Corporate Manipulation of Epidemiology

[Guest guest]

International Journal of Occupational and Environmental Health

Volume 13, Number 1

January - March 2007

118

In response to several articles on corporate corruption

of science that appeared earlier in this journal, a critic

outlined an epistemological model based on an unsupported

assertion that epidemiologic evidence is always

required to support cause–effect relationships. This

model, if adopted, would eliminate compensation to

victims of toxic exposures and impede regulation of

accepted hazards. Epidemiology is only one element in

support of cause–effect determinations. The critic's

proposal of an anti-health epidemiology was initially

developed by corporations with the goal of providing

defense in litigation, and is based not on science but on

a corporate need to enhance profits at the expense of

public health. Key words: epidemiology; industry influence;

corporate corruption of science; cause–effect

relationship.

INT J OCCUP ENVIRON HEALTH 2007;13:118–124

The October/December 2005 issue of this journal

included several articles on corporate corruption

of science.1–4 Recently, Otto Wong

responded to these articles' critiques of his work in a

seven page " editorial " published in Regulatory Toxicology

and Pharmacology, entitled, " The interpretation of occupational

epidemiologic data in regulation and litigation:

Studies of auto mechanics and petroleum workers. " *

In his editorial, Wong outlines an epistemological

model that, if adopted, would eliminate compensation

to victims of toxic exposures and impede regulation of

accepted hazards of all kinds, including drugs. He

bases his proposed method on the unsupported assertion

that epidemiologic evidence is always required to

establish cause–effect relationships. Epidemiology is

only a component of the science of cause-effect determination

and is supplemented by a variety of data sets,

including animal studies, toxicology studies, molecular

studies, case reports, studies of analogous substances,

pathologic examination of tissue specimens, and various

other studies.

While Wong claims to support " good epidemiological

practices, " these practices systematically underestimate

exposure risk. As discussed below, they were initially

proposed, developed, promoted, and funded by

tobacco companies and the Chemical Manufacturers

Association (now called the American Chemistry Council).

The specific studies of asbestos and benzene risks

that Wong cites were funded by various corporations

with the goal of providing a " scientific " defense in litigation

and a " scientific " rationale for eliminating or

limiting government regulations designed to protect

public health. Wong's proposal is an anti-health epidemiology,

based not on science but on a corporate

need to enhance profits at the expense of public

health, and this need is satisfied by limiting effective

regulation while leaving the injured uncompensated.

While epidemiology is often a component of the science

of cause–effect determination, it is never its sine

qua non. To determine general causation, researchers

evaluate a variety of data sets, including animal studies,

toxicology studies, molecular studies, case reports,

studies of analogous substances, and epidemiologic

case–control and cohort studies. Physicians use Koch's

postulates and Hill's considerations to extrapolate a

general cause–effect relationship from specific case

reports.5,6

Koch introduced his postulates in 1890 to establish

practices for determining microbial causes of disease,

including chronic and occupational (anthrax) infectious

diseases. Although limited in application, the postulates

clearly display the importance of animal models

in medical epistemology.5

Hill developed an expanded epistemological model

that demonstrated the tobacco–cancer relationship

through reliance on a number of scientific

approaches.6 Since then, Hill's " considerations " have

been used in numerous instances to establish cause–

effect relationships. However, it should be noted that

Hill's considerations do not require the use of or

Against Anti-health Epidemiology:

Corporate Obstruction of Public Health via

Manipulation of Epidemiology

DAVID EGILMAN MD, MPH, SAMANTHA HOWE

Dr. Egilman is Clinical Associate Professor of Community Health

Brown University , 8 North Main Street, Attleboro, MA 02703, U.S.A.;

telephone: (508) 472-2809; e-mail: <degilman@.... He has

served as a litigation consultant and expert witness at the request

of

victims in cases involving brake manufacturers and chemical

companies.

He has also consulted for brake manufacturers in litigation and

non-litigation contexts. Howe is a research assistant at

Never Again Consulting, Attleboro, MA.

Address correspondence and reprint requests to Dr. Egilman.

*Dr. Gori, the editor of Regulatory Toxicoloby and Pharmacology,

refused to publish our response, and we have, therefore, submitted

it to IJOEH, since Dr. Wong based his editorial on criticism that was

published in IJOEH.

espouse the superiority of epidemiologic evidence. The

scope of medical evidence that substantiates his considerations

is inclusive of all the available data.

Until the 1960s, pathology and its attendant clinical

case reports were the primary tools used to ascertain

cause–effect relationships.7–9 The first textbook of epidemiology

was published in 1960; yet physicians agreed

on cause–effect relationships for a wide variety of

chronic diseases long before that time.10 For example,

in 1942, the carcinogenicity of nickel carbonyl was

firmly established based on case reports, pathology,

and toxicology alone.11

Epidemiology is an important tool for assessing the

likely levels of risks associated with different populations,

but—as Koch and Hill's guidelines demonstrate—

it is not a requisite component in the process of

determining that there is a risk. Animal studies can also

provide information about relative risk. Epidemiologic

studies often suffer from design limitations that to do

not account for the inconstant nature of workplace

conditions and exposure levels, the often-delayed

expression of occupational disease, which can remain

latent for 50 years or longer, and the fact that studies

often omit minorities and women. By the time clear

and convincing large-scale epidemiologic evidence

emerges, exposures may span a generation or more,

with the evidence representing little more than a

belated confirmatory footnote to a fully matured

public health calamity. As the epidemiologist

Ozonoff has underscored, " a public health catastrophe

is a health effect that is so powerful that even an epidemiologic

study could detect it. "

12

WONG REJECTS CASE REPORTS

IN CAUSATION

Wong rejects the use of case reports for determinations

of causation: " Case reports, although useful in hypothesis-

generation, are of limited value in causation assessment

in chronic diseases such as cancer. Causation

assessment in chronic diseases must be based on appropriately

designed epidemiologic studies. " 13 He fails to cite

any article or empirical evidence to support this claim.

Contrary to his statement, case reports and case clusters

have consistently been used to establish causation.9

In 1939, Lanza and Vane noted, " Our knowledge of

asbestosis is based on individual case reports . . .

[although] too few to have statistical weight . . . the disease

itself and its pathology have been well demonstrated. "

14 Similarly, it is universally accepted that

Wagner's landmark 1960 publication established the

link between asbestos and mesothelioma in the consciousness

of the scientific community.15 That publication

is simply a report of a series of cases and fails to satisfy

Wong's restrictive definition of " epidemiology. "

Other exposure–chronic disease associations that have

been ascertained by case reports and disease clusters

include: aplastic anemia, as associated with radiation,

benzene, radium, and trinitrotoluene exposures;

peripheral neuropathy and n-hexane, methyl butyl

ketone and dimethylaminopropionitrile; and chemical

hepatitis from dimethylformamide, trinitrotoluene,

tetrahydrofuran, tetrachloroethane, and chlorinated

naphthalenes.9

While many of these cause-effect relationships were

later confirmed with epidemiologic studies, the epidemiologic

evidence was not a prerequisite to the

establishment of precautionary measures or compensation.

In fact, Lanes and Poole observed, " most epidemiologists

found persuasive the early evidence of

vinyl chloride and angiosarcoma of the liver, and

between diethylstilbestrol and adenosarcoma of the

vagina...The acceptance of these two cancer-exposure

relations was firmly established on case reports

alone. " 16 Thus, case reports have contributed to the

determination of causation and, at times, have supplanted

the need for epidemiologic studies. Furthermore,

an epidemiologic study is not required to establish

a harmful effect in each and every occupation

where similar exposures to established carcinogens

occur. Inferences can be made based on exposure data

alone, especially if the cohort is exposed to amounts of

the carcinogen comparable to those seen in cohorts

already studied, as is the case with brake mechanics.

IGNORING OTHER TYPES OF SCIENTIFIC

INFORMATION—ANALOGY AND

PATHOLOGY

Wong cannot and does not dispute the fact that

asbestos causes mesothelioma. He cannot and does not

dispute that no one has ever been able to scientifically

establish a level of exposure to asbestos that eliminates

the risk of cancer. With these two premises established

and unchallenged, determining cause and effect for a

particular individual with mesothelioma relies on the

demonstration of exposure to asbestos and the exclusion

of other even more rare established causes such as

thorium. The work–disease connection is especially

likely since these brake workers are exposed to

asbestos, a known cause of mesothelioma, in the act of

replacing brake linings and breathing the resulting

dust.17–19 Grinding and drilling of brake linings in manufacturing

finishing processes is known to cause high

exposure. Mechanics working similarly with asbestoscontaining

brakes and clutches are exposed to high

concentrations of asbestos; therefore, the epistemological

inquiry into mesothelioma causation ends unless

an alternative etiological factor is present. Faced with

this conundrum, Wong and others attempt to use epidemiologic

studies to refute the indisputable fact that

asbestos causes cancer. Wong relies on a variety of epidemiologic

studies that contain no information regarding

the quantum and duration of exposure to asbestos

VOL 13/NO 1, JAN/MAR 2007 • www.ijoeh.com Against Anti-health

Epidemiology • 119

as proxies for studies of exposure to asbestos in a specific

population.20 He draws inferences from these studies

to explain away the hundreds of reported cases of

mesothelioma in brake mechanics.21

Wong might assert that these workers are exposed

only to chrysotile, since it is the most common fiber

type found in brakes and—as others have tried to

argue—chrysotile causes lung cancer but not pleural

cancer. However, The World Trade Organization found

the evidence of mesothelioma risk substantial enough

to allow countries to ban the importation of

chrysotile.22 Moreover, recent pathologic studies

demonstrate that short, thin chrysotile fibers are more

prevalent than amphibole fibers in mesothelioma

tumors and pleural tissues, providing more evidence

that chrysotile is a cause of mesothelioma.23 The epistemology

of asbestos–mesothelioma causation requires

contact between the causative substance and pleural

tissue. Regardless of the carcinogenicity of chrysotile, it

is also important to note that some brake manufacturers

used amosite, and that tremolite—an undisputedly

mesotheliogenic fiber—is a universal contaminant of

the Canadian chrysotile most often used in brakes and

therefore brake workers were exposed to both amphiboles

and chrysotile.24,25

MISCLASSIFICATION BIAS AND USE OF

INAPPROPRIATE (EXPOSED) CONTROLS

All of the studies Wong relies on in his 2001 meta-analysis

have results significantly below the null. These

results raise a red flag, simply because it is illogical to

maintain that asbestos exposure or some unknown

brake-work–associated confounder reduces the risk of

asbestos-related disease. The more likely explanation

for this counterintuitive result is systematic misclassification

bias. Choice of questions and viewpoint are of

particular import when evaluating levels of interview or

recall bias. Although some of the asbestos studies analyzed

by Wong took steps to minimize interview bias,

they could not eliminate it.

Interviews of brake workers can successfully elicit

the numbers of brake linings they have changed for

various employers, as well as how many they did in years

past in after-school jobs and for friends and family. In

post-mortem interviews, family members are much less

likely to possess such complete histories and to know all

possible sources of exposures in the work situation,

leading to an underestimation of exposures.1 Persons

not fluent in English are at special risk.

Wong claims that people are likely to overestimate

exposures due to their knowledge that exposure to

asbestos causes mesothelioma. However, many people

are still unaware of their potential sources of exposure

and—as reported by Wong—they may feel that their

exposures were trivial.13 The belief that encounters

with asbestos products resulted in trivial exposures

would lead to underreporting of these exposures; the

fact that many people are unaware that brakes

presently contain asbestos means that the underestimation

may occur in an area of particular interest to

Wong's meta-analysis, and explain his counterintuitive

findings.1

In Wong's discussion of " Misclassification of occupation

or exposure, " he criticizes the way several of the

studies reported in his 2001 meta-analysis combined

work categories. He states that grouping auto mechanics

with insulators and other end-product users is inappropriate

because their exposure levels are different.

We agree, and this is another reason we found his use

of these studies to be inappropriate. He fails to address

the clear misclassifications of occupations and exposures

and other problems in the studies he favors.

Wong (200120) ignores Teschke et al.'s statement

that " grouping of occupations was likely to result in

non-differential misclassification, usually biasing risk

estimates to the null value. " 26 The Woitowitz and

Rodelsperger study, while designed to address auto

mechanics in particular, is of limited value because they

used lung cancer patients as controls.27 As asbestos

exposure is a well-known risk factor for lung cancer, any

mesothelioma case–control study with lung cancer

patients as the control group will almost certainly produce

results at or below the null. This study demonstrates

the effect of using inappropriate controls, since

the researchers used " population controls " as well as

lung cancer patients as controls. The odds ratio (OR)

for " hospital controls " was 0.75, almost half that for

population controls, 1.32.27 Rather than report both

results, Wong presented an average OR of 0.87.20

Mc and Mc conducted a case–control

study, published in 1980, examining mesothelioma

cases from Canada and the United States. Like the

Woitowitz and Rodelsperger study, this study relied on

an inappropriate control group; control subjects were

patients in cases diagnosed by pathologists in which

" pulmonary metastases were present from a non-pulmonary

malignant tumor. " 28 The control group did not

exclude patients who had died from asbestos-induced

malignancies such as peritoneal mesothelioma, laryngeal

cancer, or colon cancer. The choice of such controls

biased results towards the null.

In the case of the Hansen (198929) study, Wong omitted

the mesothelioma findings from a relevant study

that contradicted his findings, and in his original publication,

he failed to explain his omission. He retrospectively

excuses the omission by dismissing the one

case of mesothelioma in 21,800 people over ten years

allegedly " because Hansen ignored the background

risk. " 13 This is incorrect, and in his original paper

Wong relies on Hansen and quotes Hansen's SMRs for

lung cancer. 20 Table 4 in Hansen's paper presents

population SMR data for mesothelioma.29 Since no

cases were expected, Hansen did not present an SMR

120 • Egilman, Howe www.ijoeh.com • INT J OCCUP ENVIRON HEALTH

for mesothelioma but did conclude that he had found

an excess of mesothelioma in brake workers. The age

distribution of Hansen's cohort explains the low

expected number for mesothelioma. Table 1 shows that

64% of those studied were between 25 and 34 years old

at the end of follow-up. Another 21% were 35–44 years

old. Accordingly, at the end of the study, 85% of the

study population were under the age of 45, 92% were

under 55, and 96% were under 65.29 Wong used the

United States NCI annual incidence rate of ten per million

to suggest that two cases " could be expected " in

Hansen's Danish cohort. However, the incidence rate is

not synonymous with the background rate, and Denmark

is not the United States. The incidence rate

includes exposed workers, while the background rate is

meant to estimate the combined number of " idiopathic "

mesotheliomas and those caused by other

known causes, e.g., thorium and exposures to asbestos

in ambient air.

The generally accepted " background rate " is only

one or two per million, not ten per million, and even

this rate includes workers who have actual but undocumented

exposures.22 This is easily shown by comparing

" general " female with male mesothelioma rates. Since

there is no evidence that gender is a confounder for

mesothelioma risk, the true background rate is below

the female rate, since some females have direct and

indirect asbestos exposures. All the epidemiologic studies

Wong cites use this elevated " background " as a comparator.

However, even using the upper end of this

background rate, the expected rate in Hansen's nearly

22,000 auto mechanics is only 0.44. Thus, Hansen's

conclusion that even one case is indicative of possible

increased risk is valid and this study provides useful

information about possible mesothelioma causation.

MISINTERPRETING THE MEANING OF A

RATE RATIO OF 1 OR LESS

Dr. Wong and other automobile consultants misinterpret

epidemiologic studies whose results reveal a rate

ratio of 1 or less for a particular exposure. They interpret

such results to mean that the exposure in question

does not and cannot cause the disease in question.

However, as Greenland has noted, this is a complete

misinterpretation of epidemiologic results.30 First, they

failed to recognize the definition of a " caused " case. A

caused case is either a new case that would not have

occurred but for the exposure or, equally, a case that

would have occurred without the exposure but that

occurs sooner than it would have occurred due to the

exposure. For example, if a person was destined to get

breast cancer at age 80 but an exposure to a toxic substance

causes that person to develop breast cancer at

age 40 this is a " caused " case. As Greenland has noted,

the usual age stratification used in epidemiologic studies

does not account for this phenomenon.30 At the

extreme, since we are all going to die, all rate ratios that

compare total deaths between populations are always 1

after 100 (or even fewer) years of latency. This does not

mean that some of the deaths were not caused by occupational

or environmental exposures.

Figure 1 illustrates this principle; in a hypothetical

population of 100 women, we expect three deaths to

occur as a result of breast cancer at ages 70, 80, and 90.

If the exposed population also has three deaths from

breast cancer but these cases occur at ages 30, 35, and

40, it is reasonable to infer that the exposure in question

" caused " all three cases, because they occurred

sooner then they otherwise would have. The exposed

cases would have lost over 150 years of useful life and

their families and friends would have been deprived of

VOL 13/NO 1, JAN/MAR 2007 • www.ijoeh.com Against Anti-health

Epidemiology • 121

Figure 1—Age at death (exposed vs unexposed). If

three people are predisposed to a disease, but exposure

to a carcinogen causes it to occur much earlier

than it would have otherwise, the exposure is a cause

of the disease. Age stratification in epidemiologic studies

may detect the extremes such as those in Figure 1,

but not instances where the exposure causes disease

to occur only a few years earlier than might otherwise

have happened (Figure 2).

Figure 2—Age at death (exposed vs unexposed) undetected

by standard stratification.

considerable love, learning, and friendship. If the exposure

was accompanied by a specific biologic marker,

then scientists could provide unequivocal evidence to a

jury that the exposure in question was, on a more-likelythan-

not basis, a contributing factor for the deaths of

these individuals. Therefore, from a compensation perspective,

if negligent conduct caused the exposures, all

three individuals would deserve recompense, despite

the fact that the RR was 1.

Unfortunately, Dr. Wong and other automobile consultants

claim that their meta-analyses, which have

reported a statistically significant protective effect of

work with brakes, " prove " that asbestos brakes have not

and cannot cause mesothelioma in any individual.31–33

This is a misinterpretation of epidemiologic studies,

which may result in an injustice to victims. More importantly,

since some of the auto industry consultant epidemiologists

have used this misinterpretation to lobby

the government to remove warnings from brake products

and to stop training brake mechanics in how to

avoid the hazards of work with asbestos brakes, it may

serve an anti-public health purpose.34

PETROLEUM WORKER STUDIES—

CONTRACTOR BIAS

In the analysis of studies of petroleum workers, Wong's

systematic misclassification once again biases his results

toward the null hypothesis. Wong relies on company

employment records as a surrogate for exposure. Petroleum

refineries and chemical plants frequently use contractors

to perform the most hazardous jobs, such as

asbestos insulation and tank cleaning, during turnarounds.

As a result, the most heavily exposed workers

are often not official " employees " but are contract

workers who are not included in employment records.

Ironically, these workers may work for decades at a

single plant.35 They are not subject to exposure monitoring,

and they are often not aware of potential exposures

because the plant health and safety officers do

not provide them with safety training or equipment.

These most heavily exposed individuals are included in

Wong's studies as unexposed controls (i.e., as part of

the general population). Additionally, many plant

employees whom Wong classifies as exposed are never

subjected to chemicals at work (e.g., lawyers, office personnel,

etc.). Wong empties the top left square of the

classic 2-by-2 epidemiologic contingency table (the " A

box " ) by moving the exposed–diseased cases that

belong there into the " D box " (unexposed–diseased).

Such flaws in method bias Wong's results to the null.

In fact, as with his asbestos studies, this systematic

reclassification

of exposed–diseased to unexposed–diseased

often results in the counterintuitive result that exposures

to benzene and other known carcinogens reduce

the risks of developing cancer. In 1958, Wong's predecessor

consultant to the petroleum industry,

Kehoe, told the American Petroleum Institute that

" turnarounds " (refurbishment of refinery installations)

were being subcontracted by many companies and that:

It was quickly discovered as various industrial situations

were studied also that job titles differed greatly

from company to company and indicated little if any

relationship to the degree and kind of hazard. Men

charged with the actual operation of the equipment,

for example in the petroleum industry, may have

little or no contact with either petroleum or the various

refined products, but the samplers and maintenance

personnel may have frequent and heavy exposures.

At the same time, maintenance and labor

classifications may work in all parts of the refinery

or, equally common, be assigned for a specific

period of time to a single area. . . . Maintenance

problems are reduced and even presently widely

spaced " turn arounds " are being substantially subcontracted

by many companies. It must be admitted

that, although this development may diminish the

exposure hazard for the refinery worker, it may be

concentrating and augmenting the danger to

employees of companies specializing in this servicetype

operation.35

In response to the criticism that he has misclassified

study subjects, Wong says, " the issue of cohort definition

and misclassification bias in occupational studies

should be examined in terms of the availability of

records for cohort identification and subsequent analysis

within the cohort. " 13 The petroleum and chemicals

companies' cohorts are large, and these companies can

distinguish the lawyers from the operators. Thus, Wong

could have and should have compared cancer rates

between exposed and truly unexposed workers from

the same company, in the same location. He has, in

fact, done this in the past. In one of his previous studies

he noted:

[T]he provision of an internal comparison group

was designed to minimize several problems resulting

from comparing an occupational cohort with the

general population. Such an internal comparison

group is regarded as the most appropriate basis for

comparison in occupational mortality studies. In

particular, comparisons based on internal groups

were found to be more sensitive in detecting the

association between leukemia and occupational

exposure to benzene, and in demonstrating a suggested

increase in lymphoma mortality with

increased exposure among refinery workers.36

In his earlier studies, Wong even compared the use

of internal and population control groups and documented

the dramatic lowering of the apparent risk

caused by the use of population controls. He wrote:

The group exposed to benzene in this study did not

show any significant mortality excess from all lym-

122 • Egilman, Howe www.ijoeh.com • INT J OCCUP ENVIRON HEALTH

phatic and haematopoietic cancers combined when

compared with the United States male population,

although some of the lymphopoietic cancer SMRs

were slightly raised. When the SMRs for the occupationally

exposed group (continuous and intermittent)

were compared with those for the occupationally

unexposed, however, a considerable difference

was observed for all lymphopoietic cancer (121.1 v

34.6), lymphosarcoma and reticulosarcoma (112.8 v.

50.8), leukemia (117.4 v 0), and other lymphatic

tissue cancer (163.8 v 53.0). . . . Hence, mortality

from all non-Hodgkin's lymphopoietic cancers

(non-Hodgkin's lymphoma and leukemia) was

higher in the benzene exposed group than in the

comparison group.36

Despite his knowledge that an internal control

group led to the most sensitive study, all of his subsequent

refinery and chemical plant cohort studies used

general U.S. population controls.

Wong asserts that the critique of his systematic misclassification

should be ignored because evidence for it

came to light during cross-examination in a lawsuit. In

this lawsuit, a worker with acute myelogenous leukemia

(AML) sought compensation from a petroleum company.

13 At the request of the company, Wong argued

that his studies proved that benzene exposure had not

contributed to cause this benzene-exposed worker to

develop AML. We are unaware of any " methodologic "

construct that maintains that the source of a critique

invalidates the criticism. " Consider the source " is not a

scientific defense—it is a baseless, ad hominem affront.

Wong claims he is defending proper methodology in

epidemiologic practice. This claim is common among

scientists defending corporate interests. For the past

decade, the chemical and tobacco industries have

funded the promotion of what they refer to as " good

epidemiological practices. " These " practices " seek to

prevent plaintiffs from prevailing in court and thwart

government regulations.37 In Wong's case, his " proper

methodology " systematically underestimates disease

among exposed workers. Just as following a flawed

recipe will yield a foul-tasting cake, no matter how well

you follow it, a faulty method will always yield unreliable

results. In following his bad recipe to the letter,

Wong ends up with a study marred by interview and

misclassification bias. His manipulation of the " A Box "

is akin to the effect that substituting salt for sugar has

on a cake recipe; it makes for a bad-tasting cake. In epidemiology

this makes for a very bad study. Although, in

both cases the seller can represent that he followed the

recipe to the " T. "

CONCLUSION

Wong has repeatedly conducted studies on behalf of

chemical and manufacturing companies that use these

studies to defend against lawsuits and to influence government

regulation.1 Auto-parts manufacturers,

including General Motors, Ford, Chrysler, and Bendix,

have paid Wong and others tens of millions of dollars

to promote views that will protect their profits.† They

also have harassed experts whose views conflict with

their economic goals.38

As Nobel laureate Gunnar Myrdal observed, " There

can be no view except from a viewpoint. Prior to

answers there must be questions. In the questions

raised, the viewpoint has been chosen and the valuations

implied. " Wong chose his viewpoint before his

questions, and his valuations biased his studies. He may

follow his method to the letter, but it is a flawed

method that leads to flawed results. Those results ultimately

help his sponsors pollute the environment and

injure workers and consumers.

References

1. Egilman DS, Billings MA. Abuse of epidemiology: automobile

manufacturers manufacture a defense to asbestos liability. Int J

Occup Environ Health. 2005;11:360-71.

2. Egilman DS, Bohme SR. Over a barrel: corporate corruption of

science and its effects on workers and the environment. Int J

Occup Environ Health.2005;11:331-7.

3. Bohme SR, Zorabedian J, Egilman DS. Maximizing profit and

endangering health: corporate strategies to avoid litigation and

regulation. Int J Occup Environ Health.2005;11:338-48.

4. Gennaro V, Tomatis L. How epidemiologic studiesmay underestimate

or fail to detect increased risks of cancer and other diseases.

Int J Occup Environ Health. 2005;11:356-9.

5. AS. Causation and disease: the Henle–Koch postulates

revisited. Yale J Biol Med. 1976;4):175-95.

6. Hill SAB. The environment and disease: association or causation?

1965. Proc Royal Soc Med. 1965.

7. Oppenheimer GS. Heart Disease and the Emergence of

Modern Epidemiology. Research Foundation of The City University

of New York Annual Report, 532003.

8. Egilman DS. Public health and epistemology. Am Jf Indl Med.

1992;22:457-9.

9. Fleming LE, Ducatman AM, Shalat SL. Disease clusters: a central

and ongoing role in occupational medicine. J Occup Med.

1991;33:818-27.

10. MacMahon B. Epidemiologic Methods. Boston: Little, Brown,

1960.

11. Hueper WC. Occupational Tumors and Allied Diseases. Springfield,

IL: C , 1942.

12. Ozonoff D. National Public Radio, Talk of the Nation.

<http://www.mindfully.org/Health/Cancer-Clusters-NPR.

htm>. 7/27/2001.

13. Wong O. The interpretation of occupational epidemiologic

data in regulation and litigation: Studies of auto mechanics and

petroleum workers. Regul Toxicol Pharmacol 2006;44:191-7.

14. Lanza AJ, Vane RJ. Industrial dusts and the mortality from

pulmonary

disease. Am Rev Tuberculosis. 1939;39:419-20.

15. Wagner JC, Sleggs CA, Marchand P. Diffuse pleural mesothelioma

and asbestos exposure in the North Western Cape

Province. Br J Ind Med. 1960;17 October::260-71.

16. Lanes SF, Poole C. Truth in packaging? The unwrapping of

epidemiologic

research. J Occup Med. 1984;26:571-4.

VOL 13/NO 1, JAN/MAR 2007 • www.ijoeh.com Against Anti-health

Epidemiology • 123

†Goldenberg, , Hellen and Antognoli document: Plaintiffs

vs. A. W. Chesterton, Inc. et al. Payments to Dennis Paustenbach in

2004 - 2005: $275,070.99; Hessel in 2006: $3,221.24;

Goodman in 2005: $7,964.60; Garabrant in 2003–2005:

$74,034.93; Exponent, Inc., in 2002–2006: $5,407,986.77; Chemrisk,

Inc., in 2003–2006: $1,947.908.79; Epilung Consulting in 2004–2006:

$204,284.18.

124 • Egilman, Howe www.ijoeh.com • INT J OCCUP ENVIRON HEALTH

17. Castleman B, Camarota LA, Fritsch AJ, Mazzocchi S, Crawley

RG. The hazard of asbestos for brake mechanics. Public Health

Rep. 1975;90:254-6.

18. Hickish DE, Knight KL. Exposure to asbestos during brake

maintenance. Ann Occup Hyg. 1970;13:17-21.

19. Lorimer WV, Rohl AN, A, Nicholson WJ, Selikoff IJ.

Asbestos exposure of brake repair workers in the United States.

Mount Sinai J Med. 1976;43:207-18.

20. Wong O. Malignant mesothelioma and asbestos exposure

among auto mechanics: appraisal of scientific evidence. Regul

Toxicol Pharmacol. 2001;34:170-7.

21. Lemen RA. Asbestos in brakes: exposure and risk of disease. Am

J Ind Med. 2004;45:229-37.

22. World Trade Organization. Dispute Settlement Reports.

London, U.K.: Cambridge University Press, 2001. Report No.: 8.

23. Suzuki Y, Yuen SR, R. Short, thin asbestos fibers

contribute

to the development of human malignant mesothelioma:

pathological evidence. Int J Hyg Environ Health. 2005;208:201-

10.

24. Hubbard JL, Longo WE. Detection of Amphibole Asbestos in

Chrysotile Sheet Gaskets. Raleigh, NC: Material Analytical Services,

Inc.; 2002

25. Hickish DE. Report 41/68: Exposure to Asbestos during the

Servicing

of Brakes of Passenger Cars. 1968.

26. Teschke K, M, Checkoway H, et al. Mesothelioma

surveillance

to locate sources of exposure to asbestos. Can J Public

Health. 1997;88:163-8.

27. Woitowitz H-J, Rodelsperger K. Mesothelioma among car

mechanics? Ann Occup Hyg. 1994;38:635-8.

28. Mc AD, Mc JC. Malignant mesothelioma in

North America. Cancer. 1980;46:1650-6.

29. Hansen ES. Mortality of auto mechanics. A ten year follow-up.

Scand J Work Environ Health. 1989;15:43-6.

30. Greenland S. Relation of probability of causation to relative

Risk

and doubling dose: a methodological error that has become a

social problem. Am J Public Health. 1999;89:1166-9.

31. Tarry S. <http://www mcguirewoods com/lawyers/index/

samuel_l_tarry_jr asp>. 2006;

32. Egilman DS, Kim J, Biklen M. Proving Causation: The use and

abuse of medical and scientific evidence inside the courtroom—

an epidemiologist's critique of judicial interpretation of

the Daubert ruling. Food and Drug Law J. 2003;58:223-50.

33. Hardy H, Egilman D. Corruption of occupational medical

literature:

the absestos example. Am J Ind Med. 1991;20:127-9.

34. Stampher M. Letter to the EPA, Re: " Guidance for preventing

asbestos disease among auto mechanics. " EPA Docket # EPAHQ-

OPPT-2003-0074. 3/24/2004.

35. Kehoe R. An Epidemiologic Study of Cancer among Employees

in the American Petroleum Industry. 1958.

36. Wong O. An industry wide mortality study of chemical workers

occupationally exposed to benzene I: General results. Br J Ind

Med. 1987;44:365-81.

37. Ong EK, Glantz SA. Tobacco industry efforts to subverting

International Agency for Research on Cancer's second-hand

smoke study. Lancet. 2001;335:1253-9.

38. Memorandum in support of motion of Egilman, MD,

MPH, to intervene and for protective order in Daimler Chrysler

Corp. v and Janet Unden. 8/31/2006. P. Barylick,

Esq., State of Rhode Island Superior Court, Providence.