Environmental and Health Risks and Public Policy
This page intentionally left blank 
Environmental
Health Risks and
Public Policy
Decision Making in Free Societies
David V. Bates
UBC PRESS
Vancouver
Copyright © 1994 by the University of Washington Press
Printed in the United States of America
Published simultaneously in the United States of America by the
University of Washington Press, P.O. Box 50096,
Seattle, WA 98145-5096
All rights reserved. No part of this publication may be
reproduced or transmitted in any form or by any means, electronic
or mechanical, including photocopy, recording, or any information
storage or retrieval system, without permission in writing
from the publisher.
Canadian Cataloguing in Publication Data
Bates, David V. (David Vincent), 1922-
Environmental health risks and public policy
(Jessie and John Danz lectures)
Includes bibliographical references and index.
ISBN 0-7748-0506-4
1. Environmental health—Planning. 2. Environmental policy.
I. Title. II. Series.
RA566.B38 1994 363.7'0525 C94-910413-2
The paper used in this publication meets the minimum requirements
of American National Standard for Information Sciences—
Permanence of Paper for Printed Library Materials,
ANSI Z39.48-1984.
Contents
Acknowledgments xi
1
Introduction 3
2
Setting the Stage: Critical Risks 6
3
Mandated Science:
Major Issues in Health and Public Policy 57
4
A Survival Kit for the Environmental Jungle 78
5
Conclusions 89
References 105
Index 113
vii
This page intentionally left blank 
Figures
2.1 Air pollution levels and excess mortality in London in
December 1952 11
2.2 Cross-sectional study of lung function in British postal
workers in 1965 14
2.3 Hospital admissions for asthma in children in southern
Ontario, 1974-1980 17
2.4 Relative risk of death in Philadelphia in relation to
particulates 18
2.5 Relative risk of mortality in Philadelphia associated with a
100 microgram/m
3
 increase in total suspended particulates 19
2.6 Changing patterns of lung cancer mortality and cigarette
smoking in women in Britain 25
2.7 Asbestos production and the evolution of
medical knowledge 28
2.8 Annual age-specific discharge rates of asbestosis among
white males in New Jersey, 1979-1986 32
2.9 The risk of lung cancer in different occupations and with
different fibers 35
2.10 Lead concentration profile in Greenland snow strata A.D.
1750 to A.D. 1970 37
2.11 Effect of a 5 percent mean shift in IQ in a population of 100
million 40
2.12 Prenatal lead exposure vs. mental development 41
2.13 Pathways of lead exposure 42
IX
FIGURES
2.14 Decreases in blood lead and use of lead in gasoline in the
United States from 1976 to 1980 43
2.15 The electromagnetic spectrum 45
2.16 Leukemia risks among electrical workers 47
2.17 Acute myeloid leukemia in electrical workers 48
3.1 "Mandated Science" 58
5.1 Percentage of severe mental retardation among those
exposed in utero by dose and gestational age at Hiroshima
and Nagasaki 102
x
Acknowledgments
I wish to begin by thanking the University of Washington for the
honor of the invitation to give the 1993 Jessie and John Danz Lectures
on which this book is based. The Danz Lectures are concerned with
"the impact of science and philosophy on man's perception of a ra-
tional universe." The notes sent to me by the university record details
of John Danz's remarkable achievements and of his generous endow-
ment. It is stated that he was an independent and often unorthodox
thinker, with deep faith and convictions though a member of no estab-
lished church, neither a cynic nor an iconoclast, but nevertheless an
idealist.
As a Unitarian and a former dean of a medical school (in which post
optimism is essential for survival), I hope to approach in John Danz's
spirit the five public health questions I have chosen to address.
Over the past forty years, so many people have contributed
thoughts and ideas that have influenced me on the topics I have
addressed that it would be impossible to acknowledge all of them. A
partial listing of those to whom I am indebted would include Mort
Lippmann, Mark Utell, Bernard Goldstein, Paul Lioy, Jack Spengler,
Devra Davis, Scott Zeger, John Petkau, Geoffrey Rose, the late Donald
Reid, the late Ronald Christie, Terry Anderson, Ronnie Sizto, Darrell
Roberts, Bob Caton, Clyde Hertzmann, Peter Warren, Hans-Peter
Witschi, David Rall, Arthur Upton, Larry Green, Fred Bass, Margaret
Becklake, Charles Chapman, Bill McDonnell, Carl Hayes, David Mc-
Kee, Les Grant, Andrew Churg, Corbett McDonald, Joel Schwartz,
Bob Frank, Ben Ferris, Jim Whittenberger, Curtis Moore, Jim Pitts,
xi
xii ACKNOWLEDGMENTS
Tim Oke, Doug Dockery, Frank Speizer, Ira Tager, John Samet, Dun-
can Thomas, Sverre Vedal, Phil Landrigan, the late Irving Selikoff,
Paul Brodeur, Fred Lipfert, Stephen Richman, Ed Wituschek, Jack
Basuk, Ken Hare, Jim Ham, Gordon Butler, and Liora Salter.
I can recall stimulating discussions with members of the Epidemiol-
ogy 2 study section of the National Institutes of Health, on which I
served for four years; with members of the Science Council of Can-
ada, of which I was a member for six years; with my fellow commis-
sioners, Jim Murray, and the late Walter Raudsepp, on the Royal
Commission on Uranium Mining in British Columbia; with Gene Mat-
anoski, Pat Buffler, Donald Pierce, Barry Wilson, Kathleen Conway,
Richard Wilson, Clark Heath, and other members of the committee
convened by the Science Advisory Board of the U.S. Environmental
Protection Agency to review data on electromagnetic fields; and with
Don Hornig, John Bailar, Gil Omenn, and members of the Board of
Environmental Studies of the National Academy of Sciences during
my service on it.
I am also much indebted to the two reviewers to whom the Univer-
sity of Washington Press forwarded the first draft of the manuscript.
Dr. Gil Omenn was one of these; the other was anonymous. Their
questions and constructive suggestions led to important revisions in
the text, with necessary expansion of some sections.
Needless to say, none of these individuals or organizations should
accept any responsibility for the views I have expressed.
March 1993
Environmental Health Risks and Public Policy
This page intentionally left blank 
1
Introduction
World War II was characterized by unparalleled suffering of the civil-
ian population. Apart from the millions in military service who died
in combat, millions of people were the targets of intensive air attacks
in Rotterdam, London, Coventry, Dresden, Tokyo, and of course Hiro-
shima and Nagasaki; and millions of innocent civilians were sum-
marily executed or gassed. After 1945, the world struggled back to
normality from such cataclysmic events and the survivors turned
themselves to the process of repair.
As a corporal in the Home Guard, I had been a witness in Kent of
the Battle of Britain, and for most of the war was a medical student in
or around London. I was in Hiroshima with the British army in March
1946. From 1948 to 1956 I was learning my specialty of lung physiol-
ogy and lung disease, and, with others, I was looking after a special
emphysema clinic at St. Bartholomew's Hospital in the center of the
City of London. I date my interest in air pollution from the London
episode of fog and severe air pollution in December 1952. This is the
background to the chapters that follow. I was dissuaded from making
them purely autobiographical by a note in a recent New York Review of
Books that said, "Autobiography is, of course, the highest form of
fiction."
It slowly became apparent after the war that there was much unfin-
ished business that demanded attention. The four thousand excess
deaths in London in December 1952 were a preliminary warning that
air pollution could no longer be neglected. The public realized this
earlier than the government of the day. The enormous acceleration in
3
4 INTRODUCTION
the rate of cigarette smoking, which had occurred universally during
the stress of the war, was not yet appreciated as the health disaster it
would ultimately represent. Asbestos was recognized as a health risk
to a few employees, and during the war the U.S. navy had issued
some notes on the safe use of the material, but there is little evidence
that any precautions were taken in its use (1). I saw my first case of
fatal asbestosis from the Cape Asbestos plant in East Ham in 1949.
Lead poisoning, occasionally seen in small children exposed to peel-
ing, lead-containing paint, or in employees casting lead type for print-
ing, was mentioned in medical school, but I do not recall seeing a
recognized case of it. Any possible hazard of electromagnetic field
exposure was not recognized.
The principal reason why exposures to these hazards have become
a major focus of concern over the forty years since 1952 is that very
large numbers of people have become exposed to some risk from
them. Thus the catastrophic increase in cigarette smoking, which has
only now begun to decline in some countries, led to an epidemic of
lung cancer, emphysema, and heart disease. Air pollution, recog-
nized as a possible problem locally when temperature inversions oc-
curred, now affects very large numbers of people—in 1988, 86 million
people in the United States were living in areas in which the ozone
standard, based on health protection, was exceeded (2). The collapse
of totalitarian regimes in eastern Europe revealed the presence of
gross environmental degradation (3), much of it related to air pollu-
tion. At the present time in the Third World, acute respiratory disease
is now the second commonest cause of death in children under five
(4), and severe indoor and outdoor air contamination undoubtedly
plays a major role in this (5).
In New York City, two-thirds of the 800,000 buildings apparently
contain asbestos, and in many of these it is in poor condition (6). The
widespread use of lead in gasoline everywhere (by 1967 the annual
production of tetraethyl and tetramethyl lead in the United States for
use in gasoline was 685 million pounds [7]) resulted in a fourfold
increase in the lead content of snow strata samples in Greenland
between 1940 and 1960 (see figure 2.10). The development of high
voltage electrical transmission lines, together with the multiplicity of
domestic appliances using electricity, must have greatly increased the
general population's exposure to electromagnetic fields.
These observations indicate clearly that present concerns are not
INTRODUCTION 5
attributable to the energetic protestations of a lunatic environmental
fringe, nor to the persistent obsession of the media with them. These
concerns come about because of the very widespread dissemination
of such hazards. We are generally aware of health risks because of the
activities of the media, but it must be noted that media attention is
episodic and much greater in relation to some of the hazards than to
others. In addition, Rose has pointed out that a low level of exposure
for a very large number of people may constitute a more serious
public health problem than a higher level of exposure for a small
group (8, 92), and this has to be kept constantly in mind.
In Chapter 2, I sketch each of these five hazards and the evolution
of knowledge concerning them. These summaries are necessarily
brief, but they are designed to illustrate the data that lie behind differ-
ent judgments concerning each of them. I assess the role of the media,
of scientists, of industry, of legislators, and of the courts in relation to
each of the five, and some differences are noted between the free
societies of Britain, Canada, and the United States in respect of them.
Chapter 3 provides a detailed evaluation of five different problems
that underlie contemporary policy decisions. Chapter 4 illustrates
some common mistakes and misunderstandings, and examines the
ways whereby a decision is made to protect the public health in a free
society. In Chapter 5,1 attempt to draw some conclusions as a guide to
the future.
It is immediately apparent that such a review would require
knowledge not only of lung disease, which I can claim to have, but
also of child psychology (important in relation to the effects of lead),
atmospheric chemistry, biostatistics, risk assessment, epidemiology,
carcinogenesis, social sciences, political science, economics, and
jurisprudence—fields in which I cannot claim expertise. However, it
is my hope that I command a sufficient "defensive" knowledge of
them that I can draw some useful general lessons.
Setting the Stage: Critical Risks
Table 2.1 lists sequentially the five hazards under discussion and com-
pares them under twelve headings. This constitutes the underlying
structure of the following discussion of each hazard.
It must be realized at the outset that each of these five topics is
encompassed by an enormous bibliography. Many books have been
devoted to each one separately. The summaries that follow are not
designed to cover all the many aspects of the five hazards, but to
bring out aspects of them that are important in any discussion of the
public and legislative response that has been engendered.
Air Pollution
1952 was the year of the London smog episode, but concern about air
pollution in that city stretched back almost four hundred years. It was
often inferred, but not demonstrated, that air quality must be affect-
ing health of the population, and there had been certain ominous
signs. One of these was the occurrence in London in December of
1873 of an episode in which cattle died from the effects of a fog. This is
described in a letter and an editorial in The Veterinarian in January 1874
(9). A Mr. Priestman noted that he was called to attend sick animals
on the previous Tuesday and Wednesday during a heavy fog coinci-
dent with Cattle Show week in London. "On Tuesday, the first day of
the fog, as early as eleven o'clock in the morning several animals were
marked as affected with difficult breathing." By that evening "the
majority of cattle in the Hall showed evidence of suffering. Sheep and
2
6
SETTING THE STAGE: CRITICAL RISKS 7
pigs not affected. . . . On Wednesday, there was very little improve-
ment in the atmospheric condition, and a large proportion of the
cattle were so much distressed that it became necessary to do some-
thing to satisfy the exhibitors, if not to relieve the animals." By Thurs-
day things were better. Most of the animals were slaughtered, and
"the post-mortem appearances were indicative of bronchitis; the mu-
cous membrane of the smaller bronchial tubes was inflamed, and
there was present the lobular congestion and emphysema which be-
long to that disease." Many of the animals were noted to be febrile.
An editorial on page 32 of the same issue noted: "If any one, a few
weeks ago, had suggested the possibility of a London fog doing seri-
ous damage to cattle, or other animals, submitted to its influence, he
would have been looked upon as supplying in himself a melancholy
instance of intellectual fogginess." Such was the fate of Cassandra.
The great painter Claude Monet spent the winter of 1901 in Lon-
don, and captured the foggy days in his paintings of Waterloo Bridge
and Westminster. In an interview (10), he noted, "The fog in London
assumes all sorts of colours; there are black, brown, yellow, green,
purple fogs, and the interest in painting is to get the objects as seen
through all these fogs. My practised eye has found that objects
change in appearance in a London fog more and quicker than in any
other atmosphere, and the difficulty is to get every change down on
canvas." Another commentator (10) quoted him as remarking that
London was much more beautiful than the English countryside, but
only in winter, when the fogs gave it a magnificent sense of space by
obscuring detail. He painted more than a hundred canvases of Water-
loo Bridge, Charing Cross Bridge, and the Houses of Parliament; it
has been noted that (10) "the paintings show the foggy river and city
surrounding it under gray skies or when the sun penetrated the fog or
illuminated smoke." Contemporary coal-burning air pollution played
a major part in producing these effects, which Monet so successfully
caught in his paintings.
The 1952 episode in London occurred between December 5 and 8.
As the later official government report on the episode (11) noted,
"The fog was notable for its density and its duration and an important
feature was the almost complete absence of remissions, either in den-
sity or in temperature, during the four days. In a city traditionally
notorious for its fogs, there was general agreement on its exceptional
severity on this occasion." An appendix to that report noted that the
Table 2.1. Summary of Five Environmental Hazards
Year*
Significant
Outcomes
Strength of
Epidemiol Data
Low Exposures
Exposure
Measurement
Air
Pollution
1952
Episodes: long-
term effects
Indirect indices;
variable
associations
Chronic effects
Not precise
Cigarettes
1963
Lung cancer;
COPD heart
disease
Very strong
Environmental
tobacco smoke
Fair in smokers;
passive dose
difficult
Asbestos
1975
Mesothelioma; lung
cancer (asbestosis)
Very strong
As in buildings;
risk is controversial
Very difficult in
most cases
Lead
1980
Children: cognitive
defect; adults:
hypertension
Difficult outcome
to evaluate;
moderately strong
?Effects with blood
levels 15-30 dL/L
Deduced from
biomarkers
Electromagnetic
Fields
1986
Children: leukemia;
adults: cancers
Associations:
difficult to assess
causal relationship
No dose response
data
Very difficult;
deduced from
surrogates
such as wiring
configuration
Role of Acute
Episodes
Type of Risk
Public Perception
of Risk vs.
Numerical Estimate
Industry Involved
Media Coverage
Legislative
Response
Involvement of
Courts/Lawyers
+
Involuntary
?
Automobile;
electric power
Intermittent (strong
in Los Angeles)
Yes
Absent
Nil
Voluntary/
involuntary
Low
Tobacco;
advertising
Intermittent
Yes
Not major
Nil
Involuntary
High
Mining; insulation
secondary industry
Intermittent
Yes
Dominant
Nil
Involuntary
Low
Petrochemical;
automobile
Not extensive
Yes
Absent; ?starting
Nil
Involuntary
?High
Electric utilities
Recently strong
Not yet
Increasing
* Approximate year of first major public concern.
10 SETTING THE STAGE: CRITICAL RISKS
Smithfield Club's show was being held at Earl's Court in London over
those days, and sixty cattle needed veterinary attention on account of
respiratory symptoms, with one dying and twelve having to be
slaughtered. The writers of the report referred to the similar episode
in 1873. The report noted the episodes of air pollution from industrial
sources with human mortality in the Meuse Valley in 1931, and in
Donora in Pennsylvania in 1948.
In the case of the London episode, it was realized within a few days
that there had been a major increase in daily mortality. In the House
of Commons, a written question was tabled soon after the episode,
and a written reply was given on 18 December 1952. In this, the
Health Minister, Iain MacLeod, stated that the deaths in greater Lon-
don during the week ending December 13 were 4, 703, compared with
1,852 in the corresponding week of 1951. Hospital admissions for the
week ending on December 12 were 2,007 compared with 917 in the
comparable week in 1951. He stated, "The cold weather had already
caused some increase, but a large part of these increases must be
attributed to the fog." On the same day, a question was asked, "Is
there enough evidence to justify more energetic research into the
harmful constituents of the air in towns?" The mortality and air pollu-
tion data in the episode are shown in figure 2.1.
By 23 January 1953, more persistent questions were being asked.
Harold MacMillan, then the Minister of Housing and Local Govern-
ment, parried these, but his attitude must have seemed too little
concerned, since a member asked, "Does the Minister not appreciate
that last month, in Greater London alone, there were literally more
people choked to death by air pollution than were killed on the roads
in the whole country in 1952? Why is a public inquiry not being held?"
Mr. MacMillan stated: "I am not satisfied that further general legisla-
tion is needed at present. I am however keeping this aspect of the
matter in hand."
The newspapers continued their criticism of government inaction. I
can recall a meeting occurring at this time between Professor Ronald
Christie, who headed the Professorial Medical Unit at St. Bartholo-
mew's Hospital, and a couple of senior civil servants from the Minis-
try of Health. They were trying to draft a reassuring response to be
given in the House of Commons to a question on action following the
episode. One of the civil servants read out a sentence he had been
drafting that said something about a "new research initiative," which
SETTING THE STAGE: CRITICAL RISKS 11
Figure 2.1. Air pollution levels and excess mortality in London in December
1952. The smoke level of more than 1500 micrograms/m
3
 reduced visibility to a
few yards. One SO
2
 monitor recorded levels of more than 1.00 part per million
(ppm) at the height of the episode. The daily death rate did not return to its
normal level of about 270 deaths per day for a further two weeks. Redrawn
from reference 14.
had been taken "at one of London's oldest hospitals." Ronald Christie
said, "That's a bonnie phrase and has the added advantage of being
true." His candor scandalized the bureaucrats.
Rereading the exchanges in the House of Commons, one is im-
pressed by the general lack of discussion as to which constituents of
the fog might have been harmful, and where they came from. As a
result of a continuing sense of public outrage at the inaction of the
government (as expressed by the newspapers), the government fi-
12 SETTING THE STAGE: CRITICAL RISKS
nally established the Beaver Committee in July 1953; and it was this
committee's report that led to the clean air legislation in 1956. Lord
Ashby, reviewing these events (12), noted the initial apathy of the
government, and made the important comment that the Beaver Com-
mittee succeeded in proposing successful legislation that enabled lo-
cal authorities to forbid open coal burning because alternative smoke-
less fuels were available for immediate use.
In the United States, a similar growth in public awareness of air
pollution's health effects began to lead to legislation to curb its worst
constituents. Professor Joel Tarr (13), of the Carnegie-Mellon Institute,
has documented in detail the steps taken in Pittsburgh that led to
passage in 1946 of a city ordinance against air pollution. A group of
citizens, including a physician, started a campaign in 1941 for clean air
legislation. Making little headway by public argument, they decided
that the city council would have to be replaced; by 1946 they had
succeeded both in doing this and in passing an air pollution ordi-
nance. They relied on stories about the adverse impacts on public
health in their campaign, but no scientific evidence was available that
these were actually occurring. Dr. Hope Alexander, the physician
member of the citizen group, was Director of the Pittsburgh Depart-
ment of Health. He was quoted as saying, in 1941, "The medical
profession is in accord with the statement that smoke is a health
menace of major proportions. . . . For the present, however, there are
few scientific facts to definitely establish a case of cause and effect."
Various groups, including the League of Women Voters and several
women's clubs, the board of trade of a wealthy Pittsburgh neighbor-
hood, and the members of the activist group, all took a prominent role
in getting the ordinance passed. This was an era in which simple
assertion was all that was expected—that is, before epidemiological
data were required as a basis for such opinions.
In 1970, a report for the Royal College of Physicians of London on
air pollution and health appeared (14), and in the same year, after a
long period of U.S. Senate hearings and discussion, the first clean air
act was passed in the United States. Human health concerns played
a prominent part in motivating legislators to control air pollution at
this time.
The sequence—of effective legislation resulting from a single disas-
ter and public pressure on a government to deal with it—appears to
be the exception rather than the rule. As will be noted later, it is much
SETTING THE STAGE: CRITICAL RISKS 13
more difficult to get legislative action against lower level but repetitive
exposure to air pollutants. Nevertheless, by 1970, plenty of evidence
indicated that uncontrolled coal burning, with high particulate and
SO
2
 emissions, was damaging.
In the same year as the London episode, a chemist in Los Angeles
worked out for the first time the chemical reactions that led to the
formation of ozone from oxides of nitrogen in the presence of hydro-
carbons and sunlight—the generation of so-called "photochemical
smog" (15). In the 1960s, work began to establish the acute effects of
ozone on people and on plants and crops. In modern cities, most of
the oxide of nitrogen from which ozone is formed is emitted by auto-
mobiles. Initially, this problem was thought to be confined to Los
Angeles, because that is where most of the monitoring of the air for
photochemical oxidant pollution was being routinely carried out. It is
now recognized to be a serious problem over most of Europe, in
North America, and even in Australia.
Contemporary perceptions of air pollution can be summarized as
follows:
1. In eastern Europe and in much of the Third World, particularly
in China, gross particulate and sulphur dioxide pollution is occurring.
This is believed to have long-term adverse effects on both children
and adults. The first satisfactory epidemiological study of the effects
of living in this kind of pollution was published in London in 1965
(16). It took account of cigarette smoking, and, by studying employ-
ees of the British Post Office, standardized the socioeconomic level of
the population being studied. It found convincing evidence that lung
function was lower in men who lived in greater London than in men
who lived in smaller country towns, where sulfur dioxide levels and
particulate pollution were less than half what they were in London.
This kind of pollution leads to higher levels of chronic bronchitis and
possibly also of asthma. The effect of living in the pollution is interac-
tive with cigarette smoking. The basic data from this study are shown
in figure 2.2. Very similar data—as a consequence of air pollution
from open coal combustion—are now being reported from China (17).
The forced expiratory volume in liters (FEV1), a simple test of lung
function, was shown to be significantly lower in nonsmoking women
in Beijing if they lived in zones of the city with higher particulate and
SO
2
 pollution. A lower FEV1 in such cross-sectional studies in those
living in more polluted regions might be caused by a faster rate of
14 SETTING THE STAGE: CRITICAL RISKS
—A, Mean forced expiratory volume (1-0 sec.); B, mean peak
flow-rate (both standardised to age 40).
(a) non-smokers, (6) ex-smokers, (c) smokers, 1-14 g. per day,
(d) smokers, 15-24 g. per day, (e) smokers, 25 g. or more per day.
Mean value and number of subjects:
A, (a) 3-1 (30) & 3-0 (13); (6) 3-0 (77) & 2-8 (36); (c) 3-0 (142) &
2-7 (74); (d) 2-9 (134) & 2-7 (98); (e) 2-9 (40) & 2-5 (29).
B, (a) 562 (31 ) & 491 (13) ; (b) 517 (77 ) & 488 (36); (c) 510 (142 ) &
470 (74) ; (d} 485 (134 ) & 458 (98) ; («) 468 (41 ) & 448 (29) .
Figure 2.2. Cross-sectional study of lung function in British postal workers in
1965. From reference 16. At that time SO
2
 and smoke levels in the country
towns were about half those in London. Note that the largest differences in
forced expiratory volume in liters (FEV1) between the London and country
dwellers were in the heaviest smoking category. The peak flow rates are very
different in the nonsmoking category, but the numbers are small.
decline of lung function with age, or the presence of subclinical lung
changes changes, or the fact that children living in polluted air never
attain the same "normal" value for this test in early adult life; or, of
course, to all three of these effects.
2. Photochemical air pollution is now known to be much com-
moner than previously realized. The U.S. Clean Air Act of 1970 re-
quired the development of standards for the protection of public
health. In the case of ozone, this standard was initially set at 0.08 ppm
for one hour, but later was revised upwards to its present level of 0.12
ppm. The Canadian standard remained at 0.08 ppm. In 1988, 86 mil-
SETTING THE STAGE: CRITICAL RISKS 15
lion Americans were living in areas where the 0.12 ppm standard was
transgressed (2). Ozone is an intensely irritant gas, and lung function
is adversely affected in young normal subjects if they exercise in 0.08
ppm for six hours (18). Ozone acts by inducing inflammation in the
lung (19, 20). It is considered to be a special hazard to asthmatics,
because the effect on lung function and on airway responsiveness is
exerted on lungs that are already abnormal in these aspects (21).
Photochemical pollution is a major problem in many Third World
countries, and some of the highest levels now occurring are in Mexico
City. We still know very little about long-term adverse effects on peo-
ple of living in an oxidant environment. Crop yields, of such things as
alfalfa grass and beans, are depressed by ozone exposure below the
current standards during the growing season.
3. When sulfur dioxide emissions occur in the presence of ozone
formation, sulphuric acid aerosol is formed and can be directly de-
tected in the air (18). This is later neutralized by ammonia to form
sulphates. It is these aerosols that limit summer visibility over much
of the northeastern North American continent. This summer acid
haze, with ozone and ammonium sulphate, has been shown to be
associated with increased hospital admissions for acute respiratory
disease (23). These data are shown in figure 2.3 and in table 2.2. New
data from the summers of 1986, 1987, and 1988 from Toronto have
confirmed these findings (25).
4. Automobiles, trucks, and buses emit oxides of nitrogen and fine
particulates. These are less than 10 microns in size and of complex
composition. Under inversion conditions, they remain suspended in
the air, and form a visible blanket enveloping a city and obliterating
adjacent mountains. Recent studies have indicated that increases in
this form of pollution are regularly associated with increases in total
mortality; this has now been shown to be true of eleven different
locations with widely differing climatic conditions (26). Recent data
for Philadelphia are shown in table 2.3 and in figures 2.4 and 2.5. The
mechanism of this effect is currently unknown. It may be observed
that reducing general automobile pollution is a much more complex
problem than the reduction of particulate pollution from coal burning
required in 1953.
5. Most discussions of the effects of air pollution on health are cen-
tered on the interpretation of epidemiological data. Studies of con-
trolled acute human exposures have proved very important in defining
16 SETTING THE STAGE: CRITICAL RISKS
Table 2.2. Pearson Correlation Coefficients for July and August Admissions
in Southern Ontario for Years 1974 and 1976-1983 Based on Deviations from
Same Day of Week in Same Season in Same Year
Diagnostic Category
Respiratory
Variable
SO
2
LO
SO
2
 L24
SO
2
 L48
O
3
LO
O
3
L24
O
3
L48
NO
2
LO
NO
2
 L24
NO
2
 L48
COHLO
COM L24
COH L48
SO
4
LO
SO
4
 L24
SO
4
 L48
TLO
TL24
TL48
RHLO
RHL24
RHL48
Total
ns"
ns
ns
0.1217*
ns
ns
0.1137*
ns
ns
0.1586**
ns
ns
ns
ns
ns
0.1081*
ns
ns
ns
ns
ns
Respiratory
ns
ns
0.1383**
ns
0.1374**
0.1469**
ns
ns
0.1102*
ns
ns
ns
ns
0.1706**
0.1339*
ns
ns
0.1286*
ns
ns
ns
Minus
Asthma
ns
ns
0.1125*
ns
ns
0.1492*
ns
ns
ns
ns
ns
ns
ns
0.1341*
0.1290*
0.1069*
0.1344**
0.1681**
ns
ns
ns
Asthma
All ages
ns
ns
0.1063*
ns
0.1241*
ns
ns
ns
ns
ns
ns
ns
ns
0.1265*
ns
ns
ns
ns
ns
ns
ns
Asthma Non-
0-14
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
respiratory
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
"ns = Not significant.
*P=s0.01.
**P =s 0.001.
From reference 23. The simplest way of showing nonrandom associations between two vari-
ables, in this case hospital respiratory admissions (as a percentage deviation from the mean value
for that day of the week in the same season of the same year), and for the mean of the hourly
maxima for the pollutant for all monitors in the region. The associations are significant in spite of the
maximally Conservative data selection (since the mean value to which each day is compared is
affected by any pollution-related admissions).
a minimum-effects level of single gaseous pollutants, and in establish-
ing that some individuals are much more sensitive than others to cer-
tain pollutants (asthmatics being an order of magnitude more sensitive
to the acute effects of sulfur dioxide than are nonasthmatics, for exam-
ple). Animal exposure studies throw light on biological plausibility and
mechanisms of action. They are also the only way in which possible
long-term effects can be experimentally studied.
6. It has been difficult to establish with any certainty the long-term
adverse effects of living in a polluted atmosphere. The cross-sectional
SETTING THE STAGE: CRITICAL RISKS 17
NOTES'
I. For age group 0 to 14 only.
2. For summer months (July and August) only.
Figure 2.3. Hospital admissions for asthma in children in southern Ontario,
1974-1980. From reference 24. From the same data as in table 2.2. These data
for asthma admissions in children aged 0-14 show a consistent excess in
mean admissions 24 hours after the ozone level (as a mean of the hourly
maxima) exceeded the Canadian standard of 80 parts per billion.
study by Holland and Reid in Britain in 1965 (16) showed that lung
function was lower in men in a more polluted region (see figure 2.2);
this constitutes one of the more convincing pieces of evidence of long-
term effects in the general population. It is possible that the observa-
tion is to be explained by the adverse effects of pollutants on children,
ASTHMA
Mean admissions per Day
YEAR
18 SETTING THE STAGE: CRITICAL RISKS
Relative risk of death in Philadelphia by quintile
of total suspended particulates controlling by regres-
sion for year of study, time trend, and weather.
Figure 2.4. Relative risk of death in Philadelphia in relation to particulates.
From the same data as table 2.3 (reference 26). Note than an increase from
about 50 to 120 micrograms/m
3
 in pollution is associated with about a 4 per-
cent increase in mortality.
so that in polluted areas the population never reaches the same level
of "normality" of pulmonary function as do children growing up in a
clean environment. Such effects as shortened survival provide very
crude indices of effects; and major disbenefits must have occurred
before such an index would change significantly. Differences in respi-
ratory disease incidence or in mortality in cross-sectional comparisons
are difficult to interpret unless cigarette smoking is precisely known,
since this is such a dominant factor in the causation of chronic respira-
tory disease. Even passive smoke inhalation has to be considered in
this context.
This very brief summary of current knowledge leads to the follow-
ing conclusions:
SETTING THE STAGE: CRITICAL RISKS 19
Figure 2.5. Relative risk of mortality in Philadelphia associated with a 100
microgram/m
3
 increase in total suspended particulates (TSP) concentration
after controlling by regression for year of study, time trend, and weather.
Results for all-cause mortality, age-stratified mortality, and cause-specific mor-
tality. From the same database as table 2.3 and figure 2.4 (reference 26). Note
the increased risk of death from chronic obstructive pulmonary disease and
pneumonia; however, together these account for only about one-tenth as
many deaths as cardiovascular disease (see table 2.3). Numerically, the 10
percent increase in relative risk for cardiovascular disease when there is a 100
microgram/m
3
 increase in particulate pollution is more important in driving
the overall relationship.
1. Intense media attention can be generated by a single major epi-
sode of air pollution, but it is much harder to get the attention of
legislators in order to deal with long-term, unspectacular, chronic
effects. Yet, in the long run, these may be the more important.
2. As noted in table 2.1, in almost all studies of air pollution
effects the actual exposure of the subject is very poorly charac-
terized. To some extent this is inevitable if a large population is
involved.
3. Air pollution generally represents an involuntary risk. Some
avoidance may be possible by staying indoors. Some segments of the
population, such as children returning from school at a time when
ozone concentrations are highest, or young adults taking vigorous
20 SETTING THE STAGE: CRITICAL RISKS
Table 2.3. Daily Mortality Counts, Weather, and Air Pollution in
Philadelphia, 1973-1980
Weather
Temperature, °F
Dew Point, °F
Air pollutants
TSP, Mg/m
3
SO
2
, ppb
Mortality, deaths/
day
Total
Age ss 65 yr
Age > 65 yr
Cause-specific mortality,
deaths/day
Cardiovascular
Cancer
Pneumonia
Chronic obstructive
pulmonary disease
5%
25
9
37
6
35
10
21
13
5
0
0
10%
30
15
43
8
37
11
23
15
6
0
0
25%
41
28
56
12
42
14
26
18
8
0
0
50%
56
44
73
18
48
17
31
22
10
1
1
75%
71
59
94
27
54
20
36
26
13
2
1
90%
77
68
117
39
60
23
40
30
15
3
2
95%
80
70
132
46
64
25
44
32
16
4
3
Mean
54.8
42.8
77.2
21.0
48.2
16.9
31.3
22.1
10.5
1.44
0.89
exercise in polluted atmospheres, must be presumed to get much
higher exposures than others.
4. It is not possible to estimate whether the public perception of
the current risks posed by air pollution is generally higher or lower
than statistical estimates. In any community with any air pollution
problem, there is a cadre of interested citizens prepared to urge the
importance of its amelioration. There are special problems in the
Third World. A recent World Bank report (22) noted, "The potential
constituency in favor of air pollution control is weakly identifiable and
poorly informed of the causes, possible effects, and available techno-
logical alternatives. In contrast, those against pollution control en-
forcement are economically powerful and politically well organized."
5. Major industries interested in air pollution legislation are coal-
burning electric utilities, the mining industry that produces the coal,
and the automobile industry and the fuel suppliers for automobile
transport. These have generally mounted strong opposition to more
stringent controls or to tighter standards. The American Petroleum
Institute in the United States a few years ago apparently spent several
million dollars to oppose the proposed ozone standard.
6. Media coverage tends to be intermittent. In Los Angeles, a com-
plex system of public alerts in relation to the level of oxidant pollution
SETTING THE STAGE: CRITICAL RISKS 21
makes the public aware of the pollution situation on a day-to-day
basis. Similar daily reports on air quality now exist in many cities. A
recent (December 14 and 15, 1991) oxide of nitrogen pollution episode
in London led to public warnings carried in newspapers and on the
radio. In North America, readings of various "air pollution indices"
are routinely published and publicized.
7. In western Europe and North America, clean air legislation intro-
duced between 1956 and 1970 has led to a major diminution in SO
2
and gross particulate pollution. Although in 1960 pollutant emission
levels were not much different between western and eastern Europe,
by 1990 many regions in eastern Europe were grossly polluted,
whereas in western Europe these forms of pollution had been greatly
reduced and were under control. The latest U.S. Clean Air Act amend-
ments (1990) are expected to reduce sulfur dioxide emissions still
further. In most democratic societies, there has been a legislative re-
sponse to the public demand for clean air.
8. The American Lung Association has led public pressure for effec-
tive legislation to control air pollution in the United States. In Canada, a
special coalition was formed to fight acid rain. The strong public sup-
port for this exerted some pressure on the Canadian government to try
to persuade the U.S. government of the need for joint legislative action.
9. As noted in table 2.1, in general, courts and lawyers have not
been much involved in the process of air pollution control. However,
they have been extensively involved in hazardous-waste site litiga-
tion, and in some cases of localized release of toxic air pollutants. In
Japan, unique legislation permitted legal suits between victims of air
pollution (who not surprisingly proved to be very numerous) and
primary emitters.
Environmental groups, together with the American Lung Associa-
tion, have recently filed a suit against the U.S. Environmental Protec-
tion Agency (EPA) complaining that the agency has not moved to
consider revision of the ozone standard in the light of new scientific
information (some of it originating in the laboratories of the EPA
itself). It is possible that, in the future, the courts may be more in-
volved in such questions in the United States, but in a parliamentary
system, as noted below, there would be no grounds for any such
action.
We may fairly conclude that public opinion has been important in
the control of air pollution; that it is difficult to deal with noncata-
22 SETTING THE STAGE: CRITICAL RISKS
strophic events; that current air pollution is much more complex in
origin and effects than was formerly the case; that biological plausibil-
ity varies with different combinations of pollutants; that there has
been relatively little involvement of the legal process, and media inter-
est is generally local; and that there are significant present and future
problems in the Third World in relation to indoor and outdoor air
pollution.
Cigarette Smoking
In table 2.1, I have dated concern about cigarette smoking to the date
of publication of the first U.S. Surgeon General's report on the effects
of smoking, which appeared in 1964. A report of the Royal College of
Physicians of London had appeared two years earlier documenting
the adverse effects of smoking. The history of cigarette smoking is as
replete with statistical information as is baseball. In the following
summary, some of the salient landmarks are identified.
The sociological history of cigarette smoking has many interesting
facets. In about 1880, the popular Victorian painter T. J. Frith painted
a large canvas of a young man lighting a cigarette for a rather gaudily
attired young lady. Probably painted in the Tivoli Gardens in Copen-
hagen, the picture was entitled Her First Cigarette. It was one of a
series of paintings with moral messages, but was deemed too obscene
for public display. At about the same time, Van Gogh painted a re-
markable canvas of a skeleton with a cigarette between its jaws. This
picture is very rarely reproduced; perhaps the tobacco companies
managed to get it suppressed.
The design of automated machinery for cigarette manufacture was
perfected in 1913; the American Cancer Society was founded in the
same year (28). A Canadian text on human physiology for use in
schools, published in 1930 (29), contained dire warnings of the evil
consequences of smoking. It stated that smoking leads to moral degen-
eration, and asserted (without giving a reference) that the reformato-
ries and prisons were full of cigarette smokers. Unfortunately, the
dangers of lung disease were not mentioned.
Cigarette consumption rose steadily but slowly until World War II,
when in all countries there was a major acceleration of the rate of
smoking. Epidemiological studies, including a prospective study of
British doctors, began in about 1950. The first disease to be linked
SETTING THE STAGE: CRITICAL RISKS 23
unequivocally to cigarette consumption was lung cancer. The causal
significance of the association was immediately challenged. Sir Ron-
ald Fisher, a leading British biostatistician, a Fellow of the prestigious
Royal Society, and a pioneer in the statistics of experimental design,
was prominent in suggesting that there was no causal link between
cigarettes and lung cancer. Only recently has it become known that he
was being paid for his activities by the tobacco industry (30).
By the time of the publication of the U.S. Surgeon General's report
in 1964, there was overwhelming evidence that cigarette smoking was
strongly associated with lung cancer, chronic obstructive lung dis-
ease, peripheral vascular disease, and heart disease. The causal rela-
tionship was no longer doubted by the majority of physicians, and
further epidemiological studies were hardly required.
Recent data indicate that adverse effects are even associated with
passive inhalation of cigarette smoke (33). In 1986, both the National
Academy of Sciences (31) and the U.S. Department of Health and
Human Services (32) issued summary reports on the effects of passive
tobacco smoke inhalation. I was at a meeting in Germany in 1988 at
which there was a session on passive tobacco effects (33). (A distin-
guished, but very elderly, American biostatistician was flown over to
tell us that the evidence of causality in such studies was too flimsy to
be believed.) The observation that passive smoke inhalation was capa-
ble of causing ill effects had the effect of changing completely the
basis for legislation, since cigarette smoke was no longer only a volun-
tary risk. Prohibition of smoking in commercial aircraft, workplaces,
and restaurants quickly followed.
In Britain, cigarette consumption reached a peak in the mid-1960s;
thereafter it fell in men but continued to rise in women (34). In the
United States, smoking peaked in 1963, when per capita consumption
in persons over the age of 18 reached 4250 cigarettes per year (35). In
the United States in 1965, 13 percent of women smoked more than 25
cigarettes a day; by 1985 this had risen to 23 percent (36). The reasons
for this increase are largely unknown, but may well be attributable to
a clever advertising campaign aimed at the self-image of young
women.
The disease burden attributable to cigarette consumption is enor-
mous. It has been calculated that a 25 year old smoking two packs/day
loses 8.3 years of expected longevity; this works out to five and a half
minutes of life lost per cigarette smoked (37). Direct adverse health
24 SETTING THE STAGE: CRITICAL RISKS
costs of smoking in the United States have been estimated at $16
billion a year, with indirect costs more than double that figure (37).
The disease burden attributable to smoking far outweighs the com-
bined effects of the other hazards I have chosen to consider. Sir Rich-
ard Doll (38), writing in 1983, concluded:
The avoidance of smoking would alone reduce the mortality
from all cancers by about a third (including avoidance of not
only the large majority of cancers of the mouth, throat, and
lung, but also a substantial proportion of the deaths attrib-
uted to cancers of the bladder, kidney, and pancreas). It
would almost eliminate chronic obstructive lung disease
and the complications of peripheral vascular disease, would
reduce the age specific mortality from aortic aneurysm by at
least three quarters and the mortality from myocardial in-
farction by about a quarter, and would probably lead to a
small reduction in perinatal mortality in the poorer socio-
economic groups.
Furthermore, concomitant cigarette smoking greatly increases the risk
of lung cancer in asbestos- or radon-exposed workers. This question is
discussed in more detail in Chapter 3.
Since 1964, in many countries public societies have been established
to reinforce government campaigns against the tobacco industry. There
is convincing evidence that measures taken—including prevention of
television advertising and progressively more severe taxation—have
had an effect. It is now calculated, for example, that whereas in 1989
annual adult per capita cigarette consumption in the United States
would have been fifty-five hundred cigarettes without preventive mea-
sures, the actual figure was about three thousand (39). The percentage
of men who smoked reached a high of 69 percent in 1958, but by 1989
this was down to 30 percent (39). In Britain, lung cancer rates in men
have peaked and have begun to decline, but they are still rising in
young women. However, the lung cancer rates in women of all ages
began to fall in about 1970 (see figure 2.6) (34). The same is broadly true
for chronic obstructive lung disease mortality.
The diminution of cigarette consumption in free societies has not
been uniform. Reasons for the differences between these countries
are obscure, but they may be a reflection of different media emphasis.
SETTING THE STAGE: CRITICAL RISKS 25
Lung cancer rate ( ), cumulative constant tar cigarette
consumption (CCTCC ), and hypothesised "background" rate of
lung cancer unrelated to cigarette consumption ( ) ,for women aged
40-44, 1941-85.
Figure 2.6. Changing patterns of lung cancer mortality and cigarette smoking
in women in Britain. From reference 34. Although air pollution began to
decline in the UK from about 1965 onwards, the authors do not consider that
the decline in lung cancer might be due to this factor, but instead that the rise
in lung cancer in women from 1940 to 1970 occurred against a falling back-
ground rate of lung cancer unrelated to cigarettes.
Table 2.4 shows the changes in smoking rates in different countries
between 1974 and 1987 (40). It has been difficult in democratic soci-
eties to define what steps can be taken against a product that can
legally be sold. Lawsuits brought by the relatives of victims of lung
cancer against the tobacco industry have generally not succeeded. In
Canada, a recent court decision upheld the right of the government to
restrict tobacco advertising, and there are current proposals to require
licensing of sellers of tobacco and to outlaw cigarette machines from
public places. Campaigns in high schools about the dangers of ciga-
rette smoking may have had some effect, but the increasing rate of
smoking among young women has not been halted. Perhaps there
was more sense to the banning of the display of Frith's painting than
one might have thought.
26 SETTING THE STAGE: CRITICAL RISKS
Table 2.4. International Comparisons of Trends in Cigarette
Smoking Prevalence
MEN aged 20 and over: % smoking
YEAR USA GB Australia
1974-6 43.4 52 43.3
1983
1986-7
35.5
31.7
38 38.9
35 32.9
WOMEN aged 20 and over: % smoking
YEAR USA GB Australia
1974-5 31.4 41 31.9
1982-3
1986-7
29.4
26.8
34 31.4
31 30.6
Canada
45.6
35.7
32.3
Canada
32
29.3
26.6
Norway
49.6
44.6
41.3
Norway
32.1
32.3
33.3
Sweden
36
26
24
Sweden
34
28
27
In every country, th,e higher the education level, the lower the smoking prevalence.
From reference 40. Note the higher declines in women in smoking in the United Kingdom,
Canada, and the United States, compared to Australia and Norway. This might be due to differ-
ences in media impact.
Media emphasis on the health effects of cigarette smoking has been
variable. Initially, the press generally underemphasized the risk, fear-
ing loss of advertising revenue. The tobacco industry, with enormous
profits, commands such a powerful potential (used historically to
"buy" some scientific testimony or to pay for influential back-room
lobbying in political circles) that it is not easy to combat it. The income
from tobacco taxation is important for many governments, further
inhibiting effective legislation. In many countries, tobacco advertising
still succeeds in getting a message across to the insecure teenager that
smoking is an important symbol of adult independence.
Cigarette advertising and promotion in the Third World is gener-
ally unrestricted, and, as consumption has declined in the industrial-
ized countries, tobacco consumption has been aggressively promoted
by the tobacco industry in lesser developed nations. In China, the late
Chairman Mao was rarely photographed without a cigarette in his
hand, so public health emphasis on the dangers of smoking was
impossible. More recently in China, some public education on the
dangers of smoking has been initiated.
In the case of cigarette smoking, the epidemiological data are now
more or less complete; initially, the causative inference was contested;
there have been significant government initiatives in most western
countries; the issues now are psychosocial and legal, and in all juris-
dictions problems exist in limiting advertising and promotion; and
currently there is an intensive push by tobacco companies to increase
SETTING THE STAGE: CRITICAL RISKS 27
sales in the developing world. Pierce (41) has recently noted that the
overseas sales of American cigarettes doubled between 1986 and 1989.
He attributes this, at least in part, to the fact that cigarette manufac-
ture "is increasingly becoming an oligopoly controlled by four large
transnational companies and the government of China."
Asbestos
A great deal has been written about the remarkable history of asbes-
tos. Commercial production started in about 1880, and by 1920 world
production was about two hundred thousand tons annually. By 1940
this had about doubled; but a step increase to about a million tons a
year occurred during World War II. Part of this was accounted for by
the extensive use of asbestos in ships. However, general use had also
been rising steadily: the United States-manufactured steam locomo-
tive of the 1930s contained three thousand pounds of asbestos (42).
After 1950, production growth rose continuously till 1970, when it
amounted to about 4 million tons per year. In the building boom of
the 1950s, asbestos was widely applied within buildings as a sprayed
fire retardant. The growth in the number of automobiles, for which
asbestos was required for brake linings, was also part of this expan-
sion. As noted in the introduction, currently two-thirds of the eight
hundred thousand buildings in Manhattan are believed to contain
asbestos, and in many the mineral is in poor condition, being friable
and easily dispersed (6).
Becklake (43) has plotted this growth of asbestos production with
the evolution of medical knowledge in relation to its effects (see figure
2.7). That it was capable of causing irreversible lung changes was
already known in the 1930s. Its relationship to lung cancer was clearly
established only forty years later, and it was not until the mid-1960s
that the entity of pleural mesothelioma was defined. The major confer-
ence convened by the International Agency for Research in Cancer
(IARC) in 1972 (44) provides a useful milestone in the evolution of
modern knowledge. By that time, the increased incidence of mesothe-
lioma in port cities around the world as a result of ship repair work
was well established. It also became clear that a mesothelioma could
result from relatively high exposure over a short period of time. From
this date, the relative dangers of the different forms of asbestos be-
came a major question, not wholly resolved even today.
28 SETTING THE STAGE: CRITICAL RISKS
Diagrammatic representation of the growth of the asbestos industry and the recognition
of the associated biologic effects. The following symbols indicate the association with asbestos: ? =
suspected; CHH! = probable: •••= established.
Figure 2.7. Asbestos production and the evolution of medical knowledge.
From reference 43. Since 1980, there has been a considerable reduction in
world use of asbestos, but, as noted in the text, mesothelioma incidence in
some countries is still rising.
The forced bankruptcy in 1982 of the United States' major producer
of asbestos—as a result of the many outstanding injury claims that
had been filed against it—has been documented in detail by Paul
Brodeur (45). The role of the asbestos industry and of the lawyers
representing it, in fighting allegations of responsibility, constituted a
remarkable indictment (not contested in any lawsuit against the au-
thor of the book).
As indicated in table 2.1, the epidemiological data in relation to
asbestos are very strong and come from many countries. The progres-
sive decline in maximal exposure standards for asbestos fibers from
1946 to 1990, as shown in table 2.5, indicates the progressive awareness
of the risks from inhaling the material. As shown, in 1968 the American
Council of Governmental and Industrial Hygienists (ACGIH) standard
was 12 fibers per milliliter (ml) of air; and in 1990 the U.S. Occupational
Safety and Health Administration (OSHA) standard was 0.1 fibers/ml,
a drop of more than a hundred-fold (4). The precise risk attendant on
use of buildings that contain asbestos is still controversial, and measure-
SETTING THE STAGE: CRITICAL RISKS 29
Table 2.5. Recommended Air Concentration Limits and Standards
for Asbestos
Group
ACGIH
ACGIHG
ACGIH
OSHA
OSHA
NIOSH
ACGIH
OSHA
Year
1946
1968
a
1970,* 1974
b
1972
1976
1976
1978,* 1980
b
1976
1986
1990 proposal
Limit
5 x 10
6
 particles/ft
3
12 fibers
c
/mL or 2 x 10
6
 particles/
ft
3
5 fibers/mL
5 fibers/mL
2 fibers/mL
0.1 fiber/mL
0.2 fiber/mL for crocidolite
0.5 fiber/mL for amosite
2.0 fiber/mL for chrysotile and
other forms
2.0 fiber/mL
0.2 fiber/mL
0.1 fiber/mL
"Notice of intent.
b
Adopted as threshold limit value (TLV).
C
A11 fiber limits based on phase-contrast optical determination at 400-450x magnification.
From reference 46. Over this period, standards for many other hazardous materials, such as lead
and benzene, have also fallen.
ment of exposure in this situation is very difficult (see table 2.6). The
risk is usually involuntary.
Although to my knowledge it has not been quantified, the public
perception of the risk from asbestos may be higher than the pub-
lished numerical estimates. Both primary producers and secondary
industries have been implicated in perceived risk. Media coverage
has been very intermittent; some vivid descriptive television docu-
mentaries were prepared in the early 1980s, but the subject is now
rarely mentioned in the media except in relation to local issues (usu-
ally related to contentious litigation). There has, however, been con-
siderable discussion in the media in relation to local issues, for exam-
ple the pressure on local school boards to insist on the removal of all
asbestos from schools regardless of its physical state or the dangers
it represents.
What distinguishes the present status of asbestos as a health haz-
ard in the United States is the dominant role played by the courts and
the legal profession. The financial outcomes are so large and the cost
of asbestos removal (if done safely) is so considerable that many major
lawsuits have taken place or are in process. These have required
Table 2.6. Distribution of Building Average Airborne Asbestos Concentrations for Nonlitigation Data
by Study
3
No. of Building
Study Buildings Types
b
 Minimum
Burdett and
Jaffrey 1986
C
Chatfield 1986
Gazzi and
Crockford 1987
C
Hatfield et al.
1988; Chesson
et al. 1990b;
Crump and
Farrar 1989
Pinchin 1982
CPSC 1987
McCrone 1991
(unpublished)
schools
McCrone 1991
(unpublished)
office
39
7
25
43
19
45
19
1
5S,8PC,26
R
5S,2PC
R
PC
S
R
S
PC
0
0
0
0
0
0
0
—
10th
Percentile
0
0
0
0
0
0
0
—
Median
0.0001
0.0005
0
0
0
0
0.0002
—
Mean
0.00026
0.00243
0.00030
0.00005
0.00042
0.00010
0.00022
0.00004
90th
Percentile
0.0009
0.0080
0.0008
0.0003
0.0020
0
0.0005
—
Maximum
0.0017
0.0080
0.0025
0.0006
0.0030
0.0020
0.0016
—
a
Fibers greater than 5 /JLITI.
b
S = schools, PC = public and commercial buildings, R = residences.
c
Only including buildings with asbestos.
From reference 47. Note that this is from "nonlitigation data."
SETTING THE STAGE: CRITICAL RISKS 31
scientific testimony on the exact nature of the risk. Distinguished
physicians and pathologists have testified on both sides of many of
these cases, the point at issue usually being the risk of low-level
chrysotile exposure. In the United States, the need for quotable testi-
mony, from scientific meetings, that favors one side or the other in
such litigation, has led to "stacked" quasi-scientific meetings with
speakers chosen to emphasize one side of the argument. In the recent
Health Effects Institute report on asbestos (47), one table is headed,
"Distribution of Building Average Airborne Asbestos Concentrations
for Nonlitigation Data by Study," eloquent evidence of the role of the
courts (see table 2.6). Asbestos is the dominant example among the
five hazards under discussion in which extensive litigation has oc-
curred. Its implications will be discussed later.
As I have noted elsewhere, much of the present confusion is attrib-
utable to the fact that different opinions can be legitimately held on
the basis of the current evidence (48). There is no sign that the prob-
lem of asbestos-related disease has reached its peak. The hospital
discharge rates for asbestosis for white males in New Jersey has risen
from about 30 per 100,000 per year in 1979 to more than 150 per year
in men over the age of 75 in 1986, and the rate does not appear to be
flattening (see figure 2.8) (4). A recent paper from the Netherlands
(50) describes the course of understanding the asbestos risk in that
country. It is of interest that the rate of dissemination of knowledge
and legislative action have both varied among countries in the free
world.
The scientific basis for the present controversy can be briefly sum-
marized as follows:
1. Current risk estimates are inevitably based on extrapolations
from earlier epidemiological data for various occupations; there is
considerable doubt as to the reliability of the exposure estimates that
underlie these. Table 2.7, from the Health Effects Institute report,
shows computed risks from environmental exposures; it is preceded
by a note of three factors that are responsible for uncertainty. These
are: (a) the uncertainty of historical occupational exposure data, from
which low dose risks are extrapolated; (b) uncertainty about present
levels of exposure to asbestos within buildings; and (c) uncertainty
about the comparability of different methods of measuring airborne
asbestos fibers.
2. Although mesothelioma is probably dose-response related,
32 SETTING THE STAGE: CRITICAL RISKS
Figure 2.8. Annual age-specific discharge rates of asbestosis among white
males in New Jersey, 1979-1986. Data from reference 49. Note that the hospi-
tal discharge rates for asbestosis in New Jersey have not yet started to fall.
about one-third of current cases have no known asbestos exposure
(51). Huncharek (52) has recently summarized the evidence concerning
the exposures to asbestos in cases of mesothelioma; the occupations of
twenty-four men and sixteen women with mesothelioma—who had
had no asbestos-related jobs or self-reported exposures—varied
greatly. As the disease has a very long latent period (about thirty-three
years in one series of cases), it may be questioned whether any possible
impact from environmental exposure would now be visible. As noted
above, the incidence of mesothelioma in the Netherlands is still rising.
3. Different types of asbestos fibers are known to disappear from
SETTING THE STAGE: CRITICAL RISKS 33
Table 2.7. Estimated Lifetime Cancer Risks for Different Scenarios of
Exposure to Airborne Asbestos Fibers
3
Premature Cancer Deaths
(Lifetime Risks)
Conditions per Million Exposed Persons
Lifetime, continuous outdoor exposure
• 0.00001 f/mL from birth (rural) 4
• 0.0001 f/mL from birth (high urban) 40
Exposure in a school containing ACM, from age 5
to 18 years (180 days/year, 5 hours/day)
• 0.0005 f/mL (average)
b
 6
• 0.005 f/mL (high)" 60
Exposure in a public building containing ACM
age 25 to 45 years (240 days/year, 8 hours/day)
• 0.0002 f/mL (average)
15
 4
• 0.002 f/mL (high)
b
 40
Occupational exposure from age 25 to 45
• 0.1 f/mL (current occupational levels)
c
 2,000
• 10 f/mL (historical industrial exposures) 200,000
a
This table represents the combined risk (average for males and females) estimated for lung
cancer and mesothelioma for building occupants exposed to airborne asbestos fibers under the
circumstances specified. These estimates should be interpreted with caution because of the reserva-
tions concerning the reliability of the estimates of average levels and of the risk assessment models
summarized in Chapter 8.
'The "average" levels for the sampled schools and buildings represent the means of building
averages for the buildings reviewed herein (Figure 1.1). The "high" levels for schools and public
buildings, shown as 10 times the average, are approximately equal to the average airborne levels of
asbestos recorded in approximately 5 percent of schools and buildings with asbestos-containing
materials (ACM) (see Chapters 4 and 8). If the single highest sample value were excluded from
calculation of the average indoor asbestos concentration in public and commercial buildings, the
average value is reduced from 0.00021 to 0.00008 f/mL, and the lifetime risk is approximately halved.
The concentration shown (0.1 f/mL) represents the permissible exposure limit (PEL) proposed
by the U.S. Occupational Safety and Health Administration. Actual worker exposure, expected to
be lower, will depend on a variety of factors including work practices, and use and efficiency of
respiratory protective equipment.
From reference 47.
the lung at different rates. Chrysotile is cleared the fastest (53), and
the more dangerous amphiboles, such as amosite and crocidolite,
which are thought to lead to mesothelioma at lower exposures, are
cleared much more slowly (54). This means that estimates of the num-
bers of residual fibers in the lung at autopsy may be unreliable guides
to exposures.
4. The lungs of city dwellers (in Vancouver, for example) currently
contain up to 40 million asbestos fibers (55). Churg, commenting re-
cently on this finding (56), noted, "Yet despite this fiber burden, no
epidemiological evidence exists to suggest that the general population
34 SETTING THE STAGE: CRITICAL RISKS
Table 2.8. Comparison of Lung Cancer Risks Estimated by Various Groups
or Individuals from Studies of Asbestos-Exposed Workers3
Percent Increase in Lung Cancer Per f-y/mL of
Exposure (100 x K
L
)
Study
Dement et al. (1983b)
McDonald et al. (1983a)
Peto et al. (1985)g
McDonald et al. (1983b)
Berry and Newhouse (1983)
McDonald et al. (1984)
McDonald et al. (1980)
Nicholson et al. (1979)
Rubino et al. (1979)
Seidman (1984)
Selikoff et al. (1979)
Henderson and Enterline (1979)
Weill et al. (1979)
Finkelstein (1983)
Newhouse and Berry (1979)
Males
Females
Values used for risk
extrapolation
EPA
b
2.8
2.5
1.1
1.4
0.058
0.010
0.06
0.17
0.075
4.3
0.75
0.49
0.53
6.7
1.0
CPSO
2.3
1.0
0.06
0.06
0.12
0.17
6.8
h
1.0
0.50
0.31
4.8
0.3-3.0
NRC
d
5.3
0.8
0.06
0.15
9.1
h
1.7
0.3
1.3
8.4
2.0
Ontario
Royal
Commission
6
4.2
1.0
0.058
0.020-0.046
1.0
0.069
4.2'
0.02-4.2
HSC
f
1.25
0.54
1.0
"Adapted from Nicholson (1986).
b
U.S. Environmental Protection Agency (Nicholson 1986).
C
U.S. Consumer Product Safety Commission (1983).
d
National Research Council (1984).
e
Ontario Royal Commission (1984).
f
U.K. Health and Safety Commission (Doll and Peto 1985).
SEarlier reviews cited Peto (1978) or Peto (1980), and some noted that all men employed after
1951 suffered a higher dose-specific risk (100 x K
L
 = 1.5).
h
Data from Seidman and colleagues (1979).
'Unpublished data supplied to the Commission.
From reference 47. Note the ranges of risk estimates. In the case of asbestos, there is probably
only a small error in outcome measurement; the uncertainty lies in estimates of past exposure.
suffers from any type of asbestos related disease." With a hazard
apparently now so widely disseminated, how could one evaluate its
contribution to the incidence of lung cancer?
These issues indicate that it is largely scientific uncertainty, to-
gether with the problem of extrapolating from high to low exposures,
that feeds the present highly litigous environment. The consequences
of this will be examined in Chapter 4.
Figure 2.9. The risk of lung cancer in different occupations and with different fibers. From
reference 47. Note the generally higher levels of risk in secondary industries compared to
those in mining and milling of asbestos.
36 SETTING THE STAGE: CRITICAL RISKS
Therefore, in the case of asbestos, current controversy involves the
risk of long-term, low-level exposures; biological plausibility is not an
issue; there are differences of opinion on the relative risks of different
fiber types; and the debate is complicated by problems of exposure
assessment, different rates of fiber clearance from the lung, and the
long latent period before effects appear. In the United States, the legal
process has subsumed scientific discussion of these problems, and
media interest is sporadic. Asbestos companies, actively backed in
Canada by the government, have been pushing sales of the mineral in
the Third World. It is exceedingly unlikely that the safe use of asbes-
tos can be guaranteed in these countries, though the industry asserts
that this is the case. In 1987, I published an editorial (57) that raised
the question of whether the medical profession in Canada should
consider itself responsible for the uncontrolled export of such danger-
ous material to the Third World, but to this date no effective action
has been taken by any government agency or professional organiza-
tion with respect to it.
Lead
Articles on lead commonly begin with the observation that the haz-
ards of lead mining were known to the Romans. Indeed, some histori-
ans have regarded lead poisoning as an important factor contributing
to the decline of the Roman Empire. In most modern cases, questions
of lead poisoning have arisen in connection with emissions from a
local smelter or battery recovery plant. The tortuous process of in-
quiry, political responsibility, and confused scientific evidence in the
case of a small plant in Toronto has been excellently documented by
Salter (58). In Trail, British Columbia, the long-term operation of a
large smelter has led to increased lead in the soil in the adjacent town
and to raised blood lead levels in the residents (59).
However, the widespread use of lead in gasoline was responsible
for dissemination of this mineral, even into the Greenland ice cap (see
figure 2.10; 60). The symptoms of acute lead poisoning in adults and
in children were generally recognized before 1939. What has trans-
formed the discussion is that beginning in 1943, epidemiological data
indicate that even at very low levels the blood lead in children is
related to behavioral changes and a lowering of IQ. This important
SETTING THE STAGE: CRITICAL RISKS 37
Lead concentration, profile in snow strata of Northern Greenland (EPA, 1986).
Figure 2.10. Lead concentration profile in Greenland snow strata, A.D. 1750 to
A.D. 1970. From reference 60. The increase in lead in the snow strata of
Greenland is a consequence of its wide dispersal in fine particulate form in
gasoline.
controversy has recently been summarized in a collected series of
papers that provide an excellent review of the data from different
countries (61):
1. Needleman, who pioneered the first large studies relating lead
level in deciduous teeth or blood level to behavioral and intelligence
parameters, has recently summarized twenty-four studies of this type
(62). The regression coefficients for lead levels against IQ were nega-
tive in eleven of the twelve studies in which it was calculated. This
means that the higher the blood lead, the larger is the decrement from
the predicted IQ.
2. In 1986, a small commission convened by the Royal Society of
Canada (63) at the request of the government, heard evidence relating
to lead and its effects. The summary and conclusions noted:
38 SETTING THE STAGE: CRITICAL RISKS
8. Much controversy persists concerning the nature of the
relationship between blood lead and neuropsychological
effects in children. Two major problems are:
• the difficulty in determining the relative contributions of
lead and socio-demographic factors to neuropsychological
functions of children.
• the difficulties in interpretation arising from the small-
ness of the effects commonly attributed to lead at the blood
lead levels encountered among children. Current studies
being conducted prospectively have yet to confirm the ob-
servations reported in earlier studies. However, the consis-
tency of the direction of the association, even after statisti-
cal adjustment for confounding variables, suggests that the
effect is indeed real, though small.
9. Other effects on the central nervous system have been
suggested by neurophysiological studies of the central ner-
vous system in children. These studies have yielded con-
flicting results that are hard to interpret. The Commission
considers them to be still experimental.
The commission nevertheless recommended that lead should be
removed from gasoline. The 1987 report to the Canadian Environmen-
tal Assessment Research Council (64) reviewed the epidemiological
data and argued that earlier risk estimates had underestimated the
totality of risks from exposure to lead.
2. The results from a European multicenter study (65) conducted
in Bulgaria, Denmark, Greece, Hungary, Italy, Rumania, West Ger-
many, and Yugoslavia, and involving 1879 children whose blood lead
levels varied between 5 and 60 micrograms/100 ml, concluded that
"there are detectable exposure-related neurobehavioral effects in
school age children." This was published in 1990.
3. A 1988 British Medical Research Council (MRC) advisory com-
mittee (66) review of the data then available concluded, "While the
observed statistical associations detailed in this review are consistent
with the hypothesis that low-level lead exposure has a small negative
effect on the performance of children in ability and attainment tests,
the limitations of epidemiological studies in drawing causal inferences
are such that it is not possible to conclude that exposure to lead at
SETTING THE STAGE: CRITICAL RISKS 39
current urban levels is definitely harmful." This paragraph is dis-
cussed in detail later.
4. A recent U.S. National Academy of Sciences report on neuro-
toxicology (67) used lead to illustrate the importance of a minor (5
percent) downward shift in mean IQ score (figure 2.11). In a popula-
tion of 100 million children, such a change results in a fall in the
number of individuals with an IQ greater than 130 from 2.3 million
to 990,000. Those scoring less than 70 rise correspondingly in num-
bers from 990,000 to 2.3 million. Figure 2.12 shows an illustration of
the relationship between umbilical cord lead and an index of mental
development.
The contemporary argument lies partly in the interpretation of the
epidemiological data in terms of causality (discussed later), and partly
in the variability of indices of childhood behavior, attention span, and
IQ measurements. In the recent volume on lead (68), Needleman and
Beringer contributed a paper entitled "Type II Fallacies in the Study of
Childhood Exposure to Lead at Low Dose: A Critical and Quantitative
Review." They noted, "Making causal connections in the real world is
not a pure, value-free enterprise."
It seems clear that differences in style exist among investi-
gators and interpreters of the same data base. These stylis-
tic, value-laden differences in interpreting the data seem to
be rather stable traits.
A Type II error is to deny a causal link when it is later shown to exist
(see page 65).
The many pathways of lead exposure in children are shown in
figure 2.13 (69). For some years the industry interested in keeping
lead in gasoline argued that airborne lead from vehicles was such a
minor component of the lead intake of children that its elimination
would have little effect on blood levels. However, the reduction of
lead in gasoline in the United States has been closely paralleled by a
remarkable fall in the blood lead of children, which unquestionably
indicates that automobile-derived lead was of far greater importance
than had been realized (see figure 2.14).
Although lead levels in blood and deciduous teeth are useful bio-
markers of lead exposure, they are not without limitations. Blood lead
SETTING THE STAGE: CRITICAL RISKS
Effect of a shift in mean IQ
score on the population distribution. The
top figure represents a theoretical population
distribution of IQ scores with a mean of 100
and a standard deviation of 15. In a popula-
tion of 100 million, 23 million (stippled
area) will score above 130. The bottom
figure represents distribution of intelligence-
test scores with a shift of 5%, yielding a
mean of 95. Here, the number of individuals
scoring above 130 falls to 990 thousand, with
a corresponding inflation of those scoring
below 70. Source: Weiss (1990).
Figure 2.11. Effect of a 5 percent mean shift in IQ in a population of 100
million. From reference 67. A shift in mean IQ score of 5 percent in a large
population has major implications in terms of changes at either end of the
distribution curve.
40
SETTING THE STAGE: CRITICAL RISKS 41
Prenatal exposure to lead, as measured by umbilical cord blood lead levels vs early mental
development index. Low is «3 M-g/dl, medium is 6.7 jig/dl, and high is >10 M-g/dl (Bellinger et al.,
1987).
Figure 2.12. Prenatal lead exposure vs. mental development. From reference
67. A more specific index of mental development in children shows significant
depression in relation to umbilical cord levels greater than 10 micrograms/ml.
is believed to have a half-life of about thirty-six days; therefore a
single observation cannot be taken as representative of past exposure.
New methods of noninvasive detection of the lead level in bone may
prove useful in adults, but may be difficult to interpret in growing
children.
There has been remarkably little public concern expressed at the
hazard represented by lead in gasoline. There do not appear to be any
major public health societies that have taken up the issue, though in
the United States there is an Alliance to End Childhood Lead Poison-
ing. In the United States, the debate before legislators was carried out
between scientists representing industry, on the one hand, and those
with a long interest in the mineral and its adverse health effects on the
other. This debate received little significant media coverage. Unlike
the case of asbestos, the lead issue has not been the subject of detailed
litigation, and no courts have had to rule on the likelihood of a raised
42 SETTING THE STAGE: CRITICAL RISKS
Figure 2.13. Pathways of lead exposure. From reference 69. There was much
uncertainty as to which route of exposure might be dominant in the case of
children, but the fall in blood lead levels that accompanied the decline in lead
use in gasoline in the United States (see figure 2.14) indicated that direct
inhalation or indirect contact with automobile lead must have been a domi-
nant pathway.
blood lead level being a cause of mental retardation in an individual
case. Recently, the lead industry has apparently decided to mount a
campaign questioning the integrity of the epidemiologists who have
studied its effects in children.
Before the decision was made to remove lead from gasoline, the
U.S. Environmental Protection Agency (EPA) summarized the ad-
verse health effects of lead, particularly emphasizing the relationship
between blood lead levels and hypertension that had been found in a
random sample of men in a national survey. The epidemiological data
on children were noted, but no attempt was made to base a cost/
SETTING THE STAGE: CRITICAL RISKS 43
Parallel decreases in blood lead values observed in the NHANES II Study and amounts of lead
used in gasoline during 1976-1980 (EPA, 1986).
Figure 2.14. Decreases in blood lead and use of lead in gasoline in the United
States from 1976 to 1980. From reference 69. The remarkable concomitant fall
in blood lead values—in a subset of a 50,000-person random sample in the
United States—with the reduction of lead in gasoline indicates the dominance
of this source in general public exposure to the metal.
benefit analysis on them. At the same time, the automobile industry
concluded that the newly required reductions in automobile emis-
sions, particularly of hydrocarbons, necessitated the use of a catalytic
converter on the exhaust system. This equipment could only be used
if the gasoline was lead free. There was, therefore, a conjunction
between these two aspects of lead, and both undoubtedly played a
part in the decision to reduce its use in gasoline. It does not seem
possible to answer the question of whether a decision based on health
data alone would have been to discontinue the use of lead in gasoline.
By 1985, however, the adverse health effects of lead had been more
clearly delineated, and the resulting health costs were computed at
that time by the EPA. These considerations played a role in the deci-
sion to remove lead altogether from gasoline.
The lead controversy therefore involves both the routes of uptake
and the importance of shifts in the measured IQ in groups of children;
there are special aspects in relation to biological plausibility; and there
44 SETTING THE STAGE: CRITICAL RISKS
is no significant public pressure group in any jurisdiction, and little
media interest. The belief in the importance of removing lead from
gasoline was early and strong in the United States, but later and reluc-
tant in the case of Canada and Britain. This reluctance is difficult to
explain. However, one may note that in the section on lead in the World
Health Organization publication Air Quality Guidelines for Europe (70),
published in 1987, there seems to be some confusion over the routes of
intake of lead in children, since the report does not recognize that road
dust may contain lead from gasoline. The report (on page 256) con-
cluded that in children "the contribution of direct inhalation to total
lead absorption is smaller than in adults (approximately 6% only). Most
of the absorbed dose can be accounted for by other pathways, mainly
food." This seems to have been the reasoning that underlay the reluc-
tance of British experts to believe that lead in gasoline was important.
The concomitant fall in blood lead levels with reduction of the lead in
gasoline shown in figure 2.14 indicates that lead in gasoline must be a
major factor in determining contemporary blood lead levels.
Lead in all its forms represents a signficant hazard in the develop-
ing world. According to data published by the World Bank, Karachi,
Pakistan, now has the highest lead content of gasoline of any major
city. The impact of this on the neurobehavioral development of the
children there can only be imagined.
Electromagnetic Fields
In figure 2.15 are shown the various wavelengths and energies of
electromagnetic fields (EMF) and where these fit into the energy spec-
trum (71). The fields that may be induced in such a complex structure
as the human body are extremely variable and depend on a variety of
factors too detailed for the present discussion. These considerations
underlie the difficulty of knowing how to assess exposure.
In 1979, Wertheimer and Leeper (72) published a paper in the Ameri-
can Journal of Epidemiology in which they reported finding a significant
association between leukemia in children and the wiring configura-
tion of the electrical distribution lines supplying the houses in Denver
in which the children were living. The wiring configuration was
thought to be a surrogate for the EMF exposure to which the children
"had been subjected. The authors wrote a very cautious interpretation
of their data. In table 2.1,1 have dated the initial public concern about
Figure 2.15. The electromagnetic spectrum. From reference 71. The second arrow from the
left indicates the energy range of EMF exposures to overhead power lines.
46 SETTING THE STAGE: CRITICAL RISKS
this risk to 1986, since by that date there had been a number of
supporting studies. In 1982, it was first suggested that men and
women working in occupations that might increase their exposure to
electromagnetic fields faced an increased risk of various forms of can-
cer. By 1990, the EPA had issued a draft report entitled Evaluation of the
Potential Carcinogenicity of Electromagnetic Fields (73). This report engen-
dered so much heated reaction that a subcommittee of the EPA's
Science Advisory Board was put together to review it and to make
recommendations. The subcommittee reported in December 1991
(74), and required a full rewrite of the literature review and its
intepretation.
The present status of this issue is one of exceptional interest, as will
be evident from the following observations.
1. The visual presentations of risk ratios for electrical workers pre-
pared by Theriault (75; see figures 2.16 and 2.17) indicate a clear
preponderance of studies with risk ratios greater than 1, and in some
cases as high as 2.8. Theriault points out that case-control studies
have generally given a positive result, but cohort studies have been
uniformly negative. This is unexplained.
2. In the case of childhood leukemia, two recent studies, one in
Denver and another in California, have confirmed Wertheimer and
Leeper's original observation. The recently published study by Lon-
don and Peters and their colleagues in Los Angeles (76) analyzed 219
cases of leukemia in children and compared them to 232 controls.
Measurements of magnetic and electrical fields were made in the
homes of 164 of the cases and 144 of the controls; wiring configuration
was noted in 219 cases and 109 controls. The authors reported an
odds ratio of 2.15 (95 percent confidence limits 1.08-4.28) for very
high, relative to very low, current and ground configuration data (see
table 2.9). (An odds ratio of 2 means that a risk is doubled. The
confidence limits define the possible lower and upper limits to the
incurred risk, one indicating no increase, and 4.0, for example, indicat-
ing a fourfold increase.) There was no association with the measured
electromagnetic fields in the child's home. The cases were more likely
than the controls to report use of several appliances in the home that
produce high electric and magnetic fields.
This study therefore confirmed the Denver data of Wertheimer and
Leeper showing a significant association between wiring configura-
tion and leukemia risk. The authors noted,
SETTING THE STAGE: CRITICAL RISKS 47
Figure 2.16. Leukemia risks among electrical workers. From reference 75.
Note that two studies showed a relative risk of less than one, and, in a further
six, the lower level of the 95 percent confidence limits is below one. In view of
the absence of precise exposure data, the balance of the data might be taken
to indicate a probable relationship.
Unfortunately, there is little known about the etiology of
childhood cancer, and therefore, confounding by a factor
not known to be a disease risk is possible. Yet such a factor
would need to be strongly associated with risk or extremely
tightly correlated with wiring configuration classification to
have produced the odds ratio in our study and others.
48 SETTING THE STAGE: CRITICAL RISKS
Figure 2.17. Acute myeloid leukemia in electrical workers. From reference 75.
In three of the six studies, the risk ratio is greater than 2.0.
This study involved the most detailed attempt to exclude other con-
founders that one can imagine. The confirmation of the original Den-
ver data in another city and in a larger sample surely indicates that the
epidemiological data cannot be dismissed.
3. Recently, an excess of breast cancers in men (which are rare
tumors) occupationally exposed to EMF has been reported. This
might be linked to changes in melatonin secretion. Theriault has re-
cently written (75):
It has been proposed recently that the mechanism by
which EMF may participate in the development of cancer
SETTING THE STAGE: CRITICAL RISKS 49
Table 2.9. Leukemia Risk in Relation to Denver Wertheimer-Leeper Wiring
Configuration Classification, Los Angeles County, California, 1980-1987
Exposure
category*
Underground
Very low
Ordinary low
Ordinary high
Very high
Cases
11
20
58
80
42
Controls
11
27
75
68
24
ORt
i.oot
0.95
1.44
2.15
95% Clt
0.53-1.69
0.81-2.56
1.08-4.26
p for trend
0.008 (4 categories)
0.012 (all 5 categories)
*Mixed overhead/underground removed (eight cases and two controls).
tOR, odds ratio; Cl, confidence interval.
^Reference category is underground and very low combined. Wiring configuration classification
for the residence lived in the longest was used.
From reference 76. Data from the Los Angeles study of childhood leukemia confirming the
earlier finding of Wertheimer and Leeper in Denver.
in men is the decrease in melatonin hormone in blood.
Melatonin is known to influence the development of hor-
mone sensitive cancers in animals. . . . Recent findings of
excesses of male breast cancer among electrical workers
and of excess of skin melanoma in cohorts of telecom-
munications, electrical and electronic workers may be a
serious indication that EMF may indeed contribute to the
development of cancer through the decreased melatonin
pathway.
4. Hutchison (77) considered that central nervous system cancer in
children represented the strongest association with residential expo-
sure to EMF.
5. The Nonionizing Electric and Magnetic Fields Subcommittee of
the EPA Science Advisory Board (74) convened to review the draft
report on potential carcinogenicity, concluded:
Some epidemiological evidence is suggestive of an associa-
tion between surrogate measurements of magnetic field
exposure and certain cancer outcomes. In such studies, the
existence of confounders is always a possibility, but since
no common confounder has yet been identified, the exist-
ing evidence cannot be dismissed.
50 SETTING THE STAGE: CRITICAL RISKS
6. The most relevant biological data on the effects of electromag-
netic fields seem to be those indicating some interference with the
secretion of melatonin, a hormone produced by the pineal gland that
has something to do with the circadian rhythm but that in addition
may inhibit tumor growth or formation. Physicists have argued that
EMF at 60 hertz cannot affect individual cells, but they concede that
whole glands or neural networks might be affected. This would ex-
plain why radar confuses migrating birds, and why some fish seem
capable of detecting very low energy electromagnetic fields. How-
ever, the biological plausibility for a direct carcinogenic effect in the
human cannot be said to be strong.
7. The 1992 British Medical Research Council Report (71), after an
exhaustive and excellently written review of the present status of the
whole field (including the recent Los Angeles study), came to a num-
ber of conclusions:
The experimental findings are, unfortunately, not very
helpful. It cannot be concluded either that the electromag-
netic fields have no effect on the physiology of cells, even if
the fields are weak, or that they produce effects that
would, in other circumstances, be regarded as suggestive
of potential carcinogenicity.
The results of some whole animal and cellular studies
suggest the possibility that electromagnetic fields might act
as co-carcinogens or tumour promoters but, taken overall,
the data are inconclusive.
In relation to cancer in children:
The results have been variable, but, taken at face value,
they appear to provide some weak evidence in support of
the postulated association which is less weak for brain
cancer than for leukemia and less weak when exposure is
estimated from the local "wire configurations" than from
proximity to sources of electromagnetic fields or from mea-
surements in the home.
This evidence is, however, difficult to accept as the meth-
odology of the most important studies failed to obtain sets
of control data that could be confidently regarded as repre-
SETTING THE STAGE: CRITICAL RISKS 51
sentative of the populations from which the affected chil-
dren were drawn. The major positive results may in conse-
quence be artefacts of the method of enquiry. In summary,
the epidemiological findings that have been reviewed pro-
vide no firm evidence of the existence of a carcinogenic
hazard from exposure of paternal gonads, the fetus, chil-
dren, or adults to the extremely low frequency electromag-
netic fields that might be associated with residence near
major sources of electricity supply, the use of electrical ap-
pliances, or work in the electrical, electronic, and telecom-
munications industries.
The main criticism of the Los Angeles study was directed at the method
of selecting the controls and in the uncertainty over whether wiring
configuration actually represented any significant difference in electro-
magnetic exposure. Sir Richard Doll (the chair of the committee) is
quoted later as saying, "It is, we suggest, important that more research
is carried out to clarify as soon as possible whether extremely low
frequency magnetic fields in the home and at work are capable of
causing cancer." This report therefore was significantly more negative
in relation to the existing epidemiological data than was the report of
the subcommittee of the EPA's Science Advisory Board or the reports
by Hutchison and Theriault referred to above. One might wonder why
it is important that more research be carried out on the question if the
present epidemiological data can be dismissed as nonsubstantive.
8. Two recent reports, contained in manuscripts not yet published
in the open literature, are not reassuring. In a Swedish case-control
study of people who had lived for at least one year within 300 meters
of a 200 or 400 kilovolt transmission line (78), an increased relative risk
of 2.7 for childhood leukemia was found. In another large Swedish
study of workers occupationally exposed to electromagnetic fields
(79), a significant excess of chronic lymphatic leukemia was found. It
was not considered that confounding by exposure to solvents or ben-
zene could have occurred. Though these two studies have not yet
appeared in the scientific literature, they will both reinforce previous
data when they are published.
9. There is no doubt that difficulties in measuring exposure, or in
knowing what aspect of exposure might be important (e.g., mean lev-
els, transients, lifetime averages) not only make interpretation of the
52 SETTING THE STAGE: CRITICAL RISKS
epidemiological data difficult, but also cause problems in designing
appropriate animal studies. All are agreed that refinement of exposure
data, particularly in those occupationally exposed to electromagnetic
fields, is of first importance. It must be remembered that exposure
misclassification is likely to have the effect of weakening a demon-
strated relationship. Indeed, one can view the epidemiological data
and remark that it is astonishing that any relationship should have
been found in view of the obvious and serious weakness of exposure
classification.
There has been considerable public interest in this issue. A recent
book (80) has summarized various enquiry processes and events in
relation to electromagnetic fields; the author alleges a major conspir-
acy in the scientific community to prevent the truth being known. In
many areas, there have been hearings and inquiries on the location of
high-voltage power transmission lines, and in British Columbia the
hydroelectric authority has agreed to purchase, at fair price, houses
that lie very close to necessary new transmission lines. Some scien-
tists who have long been involved with electromagnetic fields have
urged a policy of "prudent avoidance" until the facts become better
established. Major new research initiatives to clarify animal effects,
and to extend the epidemiological observations, have been launched.
New technologies are being developed to permit better exposure eval-
uation. All of these steps should be useful in resolving the present
dilemma. It should be noted that the electric blanket manufacturing
industry, once it had been shown that the blankets were capable of
generating strong magnetic fields, redesigned them, with the result
that modern blankets now deliver a negligible dose.
The electric generating industry has financed efforts to question the
validity of the epidemiological data, but these have not included overt
attempts to discredit the epidemiologists who have conducted such
studies.
Electromagnetic field effects pose, in acute form, the problem of
interpreting epidemiological associations when there is little biologi-
cal plausibility and no demonstrated dose-response relationship. This
problem receives more discussion in the next chapter.
In the case of electromagnetic fields, therefore, the reason for con-
cern is based entirely on the epidemiological evidence; there are very
difficult problems in exposure assessment; biological plausibility is a
major problem; there is much media interest; and significant indus-
SETTING THE STAGE: CRITICAL RISKS 53
trial lobbies are involved. The question of whether leukemia is a
multifactorial disease, with EMF exposure being one factor, cannot
yet be settled.
Summary
This brief perspective on the five hazards I have chosen to discuss
permits some preliminary generalizations.
As far as epidemiological studies are concerned, in the case of
cigarettes the overwhelming associative evidence is accepted as indi-
cating a causal link to lung cancer and emphysema. There is basic
biological plausibility, a clearly demonstrated dose-response relation-
ship to risk, and a very large number of studies in different countries
that confirm the strong associative relationship. Cigarette smoking is
also recognized as one risk factor in the multifactorial nature of heart
disease.
In the case of air pollution, different types of epidemiological stud-
ies are used to describe adverse effects; each of these has strengths
and weaknesses, as noted in the next chapter. Only when a major
single episode occurs, as in London in 1952 or in Donora in 1948, is
the evidence so strong that the causal link cannot be disputed. In
more general cases, cross-sectional studies or time-series analyses re-
quire detailed discussion and replication before it is reasonable to
infer a causal relationship and before general confounders can be
excluded. For some pollutants, such as ozone, there is a strong basis
of biological plausibility; in the case of particulate pollution or the
possible effects of very low exposures to sulfur dioxide, no biological
mechanisms are yet known to explain the associations that have been
described. In the case of particles less than 10 microns in size, the
regression against mortality on a daily basis is so strong, and has been
demonstrated to apply in cities so different in climate and population,
that there are strong grounds for concluding that the relationship is
causal, even in the absence of any known biological mechanism.
The general relationships of asbestos to lung cancer and mesothe-
lioma have been well established. The more difficult questions of risk
in relation to fiber type, and of defining general levels of public expo-
sure, remain to be fully elucidated. Much of the present controversy
revolves around these questions.
Epidemiological studies in relation to lead, although reasonably
54 SETTING THE STAGE: CRITICAL RISKS
comprehensive and almost unanimous in relation to the hazard it
represents for young children, still occasion different interpretations.
Some take the position that the effect is too small to be of conse-
quence; others disagree.
The epidemiological data about electromagnetic field exposure
tend to point in the same direction, but their interpretation is made
difficult by the lack of any dose-response relationship and by very
weak biological plausibility. No common confounder in all the studies
has been identified.
The general questions raised by epidemiological studies are dealt
with in detail in the next chapter. It is clear, from four of the five
hazards discussed, that the interpretation of epidemiological data is
central to understanding how each of them should be approached.
Only in the case of cigarettes is the data complete, though of course
this was not the case when the first studies were published.
Public involvement in these five hazards has varied from intense,
in the case of cigarettes and electromagnetic fields, to episodic in the
case of air pollution, to surprisingly almost absent in relation to lead.
Environmental exposure to asbestos, though this has received inter-
mittent attention centering around publicized litigation, in general
has not become a major public issue.
The role of the media generally reflects the level of public concern
(indeed it is difficult to separate these two aspects). The media also
have the effect of focusing public attention on an issue, and may be
said to stimulate public concern. In the case of cigarettes, the printed
media were initially inhibited by the threat of withdrawal of cigarette
advertising, which is an important source of revenue. Such publica-
tions as the original U.S. Surgeon General's report and subsequent
yearly follow-up reports have subsequently ensured wide media cov-
erage. Surveys indicate that most members of the public in the United
States are now aware of the dangers of cigarette smoking. The decline
in smoking, though slow to start, may be attributed to government
intervention based on public concern. It is interesting that there was a
gap of several years between the first publication of summaries of the
epidemiological data (as in the Surgeon General's report of 1963) and
the development of public interest groups to represent the interests of
nonsmokers. These groups became much more active after the effect
of passive smoke inhalation began to be publicized.
SETTING THE STAGE: CRITICAL RISKS 55
In some regions, and in relation to some types of air pollution, the
media have played an important role. This was especially true in
relation to the 1952 London episode, when media pressure drove a
reluctant government to set up a committee to study the question. In
Los Angeles, and in the question of acid rain in the northeastern
United States and Canada, the media have played an important role
in public education.
Media attention to asbestos has been confined to publicizing indi-
vidual cases of mesothelioma or lung cancer in those who have been
occupationally exposed. This publicity undoubtedly has had an effect
on juries charged with the assessment of responsibility and the award-
ing of damages. There has been little general media coverage of the
risks that may follow general environmental exposure to asbestos.
It is surprising that very little public discussion has taken place
about the question of the effects of lead at low-level exposures. This
may be because individual cases cannot be highlighted. The point will
be made in the next chapter that it is one thing to say that a commu-
nity's exposure to lead is too high and should be reduced, and quite
another to tell parents that the low IQ of their child has been caused
by their living close to a main traffic artery.
The media have been very much involved in the question of electro-
magnetic fields. Wherever effects have been attributed to high-
tension transmission lines or to the location of a transformer station at
the end of the road, the media have been willing to publicize the
public concern aroused. As noted later, this coverage has generally
not been sophisticated. The authors of the most recent study of elec-
tromagnetic fields and childhood leukemia (76) were blamed for a
newspaper's apparently biased coverage of their paper. In a letter to
the editor of the American Journal of Epidemiology in which their study
had been published, they observed:
Our paper was written for the scientific literature, not the
public media. We assumed that those seriously interested
in the issue would read the entire paper. There is limited
space in an abstract, and it is impossible to word it to avoid
all possible opportunities for the media to misrepresent the
findings. Furthermore, very few of the reporters who con-
tacted us had even read the abstract. We made every at-
56 SETTING THE STAGE: CRITICAL RISKS
tempt to clarify the results in our dealings with the media
and have come to the conclusion that the media presenta-
tion is largely beyond our control.
In the case of asbestos, the discussion of the interpretation of exist-
ing data has been carried on almost entirely in relation to litigation. The
same is true to a lesser extent in relation to electromagnetic fields. The
involvement of courts and individual law firms has had various conse-
quences that have not been evident in relation to the other hazards.
As might be expected, the role of industry and industrial consortia
has been defensive. Such groups command enormous financial poten-
tial, not just that derived from the sale of their products, but also from
the power given them over governments by virtue of taxation income,
and, in the case of tobacco, over the media by virtue of advertising
revenue.
Before any general conclusions can be reached in relation to these
five hazards and the lessons to be drawn from them, there are six
issues that require detailed discussion. These are the subject of the
next chapter.
3
Mandated Science: Major Issues in Health
and Public Policy
In her text with this title (58), Salter describes "mandated science" as
involving the relationship of science, values, and public policy, and
illustrates this by a Venn diagram (see figure 3.1). This provides a
useful framework for the discussion of the interrelationships involved.
Over the period of the present review, covering approximately the
past forty years, scientists have generally become conscious of the
importance of distinguishing between their statements that are based
solely on a description of the scientific data, and other judgments they
may make, which are just that. They have, in general, learned the
necessity of being sensitive to the difference between their opinion on
an issue and an analysis of the available data that is not weighted by
value judgment. It is evident that an inference of causality from a
demonstrated association requires a judgment; it is important to treat
it as such.
A number of authors have written on the intersection between
science and values. The valuable volume by Loren Graham (81), writ-
ten in 1981, provides an excellent starting point. A recent interesting
discussion of the relationship between values and judgments particu-
larly relevant to environmental concerns is that by Steensberg (82) in
his analysis of environmental decision making in Denmark.
It is important to recognize that both policy decisions, such as
removing lead from gasoline or protecting the public from high-
tension transmission lines or from ozone exposure, and all discus-
sions of standard setting take place within the area between "values"
and "science." Such discussions require, first, agreement as to what
57
58 MANDATED SCIENCE: MAJOR ISSUES
Figure 3.1. "Mandated Science." From reference 58. Note that "mandated
science" exists in the overlapping zone between "science," "policy and regula-
tion," and "values." Many scientists find this position an uncomfortable one.
the scientific data say and the existence of a causal relationship; and,
second, statements of the "values" or economic considerations that
underly the final decision. These must be made explicit and discussed
on their own. What often happens is that the interpretation of the
scientific data is skewed (in one direction or another) on the basis of
(unstated) "values" held by the speaker.
Salter also addresses another important question when she writes:
What seems to be true is that the problem of combining
science and advocacy occurred when scientists and their
MANDATED SCIENCE: MAJOR ISSUES 59
work were being evaluated outside the scientific commu-
nity and by non-scientists. It is this observation that leads
us to identify the problem in the relationship between sci-
ence and advocacy not in the conduct of the research but in
its presentation to a non-scientific community.
This touches on the relationship between science and the law, dis-
cussed later.
What has happened in the past forty years is that "science" has
perforce moved out into the community. Certainly it has been central
to the issues of environmental epidemiology with which this text is
concerned. In discussing this background, Salter writes, "This
quandary—the increasing dependence upon science and technology
and an increasing concern or skepticism about the kind of answers
that science itself can provide—is the heart of the matter." It is for
this reason that the precise interpretation of environmental epidemio-
logical data is central to these issues of environmental hazards.
There is another aspect to mandated science that has to do with the
implications of science that is not conducted as "free inquiry" but that
is conducted to elucidate some problem that has become important.
In this category are the controlled studies of the acute effects of air
pollutants, or the attempts to study the effects of lead or electromag-
netic fields on animals. This type of science is directly concerned with
the setting of standards for the protection of public health; indeed,
Salter's text is subtitled Science and Scientists in the Making of Standards.
This question is discussed in Chapter 5.
Statistical Associations
In his history of statistics, subtitled "The Measurement of Uncertainty
before 1900, Stigler (83) notes that Yule's first use of a correlation coeffi-
cient (later refined by Karl Pearson) was in the study of characteristics
of the poor and criminal classes. Its use for this purpose was chal-
lenged by Pigou in written testimony before a Royal Commission on
the Poor Law. Pigou based his criticism on the possibility of an unmea-
sured "lurking variable" interfering with what Yule had suggested
was a causal relationship. Thus the argument concerning the meaning
and interpretation of a statistical association began right at the start of
its formulation.
60 MANDATED SCIENCE: MAJOR ISSUES
The Pearson correlation coefficient (see table 2.2) is the simplest
way of determining that a relationship between two variables is not a
random distribution. A more sophisticated method involves the calcu-
lation of a regression coefficient (which may be shown to be statisti-
cally significant) between the two variables. Such coefficients, which
relate relative incremental changes in the variables, are called "elas-
ticities" by economists. When used in relation to air pollution data, as
an example, these elasticities become changed to the form, "A 50
percent fall in the level of atmospheric sulphates would lead to a 15
percent reduction in hospital admissions for acute respiratory dis-
ease." Causality is assumed.
In most modern epidemiological work, the design of the study
involves 1) data collection to account for all known variables; 2)
correction of the basic data for these variables by filtering of one kind
or another; and 3) the model's prediction of the relationship between
the two variables of interest, such as the wiring configuration of
houses and leukemia, or particulate pollution and mortality, or blood
lead and IQ level. This application of logistic regression and model-
ing may become so complex that it is easy to lose sight of the fact
that, at the end of it all, what is being assessed is the robustness of
an association—that is, its ability to withstand the taking into ac-
count of other factors.
In some studies, such as the study of associations between atmo-
spheric variables and hospital admissions, certain difficulties can be
avoided by the design of the database rather than by later statistical
manipulation of the data. Thus, in the case of the Ontario study (23),
each summer can be considered separately to avoid long-term effects,
or the respiratory admissions on a given day of the week can be
compared only to the mean value for the same day of the week, and
so on. Such preliminary treatment of the data avoids having long-
term effects produce a spurious relationship, which is always a haz-
ard of associative analysis.
With all environmental hazards, the existence of unidentified con-
founders is invariably a possibility. Cross-sectional comparisons of one
population with another (such as symptom prevalence in children in a
high-pollution region compared to children in a low-pollution region)
are always complicated by other differences between the two
populations—in this case in socioeconomic circumstances, in housing
quality, in gas cooking, in maternal cigarette smoking, and in other
MANDATED SCIENCE: MAJOR ISSUES 61
factors. These difficulties in cross-sectional comparisons make the lon-
gitudinal time-series studies—based on variations from day to day—
somewhat easier to interpret. These studies state that, in a given popu-
lation, there is an association between one event, such as elevated
pollution, and another, such as hospital admissions or mortality. The
possible role of confounders is often misunderstood. When I submitted
my first paper on the air-pollution-to-hospital-admissions relationship
in southern Ontario, one reviewer objected that I had taken no account
of the cigarette smoking of the population. A moment's thought shows
that in this kind of study, this objection is irrelevant. This is because the
significant association is a time relationship between pollutant and
number of hospital admissions.
In the case of electromagnetic fields, the studies on the association
between home wiring configuration and childhood leukemia might be
affected if there was a chance association between wiring configura-
tion of the house and its being adjacent to major traffic arteries. In the
occupationally exposed cohorts, exposure to organic chemicals or sol-
der in the course of electrical work might well occur but would not be
accounted for in the studies. Similarly, the association between blood
lead levels in children and their IQ has been dismissed on the
grounds that blood lead is possibly elevated in children in whom
some other factor is operative, such as low family income, and it is
this that drives the association. This possibility appears to have influ-
enced the conclusions of the committee of the Royal Society of Can-
ada that considered the issue.
Inferring Causality from Associations
It is clear that the; problem of inferring causality from demonstrated
associations lies at the heart of the debates on the five hazards. This
question was first discussed in any detail in a series of articles in the
Journal of Chronic Diseases in 1959 and 1960 (see Labarthe and Stallone's
article in reference [84] for this). However, the problem of concluding
that causality underlies a demonstrated association was brought into
prominence in a presidential address by Sir Austin Bradford Hill (85) to
the Royal Society of Medicine in London in 1965. Hill had made many
distinguished contributions to statistical theory and practice. He did
not lay down specific or rigid "laws" that should govern this process,
but rather listed factors that he suggested were helpful in formulating a
62 MANDATED SCIENCE: MAJOR ISSUES
judgment that the associations might be based on a causal relationship
(see sidebar). He wrote: "Here then are nine different viewpoints from
all of which we should study association before we cry causation. What
I do not believe—and this has been suggested—is that we can usefully
lay down some hard and fast rules of evidence that must be obeyed
before we can accept cause and effect." He noted that the mere postu-
late of confounders, unidentified, was not a valid reason to discount
causality. He also stressed that a public policy decision on a hazard
might have to be made before the evidence of a causal relationship was
complete.
The criteria he suggested have recently been discussed in some
detail by many authors. In 1988, a collection of essays on this question
entitled Causal Inference was edited by Kenneth Rothman (84), and this
book should be consulted by anyone interested in this important ques-
tion. The discussion of each of Hill's criteria by Douglas Weed, Ste-
phen Lanes, and others in Rothman's volume is too detailed for my
present purposes. However, some observations will clarify the pro-
cess in relation to the hazards we are considering.
In terms of the strength of an association, it is true that the stronger
it is, the better evidence it provides toward causality. However, such
factors as poor exposure measurement (as in the case of EMF), variabil-
ity in outcome (such as in IQ determinations in children), or the fact
that we are studying multifactorial disease (such as asthma), all will
act to weaken what might otherwise be a strong association. In other
words, unless exposure and outcome are well characterized, we can-
not dismiss a "weak" association as necessarily representing reality.
Indeed, Lanes (84) has argued that "the strength-of-association crite-
rion, like other causal criteria, is an untestable tautology."
The criterion of temporality requires that the exposure precede the
outcome. In arguing against a causal inference in relation to cigarettes
and lung cancer, Sir Ronald Fisher suggested that the lung irritation
caused by the cancer induced the desire to smoke, and hence that the
cancer preceded the smoking (30).
In air pollution epidemiology, consistency is an important criterion.
If particulate levels in time-series analyses are related to daily mortal-
ity in Chicago for example, the same should be true in Detroit, or
Philadelphia, or wherever else the type of pollution is similar.
I have recently argued that coherence within different indices of
adverse air pollution effects should be an important criterion of causal-
MANDATED SCIENCE: MAJOR ISSUES 63
Criteria for a Causal Inference (from reference 85)
STRENGTH OF ASSOCIATION
CONSISTENCY: Observed by different people, in different places, circum-
stances, and times
SPECIFICITY: As in occupational exposures
TEMPORALITY: Exposure predates the outcome
BIOLOGIC GRADIENT: Dose-response curve
BIOLOGICAL PLAUSIBILITY: "This is a feature I am convinced we can-
not demand" (Bradford Hill)
COHERENCE: Histopathological evidence (from smokers)—consistent
with biology of the disease
EXPERIMENT: Effect of preventive action
ANALOGY: Argument from known effects from another agent
ity (86). Thus, if a particular pollutant is shown to be related to mortal-
ity in a time-series analysis, it would follow that the same pollutant
should be related to hospital admissions. This would necessarily be
the case unless all the excess mortality occurred at home, or if no
disease was caused that was not fatal, both unlikely circumstances. In
the 1952 London episode, not only was mortality increased, but there
was a major concomitant increase in the demand for hospital beds,
and visits to a family practice in the London area for acute respiratory
disease increased (11). This type of coherence between different indi-
ces of morbidity and mortality should play an important part in the
assessment of the causal relationship that may underlie time-series
associations.
Biologic gradient, or the demonstration of a dose-response effect, is
important, and was clearly evident in relation to cigarette smoking
and outcomes such as lung cancer and emphysema. In general, it
should be possible to show that the higher the exposure, the greater is
the risk. It also applies to the effects of raised blood lead levels.
Biological plausibility commonly causes difficulty in the interpreta-
tion of epidemiological evidence. Bradford Hill (85) wrote, "This is a
64 MANDATED SCIENCE: MAJOR ISSUES
feature I am convinced we cannot demand." Some observers have
declined to entertain a causal hypothesis in the case of EMF, because
there is no biologic plausibility. Others argue that our knowledge of
the process of induction of cancer is not yet complete enough that we
can discount the influence of modifications in body function induced
by EMF, even though these do not in themselves cause cancer. In the
case of air pollution, damage from ozone is inherently biologically
very plausible (18) because it has been shown to be an irritant gas that
produces inflammation in the lung at very low concentrations. How-
ever, we do not know the mechanism that might link low levels of
sulfur dioxide to adverse health effects, nor do we understand how
small particulate pollution may produce its biologic effects. It is the
different emphasis placed on biological plausibility that underlies
many of the differences in judgment in relation to these issues.
Lead, in my opinion, is a special case. Although in general it is
accepted that lead is a neurotoxin, we cannot expect to understand
the precise way in which it may affect children's behavior or the
development of intelligence, because our basic understanding of the
growth and development of the brain is still too primitive. In this
instance, therefore, it is reasonable to argue that the lack of biologic
plausibility—if by that we mean a full understanding of the basic
mechanism—should not be used as an argument against the causality
of the association.
It is undeniable that people differ in their readiness to assume a
causal relationship from an association. In table 3.1 I have suggested
some biases in this process. If self-knowledge is the beginning of
wisdom, this table should be useful. The public is generally biased
towards making a causal inference. This comes about, I think, be-
cause we always search for explanations of events that occur to us.
Samuel Pepys in his diary attributed a heavy cold to the fact that he
had washed his feet two weeks before. Such a human tendency can
be taken to extremes by those members of the public who have a
"conspiracy theory" handy for every event. The media reflect this
tendency and amplify it. Being unaware of the problems inherent in
arguing causality from associations—and having too little time or
space for a measured discussion—the causality is assumed if anyone
postulates it. In my experience, bureaucrats in the parliamentary sys-
tems of the United Kingdom and Canada are generally biased against
making a causal inference; the slow acceptance of the need to remove
MANDATED SCIENCE: MAJOR ISSUES 65
lead from gasoline in both countries illustrated this. Their "cousins"
in the United States are, in my experience, more likely to be biased the
other way—at least this was the case in relation to EMF. I have noted
that animal toxicologists, who are able to control all variables in their
experiments, generally seem to be biased against making causal infer-
ences unless the mechanisms are fully understood. They, and others,
may also be influenced by the idea that, since good science is necessar-
ily cautious, being more cautious represents better science. This kind
of primitive thinking often passes as wisdom.
The phrase "type 1 error" is often used to describe the error of
saying that an association is significant when it is not; a "type 2 error"
is to fail to reject a false null hypothesis—that is, to deny it is signifi-
cant when it really is. Both of these phrases, strictly defined, apply
only to the assessment of the strength of associations, but they are
often loosely applied to denote that a true causal relationship has
been rejected (type 2 error), or that causality has been falsely con-
cluded from an association (type 1 error).
The Problem of Multifactorial Disease
Synchronous with the wider dispersion of all the hazards under con-
sideration was the growing realization that a great deal of disease was
"multifactorial" in origin. The belief that a single disease necessarily
had a single cause died slowly. In one sense, the Western world,
having successfully combated diseases such as pneumococcal pneu-
monia, malaria, and tuberculosis, in which single dominant agents
could be identified, was left with a large residuum of diseases that
have many origins.
Thus, in the case of coronary artery disease, the leading cause of
mortality in the Western world, it was recognized that genetic factors
expressed as familial predispositions increased the risk; that prema-
ture death was related to body weight; that hypertension worsened
the prognosis; that cigarette smoking increased the risk; that dietary
fat intake, in some people at least, also increased the risk; and so
forth. Lung cancer incidence is dominated by cigarette smoking, but it
is now recognized that the risk of lung cancer in a smoker is increased
by exposure to radon, to asbestos, to arsenic, and possibly to air
pollution. Bronchial asthma certainly has a genetic basis in children,
but it is now recognized that this chronic condition may be acquired in
66 MANDATED SCIENCE: MAJOR ISSUES
Table 3.1. On Making a Causal Inference from Associations
Public at large Biased in favor
Media Biased in favor
Bureaucrats (Parliamentary) Biased against
Bureaucrats (US) Biased in favor
Affected industry Biased against (vested interest)
Scientists (who have worked in May be biased either for or
the field in question) against
Scientists (who have not worked Should be able to form indepen-
in the field) dent judgment
Animal toxicologists Biased against
Lawyers Depends who the client is
Juries Likely to give benefit of the
doubt to any victim
later life as a result of occupational exposures to such agents as west-
ern red cedar dust or diisocyanate compounds, or as a consequence of
respiratory infections. A worsening of status in this disease might
occur through a number of different influences, of which air pollution
is but one.
These diseases, and many others, represent a much more complex
world than the public or the media currently understand. In relation
to known multifactorial disease, there are a number of important
constraints, as follows:
1. It is not possible to assign percentage proportional "responsibil-
ity" for different factors. This has been excellently illustrated by
Saracci (87) in the case of cigarette smoking and asbestos exposure.
Using table 3.2, which gives the proportional incidence of lung cancer
in smokers and nonsmokers who have been exposed or not exposed
to asbestos, Saracci noted that the excess rate in those exposed only to
asbestos is 47.1, but it is 590.3, or 12.5 times greater, in the group
exposed to both agents. He continued:
Should the analysis stop here, the conclusion would al-
most automatically follow that in a situation as that por-
MANDATED SCIENCE: MAJOR ISSUES 67
Table 3.2. Lung Cancer Risk with Smoking and Asbestos Exposure
% Excess Rate
Exposure
Group
A- S-
A+ S-
A- S+
A+ S+
Death
Rate
(1)
11.3
58.4
122.6
601.6
Excess
Rate
(2)
0.0
47.1
111.3
590.3
Removable by
Eliminating
Smoking
(3)
—
100.0
92.0
Asbestos
(4)
100.0
—
81.2
From reference 87. See text for the correct interpretation of these data.
trayed by the death rates in column (1), smoking is by far
the dominant factor, and its removal is, correspondingly,
by far more effective as a preventive measure than asbestos
removal. That things are not so simple is indicated by col-
umns (3) and (4). If smoking is removed, the excess rate
will go down from 590.3 to 47.1 i.e. some 543 lives will be
saved (per 100,000 person years) or 92% of the excess rate
will be removed. If, on the other hand, asbestos is re-
moved, the excess rate will go down from 590.3 to 111.3,
i.e., 479 lives will be saved, or 81.2% of the excess rate will
be removed. This means that removal of smoking is more
effective than removal of asbestos, but only 1.14 times (543/
479) and not 12.5 times as it superficially appeared.
2. It follows from this that, when we know that a disease is
multifactorial in origin, we should be very cautious of any statement
telling us that 90 percent of the cases are caused by smoking. What
might be the case is that 90 percent of the cases would not have
occurred without smoking, but this does not mean that only 10 per-
cent of cases are caused by something else. It might be that 85 percent
of the cases would not have occurred if another agent had not been
present, but people continued to smoke. One finds statements in the
press or on television programs that do not recognize this basic limita-
tion when a disease is known to be multifactorial. We are confidently
told every day, "Obesity is responsible for 20 percent of all heart
attacks"; or, "5 percent of all cancers are occupational in origin." Harri-
son and Hoberg (88) have recently documented this error in the case
68 MANDATED SCIENCE: MAJOR ISSUES
of radon, noting that a Canadian Health and Welfare official "as-
sumed that 90% of lung cancer deaths are caused by smoking, and
that 5% of lung cancer deaths occur 'spontaneously/ leaving a resid-
ual of up to 5% which could possibly be ascribed to radon exposure."
3. Checkoway and Hickey (89) have presented an elegant statisti-
cal and algebraic analysis of this problem in multifactorial disease.
Their paper should be required reading by students in environmental
health sciences.
4. There is no doubt that this problem leads to serious difficulties
in disease attribution. All those who have developed risk estimates of
cancer from asbestos exposure agree that the number of lung cancers
induced will be at least twice as many as cases of mesothelioma. In
British Columbia, in the years 1980 to 1988 inclusive, the Workers'
Compensation Board recognized sixty-nine cases of mesothelioma.
During the same period, only eleven cases of lung cancer were recog-
nized in which asbestos exposure was considered to have contributed
to the risk. This disparity was not because cases were rejected by the
board, but because cases of lung cancer were never brought forward,
since the doctors presumably thought that smoking alone had caused
the disease. The question of whether the patient would have devel-
oped lung cancer without asbestos exposure was presumably never
asked. It was calculated in 1986 that approximately two thousand
mesothelioma and between four thousand and six thousand lung
cancer deaths that would not have occurred without asbestos expo-
sure were occurring annually in the United States (90).
5. In both the United States and Canada, mortality from heart
disease has fallen over the past eight years. Do we know how much of
this change is to be attributed to reduced smoking, or to jogging, or to
less fat in the diet, or to better immediate care? All we know is that we
cannot ascribe the effects in percentages so that they will add up to
100 percent. For similar reasons, we cannot apportion the factors
responsible for asthma in order to make statements like, "5 percent of
asthma is due to air pollution." This principle operates in other fields.
In a recent article (91), Stephen Jay Gould wrote the following note in
an article on evolution:
Incidentally, the concept of emergence helps us to under-
stand why the nature-nurture issue is such a false dichot-
omy. Genes influence many aspects of human behavior,
MANDATED SCIENCE: MAJOR ISSUES 69
but we cannot say that such behavior is caused by genes in
any direct way. We cannot even claim that a given behavior
is, say, 40 percent genetic and 60 percent environmental,
and thereby defend at least a partial old-fashioned genetic
determinism. Genes and environment interact in a
nonadditive way, yielding emergent features in the result-
ing anatomies, physiologies, and behaviors.
6. Rose (92) has pointed out that, if we imagine a country in which
everyone smoked ten cigarettes a day, we would conclude that lung
cancer had a genetic basis. In a sense this would be true, since with
everyone smoking but only some getting lung cancer, some factors
must determine which members of the population get the disease and
which do not. The more widely distributed a factor is, the more diffi-
cult it is to determine its role in any multifactorial disease in which we
know it exerts an influence. In Chapter 2, the question was asked,
"How we can determine the present role of environmental asbestos
exposure in relation to lung cancer, when the lungs of city dwellers
not occupationally exposed to the material are found to contain mil-
lions of asbestos fibers?" With 86 million Americans currently ex-
posed to levels of ozone higher than the EPA standard, how are we to
determine whether this is influencing the prevalence or severity of
asthma? These are complex questions because of the difficulty of de-
scribing multifactorial disease.
7. Understanding the problem of multifactorial causation is very
important when media reporters, with their short attention span and
thirst for the dramatic statement, try to badger anyone not on his or her
guard into saying, "50 percent of all cases of asthma are due to . . ."
8. Another issue is the common, but almost certainly false, as-
sumption that diseases we do not currently understand are not
multifactorial in origin. Have we any reason to assume that childhood
leukemia has one cause? Theriault (75) noted the possibility that EMF
was one of a number of factors that increased the risk of cancer.
Possibly it is the combination of EMF exposure and concomitant expo-
sure to hydrocarbons from automobile or diesel exhaust. If so, these
factors might not be "confounders" in the ordinary sense, but addi-
tional factors contributing to the risk.
Once the reality of multifactorial disease is grasped, many oversim-
plified statements (even some made by distinguished people) can be
70 MANDATED SCIENCE: MAJOR ISSUES
identified as essentially misleading. There is no doubt that a sophisti-
cated approach to multifactorial disease is required when most issues
in environmental health are being considered.
The Role of Risk Assessment and Perception
The past twenty years have seen a major expansion in formal risk
assessment activity. Corn (93) has recently written an excellent review
of the developing role of risk assessment in relation to occupational
hazards between 1970 and 1990. A complete review of this discipline
would not be germane to my present purpose. However, the process
of risk assessment has become so central to any consideration of pub-
lic hazards, and is so likely to affect future policy decisions, that some
description of it is obligatory. The following points will serve as an
introduction.
1. It is not possible to conduct a formal risk assessment unless there
is some capability of constructing a dose-response relationship. The cal-
culations of public risk from environmental asbestos exposure shown
in table 2.7 depend on downward extrapolation from occupational ex-
posure data. Risk calculations in the absence of such information (as is
the present case in relation to EMF) can only be expressed as risk ratios.
2. As shown in table 2.8, there may be a considerable range of risk
estimates. In such instances, a mean value may be taken as broadly
representative of the whole. Rosenthal, Gray, and Graham (94) have
recently published a detailed and scholarly review of the process of
risk assessment as applied to carcinogenic chemicals. They wrote:
In summary, in spite of its appearance of precision, QRA
(quantitative risk assessment) is fraught with gaps in
knowledge that are filled with guesses and assumptions.
Risk assessors have a great deal of analytical discretion in
the conduct of cancer risk assessments. As we have dis-
cussed, the choice and interpretation of data, the choice of
extrapolation models, and the choice of exposure assump-
tions and models can make a huge difference in the out-
come of a risk assessment. .. .
Regardless of one's theory of democracy, our analysis of
the risk assessment process suggests that mandated risk
levels per se would do little to assert democratic control
MANDATED SCIENCE: MAJOR ISSUES 71
over the standard setting process. The numerous semi-
technical, semipolicy judgements pervasive in the calcula-
tion of carcinogenic risk could frustrate any congressional
attempt to control regulatory decisions through specifica-
tion of risk levels.
We saw, for example, that alternative choices of exposure
assumptions and dose-response models can lead to plausi-
ble risk estimates that vary by several orders of magnitude.
If agency officials believe that a statutory bright line is too
stringent in a particular case, they can manipulate the risk
calculation to produce a numerical estimate of risk that will
allow them to justify their desired level of stringency.
3. Table 3.3 shows the odds ratios for two different levels of cumu-
lative exposure to coal dust in a large cohort of United States coal
miners (95). It can be seen that, in relation to symptom prevalence
and lung function, increased cumulative exposure leads to a greater
risk of depressed lung function and of various symptoms. When large
data banks exist that can be analyzed in this way, a convincing esti-
mate of risk relation to exposure can be displayed. Such estimates,
based on reliable exposure and outcome data, do not suffer from the
difficulties in risk estimation noted in the paragraph above.
4. Major difficulties in risk estimation occur when animal toxico-
logical data (often concerning induced cancers) are to be converted to
a human risk estimate. In view of known species differences, it is to
be expected that these limitations will continue.
5. Long latency periods, such as that existing in the case of
mesothelioma following asbestos exposure, present difficulties, as the
period of observation may be too short for all the induced cases to be
detected. For example, a recent report from Goteborg (96), in a pro-
spective study of shipyard workers seven to fifteen years after all
exposure to asbestos had ceased, found that mesotheliomas were still
occurring at approximately ten times the expected rate.
The development of risk estimates has been paralleled by interest
in the public perception of risk and what influences it, and more
recently in "risk communication." Upton (97) has recently summa-
rized much of this material, and table 3.4 is taken from his recent
discussion. The differentiation between characteristics that seem to
72 MANDATED SCIENCE: MAJOR ISSUES
Table 3.3. Odds Ratios for Exposure to 1 and 20 mg/m
3
-Years
Coal Mine Dust Based on Logistic Regression Models
FEVj <80%
FEV/FVC <80%
Cough
Phlegm
Chronic bronchitis
Obstructive bronchitis
Breathlessness
Wheeze
Wheeze with SOB
ORj
a
 (95% C.I.)
a
1.03(1.01,1-06)
1.05(1.01,1.09)
1.02 (0.99,1.04)
1.03 (1.01,1.05)
1.02 (1.00,1.05)
1.04 (1.00,1.08)
1.03(1.01,1.05)
1.03(1.01,1.05)
1.03 (1.01,1.06)
OR
20
a
1.9
2.5
1.4
1.8
1.6
2.0
1.8
1.7
1.9
a
Odds ratio as determined by the logistic models given in Table III for the
increase in exposure of 1, and 20 mg/m -years.
From reference 95. Note the progressively higher risks of respiratory symp-
toms or a lowered FEV1 with a twenty-fold increase in cumulative coal dust
exposure. Where exposure data is reasonably precise, such calculations provide
an excellent guide to the maximal exposure that should be permitted.
increase the public acceptability of a risk, such as when it is voluntar-
ily undertaken, or familiar, and characteristics that tend to decrease
the acceptability of risk, such as when it is presented by an untrusted
source or is undetectable by the individual, provide a useful summary
of much previous investigation. Such characteristics strike a chord in
most thinking people.
What is often not emphasized is that all of us are used to making
personal decisions that are based on our perception of risk, but are
uninfluenced by statistical considerations. I may decide, at the age of
sixty, that it is no longer "safe" for me to use a chain saw when up a tree.
I base this on my perceptions of my own strength and reaction time; I
may make such a reasonable judgment in the absence of any knowl-
edge of the statistics of injuries to people more than sixty years old
under these conditions. In a perhaps more familiar example, we may
go out to dinner in the winter and visit a couple who live ten miles up a
remote and difficult road. While we are having dinner, it begins to
snow hard, and when it is time to leave, six inches have fallen. Our host
suggests that we stay the night, and return in the morning. An immedi-
ate "yes" or "no" decision has to be made. In milliseconds, we assess
the state of our tires, our knowledge of the road, our commitments
early tomorrow, and even our estimate of the state of our blood alcohol,
and either accept his invitation or decide to make the journey.
MANDATED SCIENCE: MAJOR ISSUES 73
Characteristics That Tend to Increase
Acceptability of a Risk
Voluntary
Familiar
Immediate impact
Detectable by individual
Controllable by individual
Fair
Noncatastrophic
Well understood
Natural
Trusted source
Visible benefits
Characteristics That Tend to De-
crease Acceptability of a Risk
Involuntary
Unfamiliar
Remote impact
Nondetectable by individual
Uncontrollable by individual
Unfair
Catastrophic
Poorly understood
Artificial
Untrusted source
No visible benefits
Source: Slovic et al. (1979); Plough and Krimsky (1987).
From reference (97). Although based on a considerable volume of carefully collected data, these
categories generally correspond with our intuitive feelings.
We are therefore used to having to make personal choices based on
the best information available to us, but uninformed by statistical
data. What purpose do the statistical estimates of risk shown in table
2.8 serve? The answer is that they provide a necessary basis for public
policy decisions, such as whether and if asbestos should be removed;
whether present coal dust standards are adequately protective; wheth-
er automobile lead represents an unacceptable hazard; and so on.
There can be no doubt that such estimates are essential prior to policy
determination, but they have always to be interpreted with knowl-
edge of their limitations.
In instances in which the public expresses a greater fear than is
justified on the basis of comparative risk estimates (as in the case of
nuclear power, for example), can it be assumed that all that is required
is more education, and people will then change their opinion on
policy? On the basis of the public response to the Chernobyl disaster,
this seems to me very unlikely. Indeed, when that event occurred,
many members of the public took the view that this catastrophic event
confirmed their "hunch"—as against the formal risk estimates of the
experts—that the technology was too potentially dangerous to be
widely deployed. In this case, therefore, risk communication is very
likely to be viewed by some members of the public as biased "educa-
Table 3.4. Psychosocial and Cultural Characteristics Affecting the Perception
of Risk
74 MANDATED SCIENCE: MAJOR ISSUES
tion" by a highly partisan group of interests, and this is unquestion-
ably what it can become. At its best however, risk communication can
be a consensus-building and consultative exercise, and can play an
important part in decision making as a result.
There is no doubt that objectivity, as represented by formal risk
assessment, will remain an important component of decision making.
Where an environmental impact assessment is to be conducted (on
hazards from a new incinerator for example), certain prior conditions
should be met. The first is that the objective risk assessment should be
prepared by individuals who have no financial stake in the project;
the second is that this risk assessment should be critiqued in public by
a second independent group; and the third is that this process should
be in the public domain, with those members of the public who are
likely to be impacted able to ask questions of both the proposer and
the risk assessors.
Economic Aspects of Environmental Policy Decisions
A few weeks after the London air pollution episode in December
1952, the government expressed concern that it would be far too
expensive to take the steps that might be required to prevent a repeti-
tion of the incident. However, in spite of this, for the next fifteen
years or so, it was common to hear the attitude expressed that, if
public damage could be shown to be occurring, then it was incumbent
upon government to take the requisite action to put a stop to it. I can
recall assuming that this would necessarily be the position of every
right-minded citizen.
It later became popular to argue that, before any preventive action
should be taken, it had to be proved that the benefit from the pro-
posed action outweighed the costs of taking it. However, a special
report to Congress in 1989 (98) examined the health benefits from the
reduction of air pollution and wrote: "We conclude that though there
is no doubt that significant health benefits result from controlling
some air pollutants, it is not currently feasible to produce an unam-
biguous evaluation of the health benefits."
Nevertheless, detailed attempts have been made to compute some
benefits of reducing air pollution. Thus, the United States Environ-
mental Protection Agency argued that the reduction in male blood
pressure that would follow the removal of lead from gasoline more
MANDATED SCIENCE: MAJOR ISSUES 75
than compensated for the costs of taking the action. They did not
place a specific dollar amount on a unit of IQ in a child (which, in my
opinion, would hardly have been ethically possible), but they did
calculate the costs of the supplemental education required for im-
paired children. In 1985, the EPA issued a paper on the health benefits
calculable from removing lead from gasoline (99). For the year 1988, it
was computed that, from the point of view of children's health, the
benefits would be $502 million; by reduction of blood pressure in
adults, the benefits would total $5 billion; and the total monetized
benefits amounted to $6.2 billion. By contrast, the costs to the refining
industry were calculated to be only $532 million. Thus, the EPA ar-
gued that the economic benefit of the reduction in male blood pres-
sure that would follow the removal of lead from gasoline more than
compensated for the costs of taking the action.
In relation to photochemical air pollution, Hall and her colleagues
in Los Angeles (100) have recently published a very detailed estimate
of health costs attributable to air pollution in that city. This required
placing dollar costs on symptoms such as cough and chest tightness
consequent upon photochemical air pollution. Even without taking
any account of possible long-term effects, they computed the health
benefits of attaining the United States ozone standard as between $1.2
and $5.8 billion annually, with a best estimate of $2.7 billion. For the
particulate PM10 (fine particles less than 10 microns in size) standard,
since they computed mortality costs in relation to this, the best esti-
mate total was $6.4 billion. Another economist has recently argued
that ozone controls are currently too expensive for the United States
to contemplate (101).
Such exercises have the merit of indicating the extent of air pollu-
tion problems by putting a dollar figure on public discomfort or actual
morbidity. In the case of asbestos, no attempt has been made to put a
dollar figure on the health effects of environmental asbestos. This
would require some basic assumptions, such as attributing all of the
30 percent of mesothelioma cases without prior occupational expo-
sure to environmental asbestos inhalation, and this would immedi-
ately be challenged. On the basis of our present knowledge, the mat-
ter could evidently not be settled.
The report to Congress quoted above (98) indicated that there was
a third perspective on economic costs in relation to pollution. In this,
benefits in terms of health effects avoided are compared, so that the
76 MANDATED SCIENCE: MAJOR ISSUES
best decisions can be taken about which pollutants require most
emphasis.
It is not clear whether strict economic analyses can, or for that
matter should, provide the basis for environmental decision making. I
have always thought that, once it has been agreed that raised blood
lead levels in children impair intellectual development, and that there
is no demonstrable threshold, it is mandatory that the exposure of
children to lead should be reduced as much as possible. I find it
immoral to try to put a dollar cost on the disbenefit of a lowered IQ in
a child. The only question, therefore, is whether the association dem-
onstrated between mental effects and blood lead represents a causal
relationship; once that judgment has been made, the need for the
removal of lead from gasoline should not have to be argued on other
grounds. For this reason, in 1972, I wrote that, on the basis of the
evidence then, lead had to be removed from gasoline (7).
Events in eastern Europe have shown that it is possible at the same
time to have both a grossly degraded environment and economic
collapse. It used to be argued that the choice was always between
pollution control and "jobs," representing a flourishing economy.
M.E. Porter of the Harvard University Business School has recently
called this philosophy into question by pointing to examples in which
the countries with the tightest pollution controls have the most effi-
cient industries (102). This may come about because an industry may
be forced to modernize to meet new pollution standards, and in the
process may develop a far more efficient manufacturing plant. Curtis
Moore (103) has also criticized the "jobs versus the environment"
argument, still believed by many in the business community.
These arguments will no doubt continue to evolve in free societies.
They will remain important in relation to policy, but the attitudes
taken may well differ between countries, and within individual coun-
tries, with different government leadership at different times.
The 1990 amendments to the Clean Air Act of 1970 in the United
States are 328 pages long, and their detail and complexity are far
greater than in any previous legislation. Curtis Moore, in an exhaus-
tive analysis of the amendments, has recently written (104):
Tension between protection of the environment and public
health on the one hand and considerations of cost and
convenience on the other are common in environmental
MANDATED SCIENCE: MAJOR ISSUES 77
decision making. The 1970 Clean Air Act Amendments re-
solved that tension in favor of protection of health and the
environment, forcing changes in technology and behavior
where necessary. The principal author of the 1970 amend-
ments, Sen. Edward S. Muskie, said as the Senate began
consideration of the proposals:
"The first responsibility of Congress is not the making of
technological or economic judgments—or even to be lim-
ited by what is or appears to be technologically or economi-
cally feasible. Our responsibility is to establish what the
public interest requires to protect the health of persons.
This may mean that people and industries will be asked to
do what seems to be impossible at the present time. But if
health is to be protected, these challenges must be met. I
am convinced they can be met."
In contrast, the 1990 amendments resolve that tension
between protection and costs and convenience in favor of
the latter, with the result that alleged limits of technology
constrain the level of protection. Thus the law as altered by
the 1990 amendments is a significantly different creature
from the 1970 Act. After amendment, protection of human
health has all but ceased to be the basis of the law. . . .
If this reading of the 1990 amendments is correct, it signals an
important (but little noted) change of principle behind environmental
legislation in the United States. That this reflects the attitude of the
current (1992) administration is supported by other recent actions,
such as the refusal of the EPA to review the ozone standard, and the
government's refusal to sign certain international agreements at the
recent United Nations conference in Rio de Janeiro.
4
A Survival Kit for the
Environmental Jungle
Following the examination of the five hazards, and the description of
some special problems associated with environmental health risks,
special attention can now be directed to the main players. These are
individual scientists, legislators, lawyers, industry, media, and the
public. It is clear that one of the main problems of public health
protection is in the accurate description of epidemiological data; since
this involves all the players, it will be discussed first.
Interpreting Epidemiological Data
Before a body of epidemiological data can be expressed in a general
summary, words have to be chosen to describe as exactly as possible
what the speaker judges to be the strength or weakness of the data.
In the sidebar, six different phrases are used to describe epidemio-
logical results. The first, that the data "unanimously indicate a
highly significant association" might apply to a review of studies of
cigarette smoking and lung cancer. It indicates that all studies show
a strong association. The second statement, "generally indicate a
significant association," would be suitable for certain air pollution
and general morbidity studies; it might also be used for the studies
of association between blood lead and neurobehavioral changes in
children. The last four statements (see box) are taken from different
comments on the health data relating to EMF, and suggest greater or
lesser degrees of skepticism.
Such statements deal only with the degree of confidence that the
78
A SURVIVAL KIT FOR THE ENVIRONMENTAL JUNGLE 79
Interpreting Epidemiological Data
Comparative strength of the words used:
The epidemiological data we have reviewed—
"unanimously indicate a highly significant association"
"generally indicate a significant association"
"indicate a weak association between ... "
"tend to suggest an association"
"while indicating a possible association, have demonstrated only a
weak relationship"
• "results . . . appear to provide some weak evidence in support of the
postulated association"
The first of these might apply to the relationship between cigarette smoking and
lung cancer. The last three are all taken from recent reports on the possible carcino-
genicity of electromagnetic fields.
speaker (or committee) may have in relation to the association. It is
important that any comment on policy or its implications be separated
from the statement on the force or weakness of the data. The follow-
ing sentence was quoted in Chapter 2 from the British Medical Re-
search Council Advisory Committee report on lead (50):
While the observed statistical associations detailed in
this review are consistent with the hypothesis that low-
level lead exposure has a small negative effect on the per-
formance of children in ability and attainment tests, the
limitations of epidemiological studies in drawing causal
inferences are such that it is not possible to conclude
that exposure to lead at current urban levels is definitely
harmful.
Such a paragraph requires detailed analysis. The first sentence con-
cedes that associations of lead and lowered IQ and performance have
been shown to occur, but by using the word "small" encourages the
view that it can be neglected. The second sentence suggests that these
results can be ignored on account of "the limitations of epidemio-
logical studies." One wonders which aspect of Hill's suggested causal
criteria is being referred to here. The end of the paragraph might
mean either that blood lead levels in children are not adversely af-
80 A SURVIVAL KIT FOR THE ENVIRONMENTAL JUNGLE
fected by the lead in automobile exhaust or that the studies show
effects only at levels above those that might be acquired in urban
living (i.e., that a threshold has been found). It is obvious that, with-
out much more detailed discussion, the suggested conclusion about
urban lead levels cannot be assumed to follow the earlier conclusion
drawn from the epidemiological studies that were reviewed. If the
paragraph was intended simply to reassure the inexpert reader, it
might succeed in its purpose. As a precise statement of the current
issues raised by lead exposure, it leaves a lot to be desired.
This example indicates the importance of very precise and exact
wording when describing the results of epidemiological studies and
the conclusions the writer is drawing from them. If it is suggested that
a conclusion has been reached as a result of reviewing such studies,
which has certain policy implications, these studies should be dis-
cussed separately. Furthermore, as I have noted earlier, the set of
values behind the final opinion must be explicitly articulated. The
reader should not be left to guess why he or she might come to a
different conclusion.
In table 3.1,1 suggested that certain biases existed in relation to the
interpretation of epidemiological data. One can make a case for con-
cluding that special committees in the United Kingdom, and possibly
also in Canada, are less likely to conclude that an association is causal
than are their counterparts in the United States. But if a committee
anywhere is large enough, on any controversial issue its members are
very likely to differ in opinion on the question of causation. Further-
more, the larger the representation of "basic" scientists, the greater
the tendency to dismiss associative data (this follows from table 3.2).
This is largely because scientific methodology, as in animal toxicology
studies, is designed to eliminate variables and produce direct evi-
dence of causation. Finding words to accommodate a range of opinion
within such a committee can be time consuming and difficult. Yet the
task is unquestionably important.
Mistakes Made by Scientists
Leaving aside all cases in which scientists are paid to express opinions
favorable to one view of epidemiological data, one can identify four
classes of mistakes commonly made (see sidebar). The last of these,
"failure to recognize that emotion is an appropriate force for policy
A SURVIVAL KIT FOR THE ENVIRONMENTAL JUNGLE 81
change," is very common. On an occasion many years ago, when I
was arguing for a policy based only on reason, Bruce Doern, then
director of the School of Public Administration at Carleton University,
pointed out to me that emotion was a proper force in relation to
policy, and this is unquestionably correct, though easy to forget in the
world of experimental data and risk analysis.
Salter (58), in discussing the process of inquiry into the toxicity of
the pesticide Captan, refers to evidence given by scientists and conclu-
sions reached. She writes:
The second statement, "We have no evidence that the
pesticide captan is dangerous to human populations," is
not hypothetical, but is taken from the discussions about
the final report of the expert committee on captan. In the
scientific discourse about captan, this statement is not very
interesting and in the context of the expert committee re-
port, this statement was accurate. Since only toxicological
data existed about captan at the time this statement was
made, and since toxicological experiments used animal
populations, there could have been no scientific evidence
that captan was actually dangerous to humans. All that
was available to the expert committee were extrapolations
from animal studies by scientists who differed in their inter-
pretation of the data."
As noted in Chapter 2, most scientists are now sensitive to the fine
line between an estimate of the force of scientific data on the one
hand, and an opinion based on a value judgment on the other. It is
important that, whenever public statements are being made, the lan-
guage accurately reflects this sensitivity. As I have pointed out else-
where (105), it is common that a defensible difference in opinion
between scientists—as to how certain data are to be interpreted—is
wrongly portrayed by the media as evidence that the process of scien-
tific inquiry is thereby discredited.
It is also important to distinguish between concluding that a stage
has been reached when steps should be taken to reduce the levels of
lead exposure in a community, and telling individual parents that
their child has been adversely affected by lead. This latter situation is
what the television crew would like, as the family can then be inter-
82 A SURVIVAL KIT FOR THE ENVIRONMENTAL JUNGLE
Mistakes Made by Scientists
Mistaken judgment on probability of causation
Mistaken assessment of reliability of the data (Note bias against credibility
of epidemiological data in most experimental scientists)
Failure to distinguish between scientific knowledge and policy judgment
Failure to recognize that emotion is an appropriate force for policy change
Examples of all of these are noted in the text.
viewed and the whole complex problem reduced to "one-on-one"
terms. It is perfectly proper to say publicly that such a stage has been
reached in terms of the community, but to refuse to identify individ-
ual cases that may have been affected by the risk. Environmental
scientists must clearly understand this differentiation.
Role of the Media
As indicated in table 2.1, in a free society the media play a dominant
role in public education on environmental questions. Historically, they
have been important in drawing attention to actual and potential envi-
ronmental health problems. Unexplained clusters of disease, apparent
increases in disease prevalence, summarizing official reports or legal
cases involving cigarette smoking or asbestos exposure—in all these
instances the media are responsible for informing the public and alert-
ing the legislator. It is clear that anyone involved in trying to under-
stand environmental health issues cannot avoid the responsibility of
responding to requests from the media for information or explanation.
However, the four negative points shown in the sidebar must be
fully understood by anyone who collaborates with the media. I have
rarely said anything publicly about an environmental issue without
the first question from reporters being: "Is air pollution a more impor-
tant problem than . . . ?" Arranging all problems in some order of
priority seems to be a preoccupation of some interviewers. Such ques-
tions should never be answered. If you are an asthmatic living in a
modern city, air pollution may be more important to you than the
A SURVIVAL KIT FOR THE ENVIRONMENTAL JUNGLE 83
Role of the Media
Dominant role in relation to public perceptions: (Incoming medical stu-
dents identify the media as the main source of their information on envi-
ronmental health)
Generally reflect and amplify all signals
Always oversimplify complex issues
Often stress false "priority" criteria
These factors, especially the last three, must be well understood (and expected)
by scientists who try to provide balanced input to the media.
effect of lead on the intelligence of children; if you live close to a lead
smelter, questions of the effect of EMF might well be secondary. There
is, therefore, no generic order of priority applicable to everyone to
which all questions conform.
Media personnel have great difficulty in understanding the limita-
tions imposed on anyone who discusses multifactorial disease. Also,
they do not appreciate that answering the question "How many peo-
ple in Vancouver take The New Yorker?" is different from the answer to
"How many people in Vancouver are affected by air pollution?" Also,
in their minds all percentages have to add up to a hundred, which, as
noted earlier, is not true of potential causative factors in multifactorial
disease. Nevertheless, it is the important and essential task of the
media in a free society to assist the public, and through them the
legislators, to determine that the weight of evidence has shifted the
burden of proof in relation to a specific question, such as that of the
relationship between cigarettes and lung cancer.
Role of the Corporate Lawyer
The lawyer commissioned to make the strongest case on behalf of his or
her client cannot be expected to emphasize aspects of the evidence that
point in a different direction. In the sidebar I have listed four aspects of
a legal defense that should be anticipated. Attempts have been made
(most notably by Milton Wessel [106]) to suggest that, in some circum-
stances, ethical considerations should override these attitudes, but
84
Influence of the Corporate Lawyer
Deny any responsibility
Point out that the epidemiological data do not constitute proof
Argue that confounders have influenced the epidemiological association
Suggest the presence of unidentified confounders
Argue that there is no biological plausibility for the alleged effect
NOTE: Will often use any method (short of force) to discredit the evidence
of opposing expert witnesses
Reference 106 should be consulted for a scholarly discussion of some of these
factors. These points are based on my own experience, however.
there is little evidence that this advice has been heeded. The scientist
who strays into this minefield must be aware that attempts to throw
doubt on his or her credibility are to be expected. It is apparently
inherent in all legal training that any points you can score against the
character of the witness should be scored, even if these have nothing to
do with the question in hand. More subtle is the attempt to get an
expert scientific witness to go beyond the strict questions on which he
or she is expert, or to get him or her to represent as a "fact" something
that is really only an opinion. When the financial stakes are enormous,
as is the case with asbestos, all of these devices will be used without
scruple. The relationship between "science" and the "legal process"
has been viewed as a shotgun marriage, and there is no doubt that
there are aspects of each that contravene the standards of the other.
The environmental scientist requires considerable judgment in how
answers are phrased. It is all too easy to be pushed into a statement that
is too dogmatic for the weight of evidence, or to be so cautious that a
judge is entitled to the conclusion that the court is unlikely to learn
anything of value from the witness.
The role of industrial consortia (no doubt guided by their corporate
lawyers) has varied in relation to these five hazards. The tobacco
industry has fought efforts to limit or reduce cigarette smoking in
every free society, and has no conscience about promoting cigarette
consumption in the Third World. The asbestos industry's attitude to
A SURVIVAL KIT FOR THE ENVIRONMENTAL JUNGLE
A SURVIVAL KIT FOR THE ENVIRONMENTAL JUNGLE 85
its unquestionable responsibility is a matter of record (45). The re-
moval of lead from gasoline was, quite understandably, strongly op-
posed by the industry that made tetraethyl lead. Interested industries
have generally opposed stricter air pollution legislation. In the case of
electromagnetic fields, the industry has not engaged in "dirty tricks"
or in efforts to discredit epidemiologists. If stronger epidemiological
data are published, or if restrictive legislation is proposed, this atti-
tude might be expected to change.
Societies supported by the public stand on the other side of the
argument. They are subject to the temptation to exaggerate the data,
as this is required if fund raising from a concerned public is to con-
tinue. The scientist or physician who attempts some kind of balance
or objectivity in relation to these questions often finds him- or herself
in a "no-man's land" between these opposing forces, cowering in a
shell hole while the missiles pass overhead.
Phrases to Watch Out for
The four phrases listed in the sidebar occur repeatedly in hearings,
debates, and press releases on all environmental hazards in which the
data are either incomplete or contradictory.
If a counsel states, "There is no scientific proof that . . . ," he or
she should be immediately asked what, in his or her opinion, would
constitute scientific proof that the effect existed. Is a scientific consen-
sus meant? Is "proof" used in a mathematical sense? What proof
commonly exists for any biological phenomena? Does he or she con-
sider that it has been proven that cigarettes cause lung cancer?
Much the same line of questioning should follow a statement that
"there is no statistical proof that ... "
The phrase "No significant excess of ... " is commonly used to
describe an increase in an observed incidence over what would be
expected, when this finding is inconvenient to the speaker. The inci-
dence often requires much more careful definition than the speaker
gives. If the increased cases are considered to represent a random
cluster, then this should be precisely stated. If other clusters seem to
be occurring elsewhere in relation to the same exposure, more expla-
nation is required. It is easy to dismiss one cluster as representing "no
significant excess of . . . ," but harder to explain the occurrence of
similar clusters wherever the hazard can be identified.
86 A SURVIVAL KIT FOR THE ENVIRONMENTAL JUNGLE
Phrases to Watch Out for
"No scientific proof that ... "
"No statistical proof that ... "
"No significant excess of . . ."
"No objective evidence of . . ."
Whenever any of these phrases is used in a discussion of epidemiological data,
as noted in the text, some additional questions must immediately be put to the
speaker.
Public Health Decision Making in a Free Society
The example of the very serious adverse health effects of cigarette
smoking suggests that the process of public decision making in a free
society follows the following sequence of events:
Stage 1. Initial epidemiological evidence begins to suggest a strong
association between cigarettes and lung cancer. No one study consti-
tutes "proof." Prospective studies may be stronger than retrospective
ones.
Stage 2. A stage is reached when epidemiological studies in differ-
ent countries, conducted in different ways, all show a strong associa-
tion. It becomes difficult to think of a common confounder. As Brad-
ford Hill (85) remarked in this context:
But to explain the pronounced excess in cancer of the
lung in any other environmental terms requires some fea-
ture of life so intimately linked with cigarette smoking and
with the amount of smoking that such a feature should be
easily detectable. If we cannot detect it or reasonably infer a
specific one, then in such circumstances I think we are
reasonably entitled to reject the vague contention of the
armchair critic, "you can't prove it, there may be such a
feature." . . .
What I do not believe—and this has been suggested—is
that we can usefully lay down some hard-and-fast rules of
evidence that must be obeyed before we accept cause and
A SURVIVAL KIT FOR THE ENVIRONMENTAL JUNGLE 87
effect. None of my nine viewpoints can bring indisputable
evidence for or against the cause-and-effect hypothesis and
none can be required as a sine qua non. What they can do,
with greater or less strength, is to help us make up our
minds on the fundamental question—is there any other
way of explaining the set of facts before us, is there any
other answer equally, or more, likely than cause and effect?
Stage 3. When the body of data is large enough, the burden of
proof shifts from those who have to prove causality to those (from the
tobacco industry, in the case of cigarettes) who argue that the
epidemiological data do not permit a judgment of causality. Such
proponents should be required to offer an explanation of why so
much epidemiological evidence points in the same direction. This
process can be facilitated by a carefully structured public inquiry at
which questions of this kind can be put. As noted earlier in this
chapter, the role of the media is essential in establishing that this
point has been reached.
Stage 4. When this stage is reached (perhaps in 1963 in the case of
cigarettes), government or regulatory intervention is required.
The media are so central in these questions because it is only
through media presentations and discussion that a free society can
finally reach an "opinion" that causality is no longer in doubt. Until
this stage is reached, public opinion is unlikely to support action to
curb the risk (smoking level, in this case).
This scenario seems to have been relevant to cigarette smoking, but
it is not at all clear that it applies universally. Indeed, Rose (92) has
pointed out that the example of cigarettes, with its exceptionally
strong epidemiological background, may have misled us as to the way
in which we have to regard other environmental hazards. In the case
of air pollution, the dialogue does occur in the media episodically,
particularly in the worst affected regions, but the attainment of any-
thing approaching a national consensus would seem unlikely. In the
case of lead, government action was taken after representations by
scientists that attracted very little media or public attention.
The question of the hazard of environmental asbestos has been
almost entirely taken over by the courts. Discussion of the validity of
risk estimates, and of dust levels within buildings, occurs mainly in
relation to specific litigation. This phenomenon has made scientific
88 A SURVIVAL KIT FOR THE ENVIRONMENTAL JUNGLE
dialogue in the public arena, between those who hold opposite points
of view, almost impossible. The Health Effects Institute report (47)
was mandated by the U.S. Congress in an effort to provide a relatively
objective review of contemporary data. In the case of electromagnetic
fields, it has been possible to argue that the data do not yet provide a
sufficiently strong foundation for legislative or other action, though
this might change as a result of the new studies from Sweden (78, 79).
Summary
The sidebar summarizes the mistakes and misunderstandings that
can be easily identified from the five environmental hazards. A full
understanding of them and familiarity with them is, I believe, an
essential requirement in the equipment any environmental health pro-
fessional must have for survival in the environmental jungle.
Mistakes and Misunderstandings
1. Misunderstanding the interface between "science" and "values."
2. Underrating the importance to public health of "small" changes in
mean values when the population exposed is large.
3. Denying a causal relationship in presence of significant associations.
4. Not understanding the constraints on making a causal inference from
associations.
5. Not understanding that poor exposure data and other factors may
weaken an association.
6. Misunderstanding the nature of scientific "proof."
7. Not understanding the complexity of multifactorial disease.
5
Conclusions
A contemporary visitor to the Western industrial democracies from
the Third World or from eastern Europe might well conclude that the
democratic process has shown itself over the past forty years to be
capable of containing many risks to public health. As a result of gov-
ernment initiative, publicity, and restrictions on advertising, general
cigarette consumption has fallen. Air pollution from coal burning has
been greatly reduced, and although automobile-generated pollution
has yet to be controlled, it has not been shown to have severe and
direct effects on health, though the ozone generated from these emis-
sions probably is a hazard. Indeed, reviewing air pollution, both Dr.
Merrill Eisenbud (107) in the United States and Sir Richard Doll (38) in
Britain have concluded that serious adverse health effects are no
longer being caused by air pollution. Legislation has forbidden the
use of asbestos in buildings, and its handling and removal is subject
to strict supervision. The general use of asbestos has declined dramati-
cally. Lead has been removed from gasoline, and is no longer a seri-
ous air pollutant in the Western world. Exposure to electromagnetic
fields has not yet been conclusively shown to be an important factor
in disease (though the accumulation of such evidence is disquieting).
The visitor might therefore review all these potential health risks
and conclude that the democratic process has shown itself sensitive
and responsive to each of these issues as they have arisen.
There is, however, another side to this issue. There is a difference
between dramatic events, like the excess mortality in the 1952 episode
in London, and the aggravation of a common condition such as
89
90 CONCLUSIONS
asthma. Also, it is still too early to expect that the full consequences of
oxidant air pollution, when a huge population is exposed, should be
understood. Cigarette smoking is still common in some young people,
especially in young women, in some countries. Asbestos exposure, as
judged by the load of asbestos in the lungs of city dwellers, is still
occurring; and the consequences of this are still controversial. It is still
too early to know that legislative action is required to limit exposure to
electromagnetic fields. The removal of lead from gasoline stands out,
perhaps, as an unqualified success, but in this case it is unclear how
powerful an influence the health risks were in that decision.
It would be difficult, however, to convince a perceptive visitor that
the process of decision making had been orderly, rational, and consecu-
tive. Much seems to have depended either on unpredictable and epi-
sodic media interest, or on dedicated efforts by individual scientists
and legislators. Yet, on reflection, perhaps one should expect that in a
democracy, however flawed, much of the decision making would nec-
essarily be haphazard and appear to be disorganized. In the protection
of public health from these hazards, it is the freedom of expression
within the society, together with access to information, that is crucial.
We can distinguish four distinct stages in relation to these five
hazards:
Stage 1. First epidemiological study; disbelief.
Stage 2. More epidemiological data confirm original study; search
for biological mechanisms begins. "Causality not proven until mecha-
nism is understood."
Stage 3. Significant association confirmed in all studies, burden of
proof shifts to those who deny causal relationship.
Stage 4. Government action and legislation.
Of the five examples I have chosen to study, it will be evident that
cigarettes, asbestos, and some aspects of air pollution have reached
stage 4; hazards from EMF are between stages 1 and 2; and lead has
reached stage 3 in some countries. This panorama of five prominent
public health concerns over the past forty years leads to some conclu-
sions, as follows.
Central Role of Environmental Epidemiology
It is apparent that epidemiological studies have been central to the
detection and evaluation of all these hazards. The discipline has
CONCLUSIONS 91
grown (but not kept pace with) the much more general dispersion of
the hazards than formerly, and with the growing realization that, the
control of bacterial disease in large part attained, environmental fac-
tors underlie much human disease. Yet the fraction of research expen-
diture currently dedicated to epidemiological research is a very small
component of the total medical research effort. The scientific commu-
nity currently gives far greater emphasis to research on mechanisms
of disease than to uncovering the community impact. Indeed, epide-
miologists have something of the same relationship to basic medical
scientists as psychiatrists have to surgeons. One hears people speak
as if the basic "risk" was uncovered by animal experiments first, and
then the epidemiological studies became necessary to see if the same
effect was occurring in humans. In all the examples I am discussing,
the epidemiological evidence preceded animal studies. Reviewing
this question, Omenn (108) recently wrote:
In conclusion, good public policy making about health
risks depends on epidemiologic studies, probably more
than most epidemiologists realize. Epidemiologists and toxi-
cologists need to work at finding more common ground and
more paths for integrating information from each other's
field. Epidemiologists need to develop guidelines for under-
taking studies or declining to do so, and epidemiologists
and risk analysts need to seek convergence on terminology
and descriptive language for communicating results from
epidemiologic studies.
It seems proper therefore to urge that much more potential should
be devoted to training and project funding in environmental epidemiol-
ogy. Science awards no Nobel prizes to epidemiologists, perhaps con-
sidering that the scientific component in epidemiology is too limited.
Role of Industry and Government
It is clear that, if public health protection were to be solely dependent
on dialogue between industry and government (particularly if con-
ducted in private), the health of the public would not be protected.
The lesson of eastern Europe is that government and government-
controlled industry in a closed society can very easily result in a
92 CONCLUSIONS
grossly degraded environment. In a free society, private collusion
between industry and government (if that were the sole forum for
discussion) would be likely to have the same effect.
Effective public health protection in a free society ultimately de-
pends on legislative action. In the case of cigarettes, government inter-
vention has shown itself to be an effective (though perhaps rather
slow) mechanism for public education and persuasion to change hab-
its and lifestyles. The U.S. Surgeon General's initiative in 1963 marked
the start of a slow rollback in cigarette consumption. Without that
initiative, it seems unlikely that this decline would have occurred.
Similar initiatives in relation to air pollution, lead in gasoline, and the
use of asbestos were central to a reduction in risks from these sources.
Much attention has recently been directed at ethical questions in-
volving individual scientists or groups of scientists or institutions in
which they work. This is perfectly proper. However, little attention
has so far been directed—and perhaps it should be—at ethical con-
straints that should be recognized as applying to industrial consortia.
Role of Legislators (Individual and Collective)
In all jurisdictions, and in all instances, individual legislators can be
identified who have spearheaded the introduction of legislation to
control these hazards. Their role is therefore central. It follows that
individual scientists have the responsibility of providing information,
both when requested and on their own initiative, to legislators with
an interest in the topics under discussion. In a free society, this is a
potent resource, and one likely to result in useful initiatives. In my
experience, it seems to work as well in the parliamentary system as
under the United States Constitution.
Role of Public Societies and Advocacy Groups
In some instances, these have been an important component in
decision-making. But the example of lead shows that effective action
can be initiated and carried into legislation without a strong public
advocacy group. If legislators are influenced by testimony from repu-
table individual and independent scientists, they can come to believe
in the importance of the question in spite of strong industrial advo-
cacy in the other direction. Where public societies exist, individual
CONCLUSIONS 93
scientists have the responsibility of educating them on the problems
and of ensuring that indefensible exaggeration does not occur. This is
the temptation to which public advocacy groups very easily succumb.
The need to maintain public subscriptions no doubt exerts pressure
on such groups to push the interpretation of data beyond reasonable
limits.
In spite of this limitation, one may fairly conclude that, in a free
society, the existence of an informed public opinion is very important
for the protection of public health. The pressures on government
exerted by powerful industry are so great that, unless there is some
countervailing force, the needs of public health are very unlikely to
receive the priority that the public expects they should be accorded. In
this context, the shift in emphasis detected by Curtis Moore (104)
between the philosophy behind the U.S. Clean Air Act of 1970 and the
1990 amendments suggests that this countervailing force is more nec-
essary now than it was twenty years ago.
Role of the Media in a Free Society
It is apparent that the media play an essential role in drawing attention
to health hazards and in urging government to take some action. In
spite of all the limitations in the perceptions of the media, environmen-
tal scientists and public health officials in a free society must be recon-
ciled to the fact that they must work through and with the media. They
should, I believe, undergo some instruction and guidance in relation to
their future contacts with the media, and, as I have indicated, be pre-
pared in advance for some situations. University departments that
train students in environmental health issues should recognize their
responsibility in relation to this question.
Role of the Independent Scientist
The central role of the independent scientist (who does not earn his or
her living by representing a vested interest) is self-evident. This is
why universities must be independent of the outside influence of
powerful industrial consortia (acting through industrialists on boards
of governors). Often, the scientists will, as I have noted earlier, feel as
if they are in a "no-man's land" between warring factions, each pos-
sessed of heavy artillery. Nevertheless, the scientist must understand
94 CONCLUSIONS
the difference between scientific evaluation and value judgment, and
of the probability that a loose word will be taken out of context and
"used against him" by one side or the other. We could probably do a
more thorough job of educating young epidemiologists on these ques-
tions. Salter's (58) recent observations on scientists and advocacy
would be a good place to start.
The environmental scientist may occasionally be confronted by diffi-
cult ethical questions. Merrill Eisenbud, from whose autobiography
(107) I learned a great deal, gives an interesting example of this. When
radioactive contamination of the Japanese fishing boat the Lucky Dragon
(Fukuryu Maru) occurred after the BRAVO nuclear test in the Pacific in
March 1954, he was sent at short notice by the United States govern-
ment to Japan to assess the issue and consult with Japanese authorities.
He was in the process of doing this when, back in the United States,
President Eisenhower, accompanied by Admiral Strauss, chair of the
Atomic Energy Commission, gave a televised press conference at
which he said that the boat had been well within the danger area
(contrary to instructions), and that the skin lesions on the fishermen
"are believed to be due to the chemical activity of the converted mate-
rial in the coral, rather than to radioactivity." Eisenbud knew that nei-
ther of the president's statements was true. It is perhaps surprising that
Eisenbud does not discuss in detail the moral dilemma that he con-
fronted when this occurred. Few of us are placed in such a delicate
position, but we may have encountered instances in which organiza-
tions with which we were linked took actions or issued statements with
which we disagreed. It may be difficult to decide whether to remon-
strate in private or in public when this occurs.
It is unquestionable that there are some aspects of the legal process in
a free society that facilitate the often tortuous process of arriving at
some decision on difficult questions. It is a strength of the legal pro-
cess that eventually a decision is required. Furthermore, all the evi-
dence has to be on the table, and the proceedings, particularly of
cross-examination, have to be in the public domain.
As already noted, a disadvantage of the legal process is the com-
mon attempt to discredit a witness on grounds entirely different from
the validity of his or her testimony. In the case of asbestos litigation in
ROLE OF THE COURTS AND LEGAL PROCESS
CONCLUSIONS 95
the United States, so high have the financial stakes become that the
legal process has interfered with the normal components of scientific
debate. As the National Academy of Sciences has recently noted (61)
in the case of hazardous waste site litigation, adversarial counsels
may even finance quasi-epidemiological studies; they are, of course,
free not to reveal "inappropriate" conclusions. It is a current weak-
ness of the present system in the United States that, if in such cases an
out-of-court settlement is reached, the court may order that all papers
should be sealed, and hence data from epidemiological studies rele-
vant to public health may never be revealed.
It is a unique aspect of the United States system that a court may
have to decide whether an agency, such as the Occupational Safety
and Health Administration or the Environmental Protection Agency,
has or has not set an appropriate standard for worker or public protec-
tion. This has the effect of ensuring that the agency has examined all
of the scientific evidence, and that its assessment of risk is appropri-
ate. Without any possibility of court action (as in the parliamentary
system), the standard-setting process within an agency may be very
casual and incomplete, and may omit the involvement of those with a
legitimate interest in the outcome. This is discussed further in the
next paragraph.
The Process of Standard Setting
I have published some thoughts on the standard-setting process else-
where (109). The process of setting air pollution standards in the
United States involves the following steps.
1. Preparation of a criteria document. This reviews all available
scientific evidence on the matter, but does not recommend any par-
ticular number as a possible standard.
2. Approval of this document (as being comprehensive and bal-
anced) by the Clean Air Scientific Advisory Committee of the EPA.
This has a rotating membership and consists of scientists familiar with
the field but independent of the agency. Its proceedings are publicly
available documents, and often make interesting reading.
3. Review of the criteria document by a nontechnical committee,
which sends forward specific advice (as a staff paper) to the adminis-
trator of the EPA. This committee may draw attention to particular
aspects of the criteria document, or extend its range to highlight areas
96 CONCLUSIONS
of uncertainty or to include calculations of risk for sensitive subgroups
of the population.
4. The EPA administrator makes a decision; he or she usually
makes it clear to what extent the proposed standard is driven by
purely medical or scientific considerations, and to what extent it has
had to be modified in the light of economic or other factors. A court
challenge can be mounted (by the American Lung Association, or the
automobile or petroleum industries for example) on the basis that the
proposed standard is too lax or unnecessarily stringent.
It is in the preparation of the criteria document that problems in
interpretation of data are thrashed out. Having commissioned the
writing of the basic document, EPA brings together the authors, offi-
cials from the EPA office responsible for criteria documents, technical
representatives of industry and environmental groups, and others. A
line-by-line analysis of the document occurs. In this process, there
will be detailed discussion of statistical aspects of studies, or of their
design. It is because this process is so thorough (though expensive)
and because all of the proceedings are in the public domain that the
criteria documents are uniquely useful. Cheaper alternatives in stan-
dard setting include the convening of bureaucrats in closed session to
advise on standards (as in Canada); the convening of a small group of
"experts" who have forty-eight hours to agree on a figure, and then
buttress their opinion with a few selected references (the procedure
apparently followed by the World Health Organization); or some
modification of these approaches. In the case of cancer-causing sub-
stances, a very rigorous procedure exists in Germany and has for a
hundred years. A permanent board makes recommendations, but
most of its members are designated by outside scientific and academic
organizations; that is, the government does not control the member-
ship. This system works well, and provides some of the advantages of
other methods.
In table 2.5, the progressive tightening of the asbestos standard
over a forty-year period is shown. This trend is not unique to asbes-
tos, and has also occurred in the case of lead. Bailar (110) commented
in 1989:
The last point . . . which I call the creeping center of the
information base, refers to the fact that increasing informa-
tion about some hazard is much more likely to show that it is
CONCLUSIONS 97
bigger than we thought than to show it is smaller. . . . This
is illustrated by some unpublished work I did in 1982 with
my colleagues Judy Jackson and Emmet Keeler. We tracked
the individual exposure recommendations published each
year by the American Council of Governmental Industrial
Hygienists (ACGIH), a private non-profit group of experts
from government, academic, and research institutions, and
private industry. These standards are sometimes changed,
and reasons for the changes are usually given. We found
that the ACGIH had published standards for 554 chemicals.
Most of these were recently published, with little time for
possible revisions, but 175 standards had been changed a
total of 222 times. Of these, 185 changes, or 83%, were in the
direction of reducing allowable exposures. Of these 222
changes, 114 were on the basis of new evidence (33 animal
studies, 81 human studies), 46 were on the basis of new
interpretations of old evidence, and 35 seemed to be from
concern that margins of safety were too small; we could not
classify 27 changes. The conclusion seems inescapable that,
at least in the experience in the United States, exposure
limits tend to become tighter with time, and these changes
are driven largely by increases in scientific information. It
would seem good public policy to recognize this tendency
and to set initial exposure limits somewhat lower than pres-
ent evidence justifies.
Unfortunately, it seems unlikely that this good advice will be heeded.
It may be objected that few standards have ever been relaxed because
it would not be worth the effort to go through a process to change
them, even though new scientific evidence may show that the risk is
less than previously indicated.
I conclude from this brief consideration of the role of the standard-
setting process that it can provide and has provided a uniquely useful
mechanism for considering the strength of all the evidence that bears
on the process of protecting the public health. The discussion of the
detailed data—whether they are epidemiological or experimental, in a
nonjudicial but public forum in which the temptation to score points
by character denigration is absent—can lead to a very productive
dialogue. The need to try to reconcile conflicting data, if such exist,
98 CONCLUSIONS
and to concentrate attention on the final object of the discussion—
namely, to be able to reach consensus on some definitive number as a
minimal effects level for example—together have the effect of focus-
ing the discussion in a manner that would not otherwise exist.
Such deliberations should occur within a diverse group. If a spe-
cific group of workers are at special risk, as they may be from such
hazards as asbestos or radon, their representatives should participate
in the discussion (111). The placing of the proceedings in the public
domain acts to prevent too partisan a point of view.
This process of arriving at a consensus on a safe standard for public
exposure could be structured as a "consensus conference" of the type
visualized by Wessel (106). A government agency or ministry could
well finance a process of this kind, without necessarily being bound
by the outcome, and by so doing encourage a convergence of views. I
have observed that this kind of mechanism is only feared by those
who suspect (rightly) that they cannot control the outcome.
I have noted that the British have often in the past been reluctant to
agree to the propounding of any standard for maximal permitted
exposure to anything, usually arguing that the scientific data are not
strong enough for any conclusion to be reached. Other countries,
particularly the United States and some European countries, have
taken the view that a standard is always necessary if the public health
is to be protected. Canada has oscillated between these two view-
points, often being content to publish "guidelines"; these are rela-
tively weak because they do not require specific action when they are
transgressed.
The Role of Consensus Conferences
Wessel (106) argued the value of conferences designed to produce a
consensus on a question. Another mechanism is to establish an inter-
disciplinary committee, give it enough time to work, and require it to
produce a report that represents a consensus of members. This is
what the special advisory committee to the EPA's Science Advisory
Board (74) was required to do (with the additional requirement that its
deliberations should be in public). Governments can facilitate this
process when the need is perceived. In the United States system,
hearings before special committees of the Senate play a very impor-
tant role; unfortunately there is no effective and equivalent process in
CONCLUSIONS 99
the parliamentary system. However, the conference committees that
operate in the United States to negotiate a compromise between the
House of Representatives and the Senate do not do so in public. In the
parliamentary system, on some occasions, committees have been con-
stituted with members from the different parties in the House of
Commons. This was done in the case of proposed cigarette legisla-
tion, and in connection with the problem of acid rain. These commit-
tees may suffer from lack of a strong, permanent, and well-informed
staff, and usually either produce recommendations that are never
implemented (as was the case in Canada in relation to an inter-
parliamentary committee on cigarettes), or their proceedings degener-
ate into party political squabbling.
The production of "objective" reports steered by organizations that
are perceived to be independent of interested parties (such as the
Health Effects Institute [47]) provides another mechanism. However
the disadvantage here is that there is no opportunity for public input
and no dialogue, therefore, in relation to the most difficult questions.
There is, after all, a great deal to be said for the process of cross-
examination; particularly is this the case in clarifying the basis of
different opinions on the interpretation of epidemiological data. Only
in this way can the (unstated) value assumptions that necessarily
underlie such opinions be brought out into the open.
Differences Between the Parliamentary System and the
United States Constitution
Several major differences between political processes in the parliamen-
tary system, as seen in Britain and in Canada, and the United States
system affect decision-making processes on environmental questions.
In a parliamentary country, accountability is attained by such traditions
as "question time" in the House of Commons; the election process
provides an opportunity to defeat a government with unpopular poli-
cies. Actual power is exerted by the cabinet. All ministers are elected
politicians. Challenges to government action—such as that forbidding
the use of saccharin, the spraying of asbestos, or the use of lead in
gasoline—can only be mounted on the basis that a minister has ex-
ceeded the power vested in him or her. No legal challenge is possible
on the basis that the scientific data have been misread or disregarded.
Ministers in the parliamentary system may (and do) appoint advisory
100 CONCLUSIONS
committees to themselves that may report confidentially. Under the
Public Inquiry Act, the federal or a provincial government may appoint
a royal commission required to have public input and to report in
public. This inquiry system has been commonly used in Canada, and
the process has been the subject of considerable study (112). Apart
from such inquiries, the parliamentary system encourages a closed
decision-making process, but action to deal with a problem (such as a
hazardous waste site) can be very swift and unchallengeable.
In the United States, the decision that, if a citizen wished to chal-
lenge an action by a government agent, he should have recourse to a
court process, was taken by its first president, George Washington
(113). The question involved "an aggrieved owner of a carriage who
asserted an incorrect classification of his pleasure vehicle by the inter-
nal revenue agent." Carriages were apparently taxed on the basis of
their wheelbases. In deciding that a court should arbitrate between a
citizen and a government representative, Washington was no doubt
reacting to the arbitrary exercise of government power that character-
ized the British government of that era. His decision has had pro-
found repercussions. Today, for example, it means that the American
Petroleum Institute can mount a court challenge to a proposed air
pollution standard; because of this possibility, the process of deciding
on the standard has to be much more rigorous. This process of court
challenge lay dormant until about 1965, but it has been increasingly
used since then. The administrator in charge of the Environmental
Protection Agency (who is about to become a member of the cabinet)
is not an elected official but an appointed one approved by the U.S.
Senate. Decisions on environmental action are therefore one step re-
moved from the direct two-party system, though in effect no impor-
tant action can be taken without the concurrence of the White House.
The influence of the president, and particularly of his advisers, is
considerable in determining which agenda items are pushed and
which relegated to the back burner.
The strength of the United States system lies in the power vested in
committees of Congress. All the major discussion that preceded the
original Clean Air Act, and its more recent successor, took place in
public before a Senate committee. Senators are given sufficient re-
sources to employ specialized staff. In my experience, their knowl-
edge of the technicalities of air pollution and its control is prodigious,
and generally not matched by deputy ministers in the parliamentary
CONCLUSIONS 101
system. Political horse-trading in relation to decisions affecting public
health and the environment undoubtedly occurs and will presumably
occur in the future, but the final decisions are backed by formidable
technical expertise. An act passed in the United States in 1970 ex-
pressly forbade agencies to receive from advisory committees reports
that remained confidential, except in certain defined circumstances.
Those operating within the parliamentary system tend to ridicule
the dominant role of lawyers in the United States process, but con-
cede that the process is a much more open one; those in the United
States marvel at the closed decision-making process still customary in
the parliamentary system, but envy the ability of the system to take
swift and unchallenged action.
Offsetting Public Disenchantment with the Process of
Decision Making
Anyone who has listened to a large volume of public input on a
contentious question involving public health comes to realize that a
significant segment of society views the world only in terms of con-
spiracy theories, believing that industry, government, and individual
scientists constitute a formidable cabal seeking to deny the validity of
epidemiological data and to discredit anyone who expresses a con-
trary opinion. The recent volume on the adverse health effects of
electromagnetic fields (80) is written from this standpoint. An exten-
sion of such a view is to regard all information as subjective, to deny
that any objective data can exist, and hence to treat all phenomena as
if they were the result of "magic." Salter (58) notes the basis for the
contemporary disenchantment of society with scientists—they can,
after all, be bought like anything else.
It is important to stress a contrary view of the world. It is necessary
to assert that it is possible to establish whether lead in low doses
affects the development of children, and if so, how. It is possible to do
research to form a judgment on whether current levels of air pollution
are injurious. It is necessary to maintain on all occasions that we do
know something about how to establish objectively reliable informa-
tion; and that correct answers to questions of fact do actually exist.
Such an assertion will never convince a small minority of the public,
but that is what they are, a small minority.
102 CONCLUSIONS
Figure 5.1. Percentage of severe mental retardation among those exposed in
utero by dose and gestational age in Hiroshima and Nagasaki. Vertical lines
indicate 90 percent confidence intervals. Source: NRC (1990). From reference
114. These recently published data from the Hiroshima and Nagasaki exposed
population are a chilling reminder of the vulnerability of the developing brain.
It is remarkable that they have attracted (so far) very little media attention.
Respect for the Complexity of Living Processes
Albert Schweitzer taught "reverence for life" as an important philo-
sophical and ethical attitude. Reviewing the five hazards I have
chosen to discuss, I am made aware of two entirely disparate atti-
tudes. On the one hand, I see a mystical approach to life and living,
which, in extreme form, seeks to deny the validity of any scientific
evidence; and on the other, perhaps exemplified by engineers (48), I
observe a mechanistic (and simplistic) approach to biological out-
comes. This is another "no-man's land" in which the independent
environmental scientist may well find himself.
We can at the very least, I suggest, stress the complexity of living
processes. The public is surely right in believing that our entrepreneu-
rial ingenuity in introducing new chemicals or new processes into our
CONCLUSIONS 103
environment, and disseminating them widely, far outstrips our ability
to assess them. Furthermore, our understanding of such extraordi-
narily complex processes as the growth and development of the hu-
man brain is still so rudimentary that foresight can hardly be required
of us. A chilling reminder of the reality of this is shown in figure 5.1.1
began by recalling the 1945 bombing of Hiroshima and Nagasaki; this
figure shows data published only a year ago indicating that the in
utero radiation that resulted from those events led to a major (and
dose-response related) increase in severe mental retardation.
What is currently required of us is undoubtedly to stress the limita-
tions of our biological understanding, and to counter an attitude of
"environmental bravado" whenever we detect it, for it is surely based
on ignorance.
We should have learned enough over the past forty years to know
that such an attitude is a sure recipe for disaster.
This page intentionally left blank 
References
1. Corn, J. K., and Starr, J. 1987. Historical perspective on asbestos: policies
and protective measures in World War II shipbuilding. American Journal of
Industrial Medicine 11:359-373.
2. Lippmann, M. 1992. Ozone. In Environmental toxicants: human exposures
and their health effects, ed. Morton Lippmann. New York: Van Nostrand
Reinhold.
3. Thompson, J. 1991. East Europe's dark dawn. National Geographic 179:
36-39.
4. Leowski, J. 1986. Mortality from acute respiratory infections in children
under 5 years of age; global estimates. World Health Statistics Quarterly
39:138-144.
5. Pandey, M. R., Smith, K. R., Boleij, J. S. M., and Wafula, E. M. 1989.
Indoor air pollution in developing countries and acute respiratory infection
in children. Lancet (25 February):427-429.
6. Cohen, N. 1991. Regulation of in-place asbestos-containing material. Envi-
ronmental Research 55:97-106.
7. Bates, D. V. 1972. A citizen's guide to air pollution. Montreal and London:
McGill-Queen's University Press.
8. Rose, G. 1989. Science, ethics and public policy. In Assessment of health
hazards, ed. U. Mohr, pp. 349-356. Heidelberg and New York: Springer-
Verlag.
9. The effects of fog on cattle in London. The Veterinarian XLVII:533 (Fourth
Series no. 299), January 1874.
10. Seiberling, G. 1988. Monet in London. Seattle and London: University of
Washington Press.
11. H. M. Stationery Office. 1954. Mortality and morbidity during the London fog
of December 1952. Report no. 95 on Public Health and Medical Subjects.
London.
105
106 REFERENCES
12. Ashby, E., and Anderson, M. 1981. The politics of clean air. Oxford: Claren-
don Press.
13. Tarr, J. A. 1981. Changing fuel use behavior and energy transitions: the
Pittsburgh smoke control movement 1940-1950. Journal of Social History
14:561-588.
14. Royal College of Physicians. 1970. Air pollution and health: a report for the
Royal College of Physicians. London: Pitman Medical and Scientific Publish-
ing Co.
15. Haagen-Smit, A. J. 1952. Chemistry and physiology of Los Angeles
smog. Industrial and Engineering Chemistry 44:1342-1346.
16. Holland, W. W., and Reid, D. D. 1965. The urban factor in chronic bron-
chitis. Lancet 1:445-48.
17. Xu, X., Dockery, D. W., and Wang, L. 1991. Effects of air pollution on
adult pulmonary function. Archives of Environmental Health 46:198-206.
18. Bates, D. V. 1994. The effects of photochemical air pollution on people. In
Tropospheric ozone: human health and agricultural impacts, ed. David J. McKee.
Boca Raton: Lewis Publishers.
19. Koren, H. S., Devlin, R. B., Graham, D. E., Man, R., McGee, M. P.,
Horstman, D. H., Kozumbo, W. J., Becker, S., House, D. E., McDonnell,
W. E, and Bromberg, P. A. 1869. Ozone-induced inflammation in the
lower airways of human subjects. American Review of Respiratory Diseases
139:407-415.
20. Hazucha, M., Bates, D. V., and Bromberg, P. A. 1989. Mechanism of action
of ozone on the human lung. Journal of Applied Physiology 67:1535-1541.
21. Kreit, J. W., Gross, K. B., Moore, T. B., Lorenzen, T. J., D'Arcy, J., and
Eschenbacher, W. L. 1989. Ozone-induced changes in pulmonary function
and bronchial responsiveness in asthmatics. Journal of Applied Physiology
66:217-222.
22. Keeler, G. J., Spengler, J. D., Koutrakis, P., Allen, G. A., Raizenne, M.,
and Stern, B. 1990. Transported acid aerosols measured in southern On-
tario. Atmospheric Environment 24A:2935-2950.
23. Bates, D. V., and Sizto, R. 1987. Hospital admissions and air pollutants in
southern Ontario: the acid summer haze effect. Environmental Research
43:317-331.
24. Bates, D. V. 1985. The strength of the evidence relating air pollutants to
adverse health effects. Carolina Environmental Essay Series. Chapel Hill:
University of North Carolina Institute for Environmental Studies.
25. Thurston, G. D., Ito, K., Lippmann, M., and Bates, D. V. In press. Respi-
ratory hospital admissions and summertime haze air pollution in Toronto,
Ontario: consideration of the role of acid aerosols.
26. Schwartz, J., and Dockery, D. W. 1992. Increased mortality in Philadel-
phia associated with daily air pollution concentrations. American Review of
Respiratory Diseases 145:600-604.
27. Sebastian, I. 1990. Issues in urban air pollution: Ankara diagnostic report.
Environment Working Paper, no. 38. Washington, D.C.: World Bank.
REFERENCES 107
28. Warner, K. E. 1989. Effects of the antismoking campaign: an update.
American Journal of Public Health 79:144-151.
29. Ritchie, J. W., 1930. Human Physiology. Rev. W. J. Dobbie. Canadian
Health Series. Toronto: W. J. Gage & Co., Ltd.
30. Gould, S. J. 1991. The smoking gun of eugenics. Natural History 12:9-17.
31. U.S. National Academy of Sciences. 1986. Environmental tobacco smoke.
Washington, D.C.: National Academy Press.
32. U.S. Department of Health and Human Services. 1986. The health conse-
quences of involuntary smoking. Washington, D.C.
33. Samet, J.M. 1989. Environmental tobacco smoke: adverse effects on respi-
ratory infection, respiratory symptoms, and lung function. In Assessment of
inhalational hazards, ed. U. Mohr. Heidelberg and New York: Springer-
Verlag.
34. Lee, P. N., Fry, J. S., and Forey, B. A. 1990. Trends in lung cancer, chronic
obstructive lung disease, and emphysema death rates for England and
Wales 1941-1985 and their relation to trends in cigarette smoking. Thorax
45:657-665.
35. Snider, G. L. 1992. Emphysema: the first two centuries—and beyond: a
historical overview, with suggestions for future research; part I. American
Review of Respiratory Diseases 146:1334-1344.
36. Fielding, J. E. 1987. Smoking and women. New England Journal of Medicine
317:1343-1345.
37. Fielding, J. E. 1985. Smoking: health effects and control. New England
Journal of Medicine 313:491-498.
38. Doll, R. 1983. Prospects for prevention. British Medical Journal 286:445-453.
39. Warner, K. E. 1989. Effects of the antismoking campaign: an update.
American Journal of Public Health 79:144-151.
40. Pierce, J. P. 1989. International comparisons of trends in cigarette smok-
ing prevalence. American Journal of Public Health 79:152-157.
41. Pierce, J. P. 1991. Progress and problems in international public health
efforts to reduce tobacco usage. Annual Review of Public Health 12:383-400.
42. Mancuso, T. F. 1989. Letter to the editor: response to Dr. C-G Ohlson.
American Journal of Industrial Medicine 15:353-356.
43. Becklake, M. R. 1976. Asbestos-related diseases of the lung and other
organs: their epidemiology and implications for clinical practice. American
Review of Respiratory Diseases 114:187-227.
44. International Agency for Research on Cancer. 1973. Biological effects of
asbestos; proceedings of a working conference, October 2-6,1972. Lyon, France.
45. Brodeur, P. 1985. Outrageous misconduct: the asbestos industry on trial. New
York: Pantheon Books.
46. Lippmann, M. 1992. Asbestos and other mineral fibers. In Environmental
toxicants: human exposures and their health effects, ed. Morton Lippmann.
New York: Van Nostrand Reinhold.
47. Health Effects Institute. 1991. Asbestos in public and commercial buildings: a
literature review and synthesis of current knowledge. Cambridge, Mass.
108 REFERENCES
48. Bates, D. V. 1991. Editorial: asbestos—the turbulent interface between
science and policy. Canadian Medical Association Journal 144:554-556.
49. Henneberger, P. K., and Stanbury, M. J. 1992. Patterns of asbestosis in
New Jersey. American Journal of Industrial Medicine 21:689-697.
50. Burdorf, L., Swuste, P. H. J. J., and Heederik, D. 1991. A history of aware-
ness of asbestos disease and the control of occupational asbestos exposures
in the Netherlands. American Journal of Industrial Medicine 20:547-555.
51. Shepherd, K. E., Oliver, L. C, and Kazemi, H. 1989. Diffuse malignant
pleural mesothelioma in an urban hospital: clinical spectrum and trend in
incidence over time. American Journal of Industrial Medicine 16:373-383.
52. Huncharek, M. 1992. Changing risk groups for malignant mesothelioma.
Cancer 69:2704-2711.
53. Churg, A., Wright, J. L., Gilks, B., and Depaoli, L. 1989. Rapid short-term
clearance of chrysotile compared with amosite asbestos in the guinea pig.
American Review of Respiratory Diseases 139:885-890.
54. Du Toit, R. S. J. 1991. An estimate at which crocidolite asbestos fibres are
cleared from the lung. Annals of Occupational Hygiene 35:433-438.
55. Churg, A., and Wiggs, B. 1986. Fiber size and number in users of pro-
cessed chrysotile ore, chrysotile miners, and members of the general popu-
lation. American Journal of Industrial Medicine 9:143-152.
56. Churg, A. 1991. Editorial: analysis of lung asbestos content. British Journal
of Industrial Medicine 48:649-652.
57. Bates, D. V. 1987. Editorial: asbestos; promotion or prohibition? Canadian
Medical Association Journal 26:107-109.
58. Salter, L. 1988. Mandated science: science and scientists in the making of stan-
dards. Dordrecht, Boston, and London: Kluwer Academic Publishers.
59. Hertzman, C., Ward, H., Ames, N., Kelly, S., and Yates, C. 1991. Child-
hood lead exposure in Trail revisited. Canadian Journal of Public Health
82:385-391.
60. Mahaffey, K. E. R., McKinney, J., and Reigart, J. R. 1992. Lead and com-
pounds. In Environmental toxicants: human exposures and their health effects,
ed. Morton Lippmann. New York: Van Nostrand Reinhold.
61. U.S. National Academy of Science. 1992. Environmental epidemiology; Vol
2.: data gaps, resource needs and research opportunities. Washington, D.C.
62. Needleman, H. L., and Gatsonis, C. A. 1990. Low-level lead exposure
and the IQ of children: a meta-analysis of modern studies. Journal of the
American Medical Association 263:673-678.
63. Royal Society of Canada. Commission on Lead. 1986. Lead in the Canadian
environment: science and regulation; final report. Ottawa, Ont.: Royal Society
of Canada.
64. Peterson, E. B., Chan, Y-H., Peterson, N. M., Constable, G. A., Caton, R.
B., Davis, C. S., Wallace, R. R., and Yarrenton, G. A. 1987. Cumulative
effects assessment in Canada: an agenda for action and research. Ottawa, Ont.:
Minister of Supply and Services, Canadian Environmental Assessment
Research Council.
REFERENCES 109
65. Winneke, G., Brockhaus, A., Ewers, U., Kramer, U., and Neuf, M.
1990. Results from the European multicenter study on lead neurotoxicity
in children: implications for risk assessment. Neurotoxicology and Teratology
12:553-559.
66. Smith, M. A. 1988. The neuropsychological effects of lead in children: a review of
the research report from the Medical Research Council Advisory Group on Lead and
Neuropsychological Effects in Children. London: Hospital for Sick Children,
Institute of Child Health, Department of Child Psychiatry.
67. U.S. National Academy of Sciences. 1992. Environmental Neurotoxicology.
Washington, D.C.: National Academy Press.
68. Smith, M. A., Grant, L. D., andSors, A. I., eds. 1988. Lead exposure and child
development: an international assessment. Report published for the Commis-
sion of the European Communities and the U.S. Environmental Protection
Agency. Dordrecht, Boston, and London: Kluwer Academic Publishers.
69. Lippmann, M. 1990. Lead and human health: background and recent
findings. 1989 Alice Hamilton Lecture. Environmental Research 51:1-24.
70. World Health Organization. Regional Office for Europe. 1987. Air quality
guidelines for Europe. WHO Regional Publications, European Series, no. 23.
Geneva, Switzerland.
71. National Radiological Protection Board (U.K.). 1992. Electromagnetic fields
and the risk of cancer; report of an advisory group on nonionising radiation. Vol. 3,
no. 1. Chilton, England.
72. Wertheimer, N., and Leeper, E. 1979 Electrical wiring configurations and
childhood cancer. American Journal of Epidemiology 109:273-284.
73. U.S. Environmental Protection Agency. Office of Research and Develop-
ment. 1990. Evaluation of the potential carcinogenicity of electromagnetic fields;
review draft. Washington, D.C.
74. U.S. Environmental Protection Agency. Science Advisory Board. 1991.
Potential carcinogenicity of electromagnetic fields; report of the Nonionizing Elec-
tric and Magnetic Fields Subcommittee of the Radiation Advisory Committee of the
Science Advisory Board. Washington, D.C.
75. Theriault, G. 1992. Electromagnetic fields and cancer risks. Rev Epidem et
Sante Publ 40:S55-S62.
76. London, S. J., Thomas, D. C., Bowman, J. D., Sobel, E., Cheng, T-C., and
Peters, J. M. 1991. Exposure to residential electric and magnetic fields and
risk of childhood leukemia. American Journal of Epidemiology 134:923-937.
77. Hutchison, G. B. 1991. Carcinogenic effects of exposure to electric fields
and magnetic fields. Paper prepared for Workshop on Future Epidemio-
logic Studies of Health Effects of Electric Fields and Magnetic Fields, Car-
mel, California, Feb. 6-8, 1991.
78. Feychting, M., and Ahlboom A. 1992. Magnetic fields and cancer in people
residing near Swedish high voltage power lines. Stockholm: Karolinska Insti-
tutet, Institutet for Miljomedicin.
79. Floderus, B., Persson, T., et al. 1992. Occupational exposure to electromag-
netic fields in relation to leukemia and brain tumors; a case-control study. Solna,
110 REFERENCES
Sweden: National Institute of Occupational Health, Department of
Neuromedicine.
80. Brodeur, P. 1989. Currents of death: power lines, computer terminals, and the
attempt to cover up their threat to your health. New York: Simon and Schuster.
81. Graham, L. R. 1981. Between science and values. New York: Columbia Uni-
versity Press.
82. Steensberg, J. 1989. Environmental health decision-making: the politics of dis-
ease prevention. Copenhagen: Almquist and Wiksell International.
83. Stigler, S. M. 1986. The history of statistics: the measurement of uncertainty
before 1900. Cambridge, Mass.: Belknap Press of Harvard University Press.
84. Rothman, K. J., ed. 1988. Causal Inference. Chestnut Hill, Mass.: Epide-
miology Resources, Inc.
85. Hill, A. B. 1965. The environment and disease: association or causation?
Proceedings of the Royal Society of Medicine 58:295-300.
86. Bates, D. V. 1992. Health indices of the adverse effects of air pollution: the
question of coherence. Environmental Research 59:336-349.
87. Saracci, R. 1981. Personal-environmental interactions in occupational
epidemiology. In Recent advances in occupational health, ed. J. C. McDonald.
London: Churchill Livingstone.
88. Harrison, K., and Hoberg, G. In press. Risk, science, and politics: regulating
toxic substances in Canada and the United States.
89. Checkoway, H., and Hickey, J. L. S. 1988. Estimating the potential dis-
ease rate reduction under varying conditions of combined effects. Interna-
tional Journal of Epidemiology 17:666-672.
90. Omenn, G. S., Merchant, J., Boatman, E., Dement, J. M., Kuschner, M.,
Nicholson, W., Peto, J., and Rosenstock, L. 1986. Contribution of environ-
mental fibers to respiratory cancer. Environmental Health Perspectives 70:51-
56.
91. Gould, S. J. 1992. The confusion over evolution. New York Review of Books
39:47-54.
92. Rose, G. 1987. Environmental factors and disease: the manmade environ-
ment. British Medical Journal 294:963-965.
93. Corn, J. K. 1992. Response to occupational health hazards: a historical perspec-
tive. New York: Van Nostrand Reinhold.
94. Rosenthal, A., Gray, G. M., and Graham, J. D. 1992. Legislating accept-
able cancer risk from exposure to toxic chemicals. Ecology Law Quarterly
19:269-362.
95. Seixas, N. S., Robins, T. G., Attfield, M. S., and Moulton, L. H. 1992.
Exposure-response relationships for coal mine dust and obstructive lung
disease following enactment of the Federal Coal Mine Health and Safety
Act of 1969. American Journal of Industrial Medicine 21:715-734.
96. Sanden, A., Jarvholm, B., Larsson, S., and Thiringer, G. 1992. The risk of
lung cancer and mesothelioma after cessation of asbestos exposure: a pro-
spective cohort study of shipyard workers. European Respiratory Journal
5:281-285.
REFERENCES ill
97. Upton, A. C. 1992. Perspectives on individual and community risks. In
Environmental toxicants: human exposures and their health effects, ed. Morton
Lippmann. New York: Van Nostrand Reinhold.
98. Blodgett, J. 1989. Health benefits of air pollution control: a discussion. Congres-
sional Research Service Report for Congress. Washington, D.C.: U.S. Li-
brary of Congress.
99. U.S. Environmental Protection Agency. Economic Analysis Division.
1985. Costs and benefits of reducing lead in gasoline: final regulatory impact
analysis. Washington, D.C.
100. Hall, J. W., Winer, A. M., Kleinman, M. T., Lurmann, F. W., Brajer, V.,
and Colome, S. D. 1992. Valuing the health benefits of clean air. Science
255:812-817.
101. Russell, M. 1988. Ozone pollution: the hard choices. Science 241:1275-
1276.
102. Porter, M. E. 1991. Essay: America's green strategy. Scientific American
264:168.
103. Moore, C. A. 1992. Bush's nonsense on jobs and the environment. The
New York Times, September 25, 1992.
104. Moore, C. A. 1992. The 1990 Clean Air Act amendments: silk purse or
sow's ear? Duke Environmental Law and Policy Forum 2:26-58.
105. Science Council of Canada. 1982. Regulating the regulators: science, values
and decisions. Report no. 35. Ottawa, Ont.: Minister of Supply and Services.
106. Wessel, M. R. 1980. Science and conscience. New York: Columbia Univer-
sity Press.
107. Eisenbud, M. 1990. An environmental odyssey: people, pollution, and politics
in the life of a practical scientist. Seattle: University of Washington Press.
108. Omenn, G. S. In press. The role of environmental epidemiology in pub-
lic policy. Annals of Epidemiology.
109. Bates, D. V. 1988. Standard-setting as an integrative exercise: alchemy,
juggling, or science? In Inhalation toxicology, ed. U. Mohr. Heidelberg and
New York: Springer-Verlag.
110. Bailar, J. C. 1989. Inhalation hazards: the interpretation of epidemiologic
evidence. In Assessment of inhalation hazards, ed. U. Mohr. Heidelberg and
New York: Springer-Verlag.
111 . Science Council of Canada. 1977. Policies and poisons. Report no. 28.
Ottawa, Ont.
112. Salter, L., and Slaco, D. 1981. Public enquiries in Canada. Science Council
of Canada Background Study no. 47. Ottawa, Ont.: Minister of Supply and
Services.
113. White, L. D. 1948. The Federalists: a study in administrative history 1789-
1801. New York: Free Press.
114. U.S. National Research Council. 1990. Health effects of exposure to low levels
of ionizing radiation (BIER 5). Washington, D.C.: National Academy Press.
This page intentionally left blank 
Index
Access to information, 90
ACGIH, 28, 97
Acid rain, 21, 55, 99
Acid summer haze, 15
Air pollution, 6-22, 53, 65, 78, 87, 89, 92,
101; effect on FEV1, 13 (Fig. 2.2); and
hospital admissions, 16 (Table 2.2);
exposure to, 19; public awareness and
media coverage of, 20; industries and,
20
Alexander, Dr. Hope, 12
Alliance to End Childhood Lead Poison-
ing, 41
American Cancer Society, 22
American Lung Association, 21, 96
American Petroleum Institute, 20, 100
Ammonia, 15
Amosite asbestos, 33
Animal studies, 16
Arsenic, 65
Asbestos, 53, 56, 65, 69, 75, 89, 92, 98,
99; production, 27, 28 (Fig. 2.7);
epidemiology, 28; standards, 28, 29 (Ta-
ble 2.5), 96; public perception of risk,
29; media and, 29; in buildings, 30 (Ta-
ble 2.6); chrysotile, 31, 33; fiber clear-
ance, 33; risks, 33 (Table 2.7), 34 (Table
2.8), 35 (Fig. 2.9); exposure assess-
ment, 36, 90; crocidolite, 33; industry,
36, 85; export, 36; environmental expo-
sure, 55, 70; and smoking risks, 67 (Ta-
ble 3.2)
Asbestosis, 4, 31
Ashby, Lord, 12
Asthma, 13, 15, 62, 65, 90; effect of air
pollution, 17 (Fig. 2.3)
Asthmatics, 16, 82
Australia, 13
Automobile industry, 20, 96
Automobiles, 15, 43, 80, 89
Bailar, J. C., 96
Basic scientists, 80
Beaver Committee, 12
Becklake, M. R., 27
Beijing, 13
Biologic gradient, 63
Biological plausibility, 52, 63, 64
Biological understanding, 103
Birds: effect of radar, 50
Body weight, 65
Bravado, environmental, 103
Breast cancers in men, 48
Britain, 44, 64, 80; cigarette smoking, 23,
25 (Fig. 2.6); and standards, 98
British Columbia, 52
British Medical Research Council, 38
Brodeur, P., 28
Bronchitis, chronic, 13
Bulgaria, 38
Burden of proof, 83
Bureaucrats, 64; and standard-setting,
96
Buses, 15
113
114 INDEX
California, 46
Canada, 21, 44, 55, 64, 80, 96, 99,
100; ozone standard, 14; and stan-
dards, 98
Canadian Environmental Assessment Re-
search Council, 38
Captan, 81
Carcinogenic chemicals, 70
Catalytic converter, 43
Causality, 61-65, 66 (Table 3.1), 87; bi-
ases in making causal inference, 66
(Table 3.1)
Caution in science, 65
Charing Cross Bridge, 7
Checkoway, H., 68
Chernobyl, 73
Children: air pollution exposure, 20
China, 13, 26
Christie, R., 10
Chrysotile asbestos, 31, 33
Churg, A., 33
Cigarettes, cigarette smoking, 4, 22-27,
53, 54, 61, 65, 78, 83, 87, 90; in air
pollution studies, 13; consumption,
22; prohibition, 23; disease burden,
23, 24; reductions in, 24, 26 (Table
2.4); and lung cancer, 25 (Fig. 2.6); to-
bacco advertising, 25; environmental
tobacco smoke, 54; maternal, 60; and
asbestos, 66, 67 (Table 3.2); tobacco in-
dustry, 84; legislation, 99
Circadian rhythm, 50
Clean Air Act of 1970 (U.S.), 14, 21, 76,
93, 100; 1990 amendments, 76
Clean Air Scientific Advisory Committee
of EPA, 95
Coal burning, 13, 89
Coal dust, 71; standards, 73
Coal miners (U.S.), 71; risks, 72 (Table 3.3)
Coherence, 62
Confounders, 49, 60, 69
Congressional committees, 100
Consensus conferences, 98
Consistency, 62
Conspiracy theory, 64
Copenhagen, 22
Corn, J. K., 70
Coronary artery disease, 65
Corporate lawyer: role of, 83
Cost/benefits, 74
Courts: role of, 29, 87, 94-95
Coventry, 3
Criteria document (EPA), 95
Crocidolite asbestos, 33
Crops: effect of ozone, 13, 15
Cross-examination: value of, 99
Cross-sectional studies, 61
Decision-making, 90
Denmark, 38, 57
Denver, 44
Dietary fat, 65
Diisocyanate compounds, 66
Doern, B., 81
Doll, Sir Richard, 24, 51, 89
Donora, Pennsylvania, 10
Dose-response, 63, 70
Dresden, 3
East Ham, 4
Eastern Europe, 13, 76, 91
Economic aspects of environmental pol-
icy decisions, 74-77
Eisenbud, M., 89, 94
Eisenhower, President Dwight, 94
Elasticities, 60
Electric blankets, 52
Electric utilities, 20
Electrical workers, 46
Electromagnetic fields (EMF), 4, 44-53,
55, 61, 78, 83, 85, 89, 90, 101; wave-
lengths, 44, 45 (Fig. 2.15); leukemia
risks, 46 (Fig. 2.16), 49 (Table 2.9);
exposures, 51; public interest, 52;
epidemiology, 52; biological plausibil-
ity, 64
EMF. See Electromagnetic fields
Engineers, 102
Environmental bravado, 103
Environmental impact assessment, 74
Environmental Protection Agency. See
U.S. Environmental Protection Agency
Environmental tobacco smoke. See Ciga-
rettes, cigarette smoking
EPA. See U.S. Environmental Protection
Agency
Epidemiology, environmental: impor-
tance of data, 54; interpretation of
data, 54, 78; central role of, 90-91
Ethical issues and questions, 92, 94
INDEX 115
Family income, 61
FEV1 (forced expired volume in 1 sec-
ond), 13
Fish: effect of EMF, 50
Fisher, Sir Ronald, 23, 62
Forced expired volume in 1 second
(FEV1), 13
Freedom of expression, 90
Frith, T. J., 22
Gas cooking, 60
Genetic factors, 65
Germany: standard-setting, 96
Goteborg, 71
Gould, S. J., 68
Graham, J. D., 70
Graham, L., 57
Gray, G. M., 70
Greece, 38
Greenland, 4, 35, 37 (Fig. 2.10)
Hall, J. W., 75
Harrison, K., 67
Hazards: dissemination, 5; summary, 53-
56; stages, 90
Health and Welfare Canada, 68
Health Effects Institute, 31, 88, 99
Hickey, J. L. S., 68
Hill, Sir A. B., 61, 63, 86
Hiroshima, 3, 103
Hoberg, G., 67
Holland & Reid, 17
Hospital admissions, 60, 63
Houses of Parliament, 7, 10, 99
Housing quality, 60
Huncharek, M., 32
Hungary, 38
Hutchinson, G. B., 49
Hydrocarbons, 69
Hypertension, 65
Incinerators, 74
Independent
¥
 scientist: role of, 93
Industrial consortia, 56, 84
Industry, 56, 93
Information: access to, 90
International Agency for Research on
Cancer, 27
IQ: score, 39; dollar costs, 76; effect of
lead, 79
Italy, 38
Jackson, J., 97
Japan, 21, 94
Karachi, 44
Keeler, E., 97
Lanes, S., 62
Latency, 32, 71
Lawyers: proliferation of, 101
Lead, 36-44, 53, 55, 61, 73, 74, 79, 83, 87,
89, 90, 92, 99, 101; in gasoline, 4, 42
(Fig. 2.14), 85; behavior changes, 36 ff.;
lowered IQ, 37 ff.; in snow in Green-
land, 37 (Fig. 2.10); childhood behav-
ior, 38; pathways of exposure, 39, 41
(Fig. 2.13); mental development, 40
(Fig. 2.12); public concern, 41; indus-
try, 41; in blood, 42 (Fig. 2.14); biologi-
cal plausibility, 64; cost-benefit data,
76; ethical issues, 81; in gasoline, 85
League of Women Voters, 12
Leeper, E., 44
Legal process: role of, 84, 94-95
Legal profession, 29
Legislative action, 92
Legislators: role of, 92
Leukemia, 44, 55, 69; risks in workers, 46
(Fig. 2.16), 47 (Fig. 2.17); domestic risk,
49 (Table 2.9)
Life: respect for, 102
Litigation, 31, 56, 87
Living processes: complexity of, 102
Logistic regression, 60
London: air pollution, 3, 21; 1952 fog epi-
sode, 7-12, 11 (Fig. 2.1), 63, 74, 89
London, S., 46
Los Angeles, 13, 20, 46, 55, 75
Low level exposures, 34
Lucky Dragon (Fukuryu Maru), 94
Lung cancer, 23, 53, 65, 66, 68, 78, 83
MacLeod, I., 10
Macmillan, H., 10
Mandated risk levels, 70
Mandated science, 57
Manhattan, 27
Mao Zedong, 26
Media, role of, 54, 55, 64, 69, 81, 82-83,
93; and air pollution, 19, 20, 22; and
cigarette smoking, 26; and lead, 41
116 INDEX
Medical Research Council (MRC), Brit-
ish, 38, 50, 79
Medical scientists, 91
Melatonin, 48, 50
Mental retardation, 102 (Fig. 5.1), 103
Mesothelioma, 27, 31, 53, 68, 71, 75
Meuse Valley, 10
Mexico City, 15
Mining, 20
Modeling, 60
Monet, Claude, 7
Moore, C, 76, 93
Mortality, 63; air pollution effects, 15, 19
(Fig 2.5), 20 (Table 2.3)
Multifactorial disease, 65-70, 83
Muskie, Senator Edmund, 77
Nagasaki, 3, 103
National Academy of Sciences (U.S.), 23,
95
Needleman, H. L., 37
Netherlands, 31, 32
New York City: asbestos and lead in, 4
Nitrogen oxides, 13, 21
"No-man's land," 85, 93, 102
Nobel Prize, 91
Nonionizing Electric and Magnetic Fields
Subcommittee of the EPA Science Advi-
sory Board, 49
Objectivity, 74
Occupational Safety and Health Adminis-
tration (OSHA) (U.S.), 28, 95
Odds ratio, 46
Omenn, G., 91
Ontario: hospital admissions, 60, 61
Organic chemicals, 61
OSHA (U.S.), 28, 95
Oxidant pollution, 90
Ozone, 13, 14, 19, 21, 53, 69, 77; mode of
action, 15; biological plausibility, 64
Parliamentary system, 92, 95, 99
Particulate pollution, 13, 53; effects, 18
(Fig. 2.4); mortality, 19 (Fig. 2.5), 20 (Ta-
ble 2.3); biological plausibility, 64;
PM10, 75
Pearson, K.: correlation coefficient, 59
Pepys, Samuel, 64
Peters, J. M., 46
Petroleum industry, 96
Philadelphia, 15, 18 (Fig. 2.4), 19 (Fig.
2.5), 20 (Table 2.3)
Photochemical smog, 13, 14. See also
Ozone
Pierce, J. P., 27
Pineal gland, 50
Pittsburgh, 12
Plants, 13
PM10 (particulate), 75
Porter, M. E., 76
Priestman, Mr., 6
Psychiatrists, 91
Public concern, 54
Public disenchantment, 101
Public health decisions: stages, 86
Public Inquiry Act (Canada), 100
Public opinion, 21
Public policy, 62, 73
Public societies, 85, 92-93
Quasi-scientific meetings, 31
Radar, 50
Radiation exposure, 102 (Fig. 5.1)
Radon, 65, 68, 98
Regression coefficient, 60
Research expenditures, 91
Rio de Janeiro conference, 77
Risk, involuntary, 19
Risk acceptability, 72-73, 73 (Table 3.4)
Risk analysis, 91
Risk assessment, 70-74; quantitative, 70
Risk communication, 73
Risk estimates, 68, 70-74, 87; and asbes-
tos, 31, 35 (Table 3.4)
Risk perception, 70-74
Risk ratios, 46
Rose, G., 5, 69, 87
Rosenthal, A., 70
Rothman, K., 62
Rotterdam, 3
Royal College of Physicians, 12, 22
Royal Commission on the Poor Law, 59
Royal Society of Canada, 37, 61
Royal Society of Medicine, 61
Rumania, 38
Saccharin, 99
Salter, L., 36, 57, 58, 59, 81, 94, 101