Mortality in Eight Canadian Cities'' In: Grant,
L. D., ed. PM2000: Particulate Matter and Health. Inhalation Toxicology
12(suppl. 4): 15-39.
---------------------------------------------------------------------------
Fairley (2003).\21\This study used data from Santa Clara County,
CA, over a seven year period, and during that time pollution levels
were falling dramatically. Although the annual average PM2.5
that is attributed to this study is 13.6 m3, the annual
average was as high as 18.4 m3 at the start, and fell
progressively to 9.5 m3 by the end of the 7 years
studied.\22\ Peak levels of PM2.5 were also falling,
starting at a 98th percentile of 88 m3 for the first year
and ending at 25 m3. Such a wide range of PM2.5
levels within this one city's data set begs the question: Are the
reported acute effects relationship driven largely by the high levels
in the early years, or are they also evident in the later years? This
highly relevant question is never mentioned, let alone analyzed, by the
authors. Lacking any exploration of such an obviously relevant issue,
it would seem a dubious proposition to use the annual average over the
entire time period in this one study as the basis for a national
ambient standard.
---------------------------------------------------------------------------
\21\ Fairley, op. cit.
\22\ To know this, one must go back and read the original study
that Fairley (2003) reanalyzes: Fairley, D. (1999) Daily Mortality and
Air Pollution in Santa Clara County, CA: 1989-1996? Environmental
Health Perspectives 107: 637-641.
---------------------------------------------------------------------------
Another concern with this study is that it reports PM2.5
risk estimates for two alternative methods of estimation, both of which
are reasonable. One method considers whether deaths tend to fluctuate
with the same day's PM2.5 levels and the other method
considers whether deaths tend to fluctuate with the previous day's
PM2.5 levels. The same-day estimate finds the positive
association that this study is known for, but the estimate based on
PM2.5 on just the previous day is actually in the negative
direction. Complete reversal of evidence of a PM2.5
mortality effect by considering PM2.5 levels only 24 hours
apart in time presents a concern for interpreting the study's same-day
estimate as a causal one. However, there is no discussion of what these
conflicting results might mean.
Mar et al. (2003).\23\ This study considered acute risks in
Phoenix, AZ, with annual average PM2.5 levels of 13.5
m3. There are ten estimates of PM2.5 risk in the
paper, and only three of them are significant. More importantly, this
is not the only paper that studied the ability of this same set of
PM.5 data to explain acute mortality risks in Phoenix. One
of the other studies found that PM2.5 did not have any
explanatory power, and found instead that the coarse fraction of PM had
explanatory power.\24\ The third study found evidence that there is a
threshold below which PM2.5's apparent ability to explain
changes in daily mortality disappeared.\25\ That threshold appeared to
be above 20 m3. If there is a threshold, then the rationale
for a linkage between annual average PM2.5 and acute risks
simply falls apart.
---------------------------------------------------------------------------
\23\ Mar et al., op. cit.
\24\ Clyde et al., op. cit.
\25\ Smith et al., op. cit.
---------------------------------------------------------------------------
A final concern with all three of the Phoenix studies is that none
of them considered whether the PM2.5 effect would remain if
pollutants such as CO, SO2, ozone, or NO2 were
also included in the analysis. This is a critical gap in many of the
current studies because the new body of papers on PM2.5
health effects reveals that PM2.5 effects usually disappear
when one of the gaseous pollutants is explored. (My response to
Question 2 above addresses this point in more detail.)
__________
Statement of John Stephenson, Director, Natural Resources and
Environment, U.S. General Accountability Office
Mr. Chairman and members of the committee, I am pleased to be here
today as the committee considers the science and risk assessment
supporting the Environmental Protection Agency's (EPA) proposed
revisions to the national air quality standards for particulate matter.
A large body of scientific evidence links exposure to particulate
matter a ubiquitous form of air pollution commonly referred to as soot
to serious health problems, including asthma, chronic bronchitis, heart
attack, and premature death. Under the Clean Air Act, EPA periodically
reviews the appropriate air quality level at which to set national
standards to protect the public against the health effects of
particulate matter. As you are aware, EPA proposed revisions to the
particulate matter standards in January 2006 and issued a draft
regulatory impact analysis of the revisions expected costs and
benefits.
EPA's estimates of the expected benefits from its air pollution
regulations have often been controversial, and the methods the agency
has used to prepare these estimates have been questioned. In 2000, at
the direction of the Senate Appropriations Committee, EPA asked the
National Academies (Academies) to evaluate EPA's overall methodology
for estimating the health benefits of proposed air regulations. In
2002, the Academies issued a report that made recommendations focusing
on con ducting more rigorous assessments of uncertainty, increasing the
transparency of how EPA estimates benefits, conducting more detailed
analyses of exposure, and estimating the benefits of each regulatory
option under consideration. My testimony summarizes the highlights of
our report being released today on the extent to which EPA applied the
recommendations made by the Academies to its January 2006 proposed
revisions to the particulate matter standards.\1\ Our report provides a
more detailed discussion of each recommendation, including whether and
how EPA applied it to the regulatory impact analysis on particulate
matter.
---------------------------------------------------------------------------
\1\ See GAO, Particulate Matter: EPA Has Started to Address the
National Academies Recommendations on EstimatingHealth Benefits, but
MoreProgress Is Needed, GAO-06-780 (Washington, DC: July 14, 2006).
---------------------------------------------------------------------------
summary
While the National Academies report generally supported EPA's
overall approach to estimating benefits, it included 34 recommendations
for improvements. EPA has begun to change the way it conducts and
presents its analyses of health benefits in response to the National
Academies? recommendations. In the case of the January 2006 proposed
rule on particulate matter standards, EPA applied, at least in part,
about two-thirds of the recommendations to its particulate matter
health benefit analysis; it applied 8 and partially applied 14 more.
For example, in applying the recommendations, EPA evaluated how
benefits might change given alternative assumptions and discussed
sources of uncertainty not included in the benefit estimates. In
addition, EPA applied an alternative technique for evaluating one
important source of uncertainty in its analysis the uncertainty
underlying the causal link between exposure to particulate matter and
premature death. Consistent with the National Academies recommendation
to assess uncertainty by developing ranges of estimates of benefits and
specifying the likelihood of attaining those levels of benefits, EPA
systematically gathered expert opinions about this link and developed
ranges reflecting the experts confidence in attaining reductions in
premature death expected from the proposed revisions. However, the
health benefit analysis did not assess how the benefit estimates would
vary in light of other key uncertainties as the Academies recommended.
Consequently, EPA's response represents a partial application of the
recommendation. Agency officials told us that ongoing research and
development efforts will allow EPA to gradually make more progress in
applying this and other recommendations to future analyses.
EPA did not apply the remaining 12 recommendations to the analysis,
such as the recommendation to evaluate the impact of using the
assumption that the components of particulate matter are equally toxic.
EPA officials viewed most of these 12 recommendations as relevant to
its health benefit analyses but noted that the agency was not ready to
apply specific recommendations because of, among other things, the need
to overcome technical challenges stemming from limitations in the state
of available science. For example, EPA did not believe that the state
of scientific knowledge on the relative toxicity of particulate matter
components was sufficiently developed to include it in the January 2006
regulatory impact analysis, but the agency is sponsoring research on
this issue.
background
EPA is required by the Clean Air Act to conduct reviews of the
National Ambient Air Quality Standards (NAAQS) for the six criteria
pollutants, including particulate matter, every 5 years to determine
whether the current standards are sufficient to protect public health,
with an adequate margin of safety. If EPA decides to revise the NAAQS,
the agency proposes changes to the standards and estimates the costs
and benefits expected from the revisions in an assessment called a
regulatory impact analysis. In January 2006, EPA prepared a regulatory
impact analysis for one such rule?particulate matter that presented
limited estimates of the costs and benefits expected to result from the
proposed particulate matter rule. EPA developed the estimates by, for
example, quantifying the changes in the number of deaths and illnesses
in five urban areas that are likely to result from the proposed rule.
The National Academies' 2002 report examined how EPA estimates the
health benefits of its proposed air regulations and emphasized the need
for EPA to account for uncertainties and maintain transparency in the
course of conducting benefit analyses. Identifying and accounting for
uncertainties in these analyses can help decision makers evaluate the
likelihood that certain regulatory decisions will achieve the estimated
benefits. Transparency is important because it enables the public and
relevant decision makers to see clearly how EPA arrived at its
estimates and conclusions. Many of the recommendations include
qualifying language indicating that it is reasonable to expect that
they can be applied in stages, over time; moreover, a number of the
recommendations are interrelated and, in some cases, overlapping. Soon
after the National Academies issued its report, EPA roughly
approximated the time and resource requirements to respond to the
recommendations, identifying those the agency could address within 2 or
3 years and those that would take longer. According to EPA officials,
the agency focused primarily on the numerous recommendations related to
analyzing uncertainty. As is discussed below, EPA applied some of these
recommendations to the particulate matter analysis.
EPA Applied Some, but Not All, of the National Academies
Recommendations to the Particulate Matter Regulatory Impact Analysis
EPA applied either wholly or in part approximately two-thirds of
the Academies recommendations in preparing its January 2006 particulate
matter regulatory impact analysis and continues to address the
recommendations through ongoing research and development. According to
EPA, the agency intends to address some of the remaining
recommendations in the final rule and has undertaken research and
development to address others.
Recommendations EPA Applied or Partially Applied to Its Particulate
Matter Health Benefit Analysis
The January 2006 regulatory impact analysis on particulate matter
represents a snapshot of an ongoing EPA effort to respond to the
National Academies recommendations on developing estimates of health
benefits for air pollution regulations. Specifically, the agency
applied, at least in part, approximately two-thirds of the
recommendations 8 were applied and 14 were partially applied by taking
steps toward conducting a more rigorous assessment of uncertainty by,
for example, evaluating the different assumptions about the link
between human exposure to particulate matter and health effects and
discussing sources of uncertainty not included in the benefit
estimates. According to EPA officials, the agency focused much of its
time and resources on the recommendations related to uncertainty. In
particular, one overarching recommendation suggests that EPA take steps
toward conducting a formal, comprehensive uncertainty analysis the
systematic application of mathematical techniques, such as Monte Carlo
simulation and include the uncertainty analysis in the regulatory
impact analysis to provide a more realistic depiction of the overall
uncertainty in EPA's estimates of the benefits.\2\
---------------------------------------------------------------------------
\2\ Monte Carlo simulation refers to a computer-based analysis that
uses probability distributions for key variables, selects random values
from each of the distributions simultaneously, and repeats the random
selection over and over. Rather than presenting a single outcome such
as the mostly likely or average scenario Monte Carlo simulations
produce a distribution of outcomes that reflect the probability
distributions of modeled uncertain variables.
---------------------------------------------------------------------------
Overall, the uncertainty recommendations call for EPA to determine
(1) which sources of uncertainties have the greatest effect on benefit
estimates and (2) the degree to which the uncertainties affect the
estimates by specifying a range of estimates and the likelihood of
attaining them. In response, EPA examined a key source of uncertainty
its assumption about the causal link between exposure to particulate
matter and premature death and presented a range of expected reductions
in death rates. EPA based these ranges on expert opinion systematically
gathered in a multiphased pilot project. The agency did not, however,
incorporate these ranges into its benefit estimates as the National
Academies had recommended.
Moreover, the Academies recommended that EPA's benefit analysis
reflect how the benefit estimates would vary in light of multiple
uncertainties. In addition to the uncertainty underlying the causal
link between exposure and premature death, other key uncertainties can
influence the estimates. For example, there is uncertainty about the
effects of the age and health status of people exposed to particulate
matter, the varying composition of particulate matter, and the
measurements of actual exposure to particulate matter. EPA's health
benefit analysis, however, does not account for these key uncertainties
by specifying a range of estimates and the likelihood of attaining
them. For these reasons, EPA's responses reflect a partial application
of the Academies recommendation.
In addition, the Academies recommended that EPA both continue to
conduct sensitivity analyses on sources of uncertainty and expand these
analyses. In the particulate matter regulatory impact analysis, EPA
included a new sensitivity analysis regarding assumptions about
thresholds, or levels below which those exposed to particulate matter
are not at risk of experiencing harmful effects. EPA has assumed no
threshold level exists that is, any exposure poses potential health
risks.\3\ Some experts have suggested that different thresholds may
exist, and the National Academies recommended that EPA determine how
changing its assumption that no threshold exists would influence the
estimates. The sensitivity analysis EPA provided in the regulatory
impact analysis examined how its estimates of expected health benefits
would change assuming varying thresholds.
---------------------------------------------------------------------------
\3\ Recent EPA analyses used the natural background concentrations
of particulate matter, rather than zero, for its assumption of no
threshold level. The National Academies supported the assumption of no
threshold level, but it recommended that EPA conduct a consistent and
transparent sensitivity analysis to consider various threshold levels.
---------------------------------------------------------------------------
In response to another recommendation by the National Academies,
EPA identified some of the sources of uncertainty that are not
reflected in its benefit estimates. For example, EPA's regulatory
impact analysis disclosed that its benefit estimates do not reflect the
uncertainty associated with future year projections of particulate
matter emissions. EPA presented a qualitative description about
emissions uncertainty, elaborating on technical reasons such as the
limited information about the effectiveness of particulate matter
control programs why the analysis likely underestimates future
emissions levels.
Recommendations EPA Did Not Apply to the Particulate Matter Analysis
EPA did not apply the remaining 12 recommendations to the analysis
for various reasons. Agency officials viewed most of these
recommendations as relevant to its health benefit analyses and, citing
the need for additional research and development, emphasized the
Agency's commitment to continue to respond to the recommendations. EPA
has undertaken research and development to respond to some of these
recommendations but, according to agency officials, did not apply them
to the analysis because the agency had not made sufficient progress.
For example, EPA is in the process of responding to a
recommendation involving the relative toxicity\4\ of components of
particulate matter, an emerging area of research that has the potential
to influence EPA's regulatory decisions in the future.\5\
Hypothetically, the agency could refine national air quality standards
to address the potentially varying health consequences associated with
different components of particulate matter. The National Academies
recommended that EPA strengthen its benefit analyses by evaluating a
range of alternative assumptions regarding relative toxicity and
incorporate these assumptions into sensitivity or uncertainty analyses
as more data become available.\6\ EPA did not believe the state of
scientific knowledge on relative toxicity was sufficiently developed at
the time it prepared the draft regulatory impact analysis to include
this kind of analysis. In a separate report issued in 2004, the
National Academies noted that technical challenges have impeded
research progress on relative toxicity but nonetheless identified this
issue as a priority research topic. The Clean Air Scientific Advisory
Committee also noted the need for more research and concluded in 2005
that not enough data are available to base the particulate matter
standards on composition. The Office of Management and Budget, however,
encouraged EPA in 2006 to conduct a sensitivity analysis on relative
toxicity and referred the agency to a sensitivity analysis on relative
toxicity funded by the European Commission.
---------------------------------------------------------------------------
\4\ Particulate matter is a highly complex mixture comprising
particles emitted directly from sources and particles formed through
atmospheric chemical reactions. Particles span many sizes and shapes
and consist of hundreds of different chemicals. EPA identifies the
major components of fine particulate matter as carbon, sulfate and
nitrate compounds, and crustal/metallic materials such as soil and ash.
\5\ Relative toxicity refers to the premise that different
components of particulate matter have different levels of potency
affecting premature mortality and illness. In the draft particulate
matter regulatory impact analysis, EPA assumed equivalent toxicity,
stating that while it is reasonable to expect that the potency of
components may vary across the numerous effect categories associated
with particulate matter, EPA's interpretation of scientific information
considered to date is that such information does not yet provide a
basis for quantification beyond using fine particle mass. EPA, Draft
Regulatory Impact Analysis for the PMPM-2.5 National Ambient
Air Quality Standards (Washington, DC, 2006), 3-21.
\6\ In the context of the National Academies recommendations, a
sensitivity analysis would assess how changes in one or more variables
affect the outcome, whereas a comprehensive or formal uncertainty
analysis evaluates the probability distributions of multiple variables.
---------------------------------------------------------------------------
We found that EPA is sponsoring research on the relative toxicity
of particulate matter components. For example, EPA is supporting long-
term research on this issue through its intramural research program and
is also funding research through its five Particulate Matter Research
Centers and the Health Effects Institute. In addition, an EPA
contractor has begun to investigate methods for conducting a formal
analysis that would consider sources of uncertainty, including relative
toxicity. To date, the contractor has created a model to assess whether
and how much these sources of uncertainty may affect benefit estimates
in one urban area. Agency officials told us, however, that this work
was not sufficiently developed to include in the final particulate
matter analysis, which it says will present benefits on a national
scale.
Another recommendation that EPA did not apply to the particulate
matter analysis focused on assessing the uncertainty of particulate
matter emissions. The National Academies recommended that EPA conduct a
formal analysis to characterize the uncertainty of its emissions
estimates, which serve as the basis for its benefit estimates.\7\ While
the agency is investigating ways to assess or characterize this
uncertainty, EPA did not conduct a formal uncertainty analysis for
particulate matter emissions for the draft regulatory impact analysis
because of data limitations. These limitations stem largely from the
source of emissions data, the National Emissions Inventory,\8\ an
amalgamation of data from a variety of entities, including state and
local air agencies, tribes, and industry. According to EPA, these
entities use different methods to collect data, which have different
implications for how to characterize the uncertainty. EPA officials
stated that the agency needs much more time to address this data
limitation and to resolve other technical challenges of such an
analysis. While the final particulate matter analysis will not include
a formal assessment of uncertainty about emissions levels, EPA
officials noted that the final analysis will demonstrate steps toward
this recommendation by presenting emissions data according to the level
emitted by the different kinds of sources, such as utilities, cars, and
trucks.
---------------------------------------------------------------------------
\7\ Because the precise levels of total emissions are not knowable
but rather are approximations based on a sample of measurements, there
is uncertainty about the true quantity of emissions.
\8\ EPA compiles the National Emissions Inventory, a national
database of air emissions data that includes estimates of annual
emissions, by source, of air pollutants in each area of the country on
an annual basis.
---------------------------------------------------------------------------
Finally, EPA did not apply a recommendation concerning the
transparency of its benefit estimation process to the particulate
matter analysis. Specifically, the National Academies recommended that
EPA clearly summarize the key elements of the benefit analysis in an
executive summary that includes a table that lists and briefly
describes the regulatory options for which EPA estimated the benefits,
the assumptions that had a substantial impact on the benefit estimates,
and the health benefits evaluated. EPA did not, however, present a
summary table as called for by the recommendation or summarize the
benefits in the executive summary. EPA stated in the regulatory impact
analysis that the agency decided not to present the benefit estimates
in the executive summary because they were too uncertain. Agency
officials told us that the agency could not resolve some significant
data limitations before issuing the draft regulatory impact analysis in
January 2006 but that EPA has resolved some of these data challenges.
For example, EPA officials said they have obtained more robust data on
anticipated strategies for reducing emissions, which will affect the
estimates of benefits. The officials also said that EPA intends to
include in the executive summary of the regulatory impact analysis
supporting the final rule a summary table that describes key analytical
information.
concluding observations
While EPA officials said that the final regulatory impact analysis
on particulate matter will reflect further responsiveness to the
Academies recommendations, continued commitment and dedication of
resources will be needed if EPA is to fully implement the improvements
recommended by the National Academies. In particular, the agency will
need to ensure that it allocates resources to needed research on
emerging issues, such as the relative toxicity of particulate matter
components, and to assessing which sources of uncertainty have the
greatest influence on benefit estimates. The uncertainty of the
agency's estimates of health benefits in the draft regulatory impact
analysis for particulate matter underscores the importance of
uncertainty analysis that can enable decision makers and the public to
better evaluate the basis for EPA's air regulations. While EPA
officials said they expect to reduce the uncertainties associated with
the health benefit estimates in the final particulate matter analysis,
a robust uncertainty analysis of the remaining uncertainties will
nonetheless be important for decision makers and the public to
understand the likelihood of attaining the estimated health benefits.
Mr. Chairman, this concludes my prepared statement. I would be
happy to respond to any questions that you or members of the committee
may have.
______
Responses by John Stephenson to Additional Questions
from Senator Voinovich.
Question 1. Mr. Stephenson, your testimony states ``The National
Academies'' 2002 report examined how EPA estimates the health benefits
of its proposed air regulations and emphasized the need for EPA to
account for uncertainties and maintain transparency in the course of
conducting benefit analyses. I agree that transparency is very
important. However, right now, EPA is simply telling us that we need to
wait until the final rule to see a more complete RIA, how they
addressed more of these recommendations, and the inclusion of new
science. In your opinion, has EPA ``maintained transparency'' through
this process with the PM standards?
Response. EPA has maintained transparency to the extent feasible
under federal rulemaking procedures, which include collecting and
reviewing public comments on the proposed rule and submitting the final
rule package to the White House Office of Management and Budget for
review at least 45 days before the EPA Administrator signs the rule.
Proposed and final rules can differ substantially for a number of
reasons, including changes made in response to public comments on the
proposed rule and the availability of new or more complete data.
Therefore, in developing and supporting both proposed and final rules,
it is important for agencies to be transparent that is, to provide
decision makers and the public with clear and relevant information
about the data, assumptions, methodologies, uncertainties, etc.,
underlying its regulatory impact analyses. As you are aware, the
Congressional Review Act of 1996 established an expedited process under
which Congress may disapprove a broad range of regulatory rules issued
by Federal agencies by enacting a joint resolution of disapproval
within 60 days after receiving the rule. The rule would not go into
effect without subsequent statutory authorization.
To date we have not identified major shortcomings concerning the
transparency of EPA's current rulemaking process for particulate matter
standards. For example, in the January 2006 regulatory impact analysis
for proposed rule, EPA highlighted the fact that the benefits estimates
were based only on five cities and were highly uncertain. According to
EPA officials, the agency could not resolve data limitations in time to
meet the court-issued deadline for the proposed changes. As you know,
EPA plans to provide national estimates of the health benefits in the
final rule using more robust data that it has subsequently been able to
obtain. Thus, the regulatory impact analyses supporting the proposed
and final rules are expected to be significantly different. EPA can
maintain transparency in this rulemaking process by clearly discussing
in the final regulatory impact analysis due in September 2006 the
changes in its approach to estimating the benefits and the Agency's
rationale for doing so and by clearly identifying key data,
assumptions, methodologies, and uncertainties.
In terms of transparency and the science supporting the rulemaking
efforts, EPA made a commitment in the proposed rule to conduct a review
and assessment of studies published after the completion of the
scientific review of health effects linked to particulate matter
emissions. EPA has recently made information about its review of new
science available to the public; see EPA, Provisional Assessment of
Recent Studies on Health Effects of Particulate Matter Exposure
(Research Triangle Park, N.C., 2006). The July 2006 provisional
assessment presents EPA's approach to reviewing new studies and
summarizes the findings of these studies, including those that
evaluated the links between particulate matter components and adverse
health effects.
Question 2. Mr. Stephenson, please elaborate on this very important
statement that you make at the end of your testimony (quote): "While
EPA officials said they expect to reduce the uncertainties associated
with the health benefit estimates in the final particulate matter
analysis, a robust uncertainty analysis of the remaining uncertainties
will nonetheless be important for decision makers and the public to
understand the likelihood of attaining the estimated health benefits.''
Response. In estimating potential health benefits stemming from
regulatory actions, some level of uncertainty is unavoidable, in part
because the scientific information used to develop estimates, such as
the inventory of particulate matter emissions, will never be perfect or
complete. According to the Academies, high uncertainty does not imply
that action to promote or protect public health should be delayed, but
rather that a comprehensive and rigorous assessment of uncertainty
would improve policy decisions. Many of the recommendations of the
National Academies to EPA are aimed at the agency developing a more
robust uncertainty analysis by providing comprehensive, quantitative
information about the uncertainties underlying the estimates of health
benefits. A quantitative uncertainty analysis can, among other things,
identify which sources of uncertainties have the greatest effect on the
benefit estimates and assess the degree to which the uncertainties
affect the estimates by specifying a range of estimates and the
likelihood of attaining them. While EPA has taken steps toward a more
robust uncertainty analysis, the agency has not assessed uncertainty in
the quantitative manner
recommended by the National Academies. GAO believes it is important
for EPA to continue to strengthen its uncertainty analysis. In
evaluating estimates of health benefits, decision makers and the public
can better assess the likelihood of achieving such benefits if they are
provided with qualitative and quantitative information about the
underlying uncertainties.
Question 3. Mr. Stephenson, in your testimony (quote): ``We note
that continued commitment and dedication of resources will be needed if
EPA is to fully implement the improvements recommended by the National
Academies.'' However, EPA's 2006 budget proposes to reduce funding for
the State and Local Air Quality Management Program which is used by
states and localities to conduct monitoring among other things. How
would this budget cut impact EPA's ability to address more of these
recommendations?
Response. The proposed budget cuts could adversely impact EPA's
ability to address more of the Academies recommendations, such as those
concerning particulate matter emissions data. Emissions estimates are a
key element in the analysis of potential health benefits, and EPA
relies largely on state and local entities emissions monitoring data to
develop estimates of emissions on a national scale. Thus, any
reductions in emissions monitoring could affect EPA's ability to
develop meaningful nationwide emissions estimates, increasing, for
example, the uncertainty of such estimates. Moreover, as noted in our
report, EPA has initiated long-term research and development to
understand the relative toxicity of particulate matter components and
to better evaluate the uncertainty of estimates of emissions. A
reduction in monitoring data would likely limit EPA's progress to
address these areas of research.
__________
Statement of George Thurston, Associate Professor, New York University,
School of Medicine, Department of Environmental Medicine
I am George D. Thurston, a tenured Associate Professor of
environmental Medicine at the New York University (NYU) School of
medicine. In addition, I served as the Deputy Director of the NYU-EPA
Particulate Matter (PM) Health Research Center for the past 4 years. My
scientific research involves the investigation of the human health
effects of air pollution. In this testimony, I will primarily address
three factors that need to be considered in the EPA's proposed
revisions to the particulate matter air quality standards. First, I
will address the fact that we are now far more certain of the adverse
impacts and biological mechanisms of PM health effects: most of the
uncertainties raised at the time of the initial setting of the
PM2.5 standard are far better understood. Second, I will
document that reducing ambient PM levels can and do result in
significant reductions in the mortality risk associated with this
pollutant. Finally, I will show that the adverse health impacts of PM
air pollution extend below the current PM2.5 standard, and
that there is, therefore, a public health imperative to reduce the fine
particle (PM2.5) annual standard below 15 wg/m3,
consistent with the advice of the U.S. EPA's Clean Air Science Advisory
Committee (CASAC).
Despite progress over the last few decades, Americans are still
suffering from the adverse health effects of air pollution. The adverse
health consequences of breathing air pollution are severe and well
documented in the published medical and scientific literature. Over the
past few decades, medical researchers examining air pollution and
public health, including myself, have shown that air pollution is
associated with a host of serious adverse human health effects,
including: asthma attacks, heart attacks, hospital admissions, adverse
birth outcomes, and premature death.
PM is one of the air pollutants most carefully studied in the last
decade. Small particles can bypass the defensive mechanisms of the
lung, and become lodged deep in the lung where they can cause a variety
of health problems. Indeed, the latest evidence indicates that
exposures cannot only cause respiratory damage, but also cardiac
effects, including heart attacks. Moreover, long-term exposure to fine
particles increases the risk of death, and has been estimated to take
years from the life expectancy of people living in the most polluted
cities, relative to those living in cleaner cities (Brunekreef, 1997).
The State of the science on particulate matter and health was
thoroughly reviewed in the recently released U.S. EPA Criteria Document
for Particulate Matter (U.S. EPA, 2004). Since the PM2.5
standard was last set in 1997, more than 100 new published studies,
taken together, collectively confirm the relationship between
PM2.5 pollution and severe adverse human health effects. In
the process, this new research has eliminated many of the doubts that
were raised in the past regarding the causality and size of the PM-
health effects relationships, and has now provided plausible biological
mechanisms for the serious impacts associated with PM exposure. As
outlined in Figure 1, the PM research funded since the setting of the
last PM2.5 standard has collectively shown the existence of
numerous biological pathways capable of causing damage in the human
heart, lung, nervous system, and circulatory system consistent with the
health impacts found by the PM epidemiology studies upon which the
PM2.5 standard was set.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
At the time of the last PM standard revisions, the largest landmark
studies on particulate matter and death were the Harvard Six Cities
Study (Dockery et al, 1993) and the American Cancer Society Study (Pope
et al, 1995). The American Cancer Society study examined half a million
people in over 150 metropolitan areas throughout the United States,
finding a 17 percent greater risk of mortality between the city with
the least PM and the city wit the highest levels of this particulate
pollution. The results of these studies were challenged by industry,
resulting in an independent reanalysis by the Health Effects Institute
(HEI)--an organization funded by both industry and Government. The
results of the HEI re-analyses have now confirmed the associations
found by the original investigators, increasing our confidence in the
results of these two already highly regarded studies of PM mortality.
Since the setting of the original PM2.5 standard, more
recent follow-up analyses of the Harvard and ACS studies have now
considered longer records of time, and have confirmed and extended the
conclusions from these two major studies. An extended analysis of the
Harvard Six Cities Study (to include follow-up through 1990) has now
shown that reductions in long-term ambient PM pollution results in
concomitant reductions in the health risks associated with PM. As shown
in Figure 2, large reductions in PM at four of the Harvard cities have
resulted in likewise large reductions in the relative risk (RR) of
mortality in those cities: Steubenville, OH(S), Harriman, TN(H), St.
Louis, MO(L), and Watertown, MA(W). Other published studies have
similarly found indications that reductions in ambient PM are
associated with reduced mortality risk (e.g., Clancy et al., 2002).
Thus, although we still carry very large health risks in the United
States from our present levels of PM air pollution, amounting to tens
of thousands of premature deaths per year, and although we still have a
long way to go to have what can be called ``clean air'', recent
research shows that the lowering of PM levels in the air is an
effective way to improve public health.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
In addition, a recent National Institute of environmental Health
Sciences (NIEHS)-funded extension of the ACS study, of which I was
Principal Investigator, strengthens the original conclusions of the ACM
study and, importantly, now links increased risk of lung cancer to
long-term exposure to particulate matter (Pope et al, 2002). As seen in
Figure 3, this recent JAMA study also clearly indicates that the risks
from PM2.5 exposure extend below 15 ug/m3, supporting a
reduction in the annual PM2.5 standard at this time,
consistent with the advice of the EPA's Clean Air Scientific Advisory
Committee (Henderson, 2006).
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
In conclusion, since it was the level of uncertainty about PM
biological mechanisms and effects at lower concentrations than 15 wg/m3
that limited the standard to that level in 1997 (and not some specific
acceptable level of health risk from PM), and since new sound
scientific studies have greatly reduced or resolved those
uncertainties, then concern about the health of the public clearly
indicates that the long-term PMPM2.5 standard should now be
reduced below 15 ug/m3, consistent with the advice of CASAC.
Thank you for the opportunity to testify on this important issue.
______
Response by George Thurston to an Additional Question from
Senator Voinovich
Question. Dr. Thurston, at the end of your testimony, you state the
annual standard was limited to 15 not because of (quote): ``some
specific acceptable level of health risk from PM.''
Am I correct that you are saying that you agree with Dr. Smith that
the risk level identified in 1997 has gone down? I know that you do not
think this should be the basis of setting the standard, but I am just
looking for a yes or no answer on the level of risk.
Response. No, I do not agree with Dr. Smith. As I noted in my
verbal testimony, the consensus ``best-estimates'' for the mortality
effects of long-term exposure to PM have actually increased in recent
years, relative to what the U.S. EPA has been assuming in their risk
analyses. Indeed, the Pope et al. (2002) overall mortality effect
estimate, as derived from the American Cancer Society Cancer Prevention
II study cohort, is now thought to be an underestimate of the actual
risk for the general U.S. population. In fact, preliminary results from
the EPA's recent PM Expert Elicitation evaluation of PM risk indicate
that the central effect estimate, as derived from a distillation of all
available studies by experts, yield a PM2.5 mortality effect
(some 1 percent per ug/m3 PM2.5) that is roughly
two thirds higher than given by the Pope et al. (2002) study. This
further indicates that
lowering the long-term PM2.5 standard below 15 ug/
m3 would yield significant public health benefits when
implemented across the United States.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]