AMERICAN JOURNAL OF INDUSTRIAL MEDICINE 45:382–385 (2004)
Implications of the Precautionary Principle
in Research and Policy-Making
Philippe Grandjean, MD, PhD,1,2?John C. Bailar, MD, PhD,3David Gee, BA,4
Herbert L. Needleman, MD,5David M. Ozonoff, MD, MPH,6Elihu Richter, MD,7
Morando Sofritti, MD,8and Colin L. Soskolne, PhD9
international law. The key element is the justification for acting in the face of uncertainty.
The PP is thereby a tool for avoiding possible future harm associated with suspected, but
not conclusive, environmental risks. Under the PP, the burden of proof is shifted from
demonstrating the presence of risk to demonstrating the absence of risk and it is the
responsibility of the producer of a technology to demonstrate its safety rather than
the responsibility of public authorities to show harm. Past experiences show the costly
consequences of disregarding early warnings about environmental hazards. Today, the
into disease causation, to elucidate the full scope of potential adverse implications result-
ing from environmental pollutants, and to identify opportunities for prevention. Research
and to support ourconfidence in applying the PP fordecision-making in the public policy
arena. Am. J. Ind. Med. 45:382–385, 2004. ? 2004 Wiley-Liss, Inc.
KEY WORDS: environmental health; hazardous substances; primary prevention;
public health; risk assessment
The Precautionary Principle (PP) has recently been
formally introduced into national and international law,
including the European Union [European Commission,
2000]. The key element is the justification for acting in the
face of uncertain knowledge about risks from environmental
exposures. Appropriate public health action should be taken
in response to limited, but plausible and credible, evidence
of likely and substantial harm. The PP is thereby aimed at
avoiding possible future harm associated with suspected, but
not conclusive, environmental risks. The burden of proof is
shifted from demonstrating the presence of risk to demon-
strating the absence of risk.
has always been advised in regard to decisions affecting
human health. The lack of precautionary or preventive
action despite early warnings has resulted in severe harm to
2Department of Environmental Health, Harvard School of Public Health, Boston,
7Unit of Occupational and Environmental Medicine, Hebrew University-Hadassah,
8European Ramazzini Foundation of Oncology and Environmental Sciences,Bentivoglio
9Department of Public Health Sciences and the John Dossetor Health Ethics Centre,
health and damage to ecosystems [European Environment
Agency, 2001]. Too often, unreasonable delays in preventive
responses have resulted from demands for detailed proof
of causation, including detailed knowledge of mechanistic
actions. The adverse effects on human health and ecosystem
sustainability have in some cases been catastrophic.
Asbestos is perhaps the best known example of a
preventable health hazard for which there was credible
evidence of harm long before the causal link was generally
accepted. However, asbestos is still being used in many
developing countries, with attendant increases in cancer and
pulmonary disease. The same sequence of events occurred
with other toxic materials, such as benzene. The toxic pro-
perties of these well-known hazards have now been clearly
documented, and prevention efforts are justified without any
need to employ the PP. Had appropriate action been taken
when the first evidence of adverse effects emerged, then a
truly precautionary intervention could have ensued, prevent-
ing much illness and many premature deaths.
Similarly, when tetraethyllead was synthesized in the
1920s, a strong precautionary argument was made by Prof.
Yandell Henderson of Yale University. A medical physiol-
poison the entire planet. His warning was ignored, and the
large-scale production of lead additives began. At its peak,
250,000 tons of lead were released into the atmosphere each
year in the US alone. The cessation of this practice began in
the 1970s and was completed in the US, most of Europe, and
Japan 20 years later. The US average blood-lead concentra-
(0.10 mmol/L) in 1999. For the US birth cohort of children
born in 1998 alone, the monetized benefit from this lowering
of their blood-lead concentration was estimated by the US
Centers for Disease Control and Prevention to range in the
hundreds of billions of dollars. Still, while this primary
prevention effort was highly successful, it was not precau-
tionary, because it was initiated only after a preponderance
of evidence had become available.
Many new and potentially hazardous chemicals are
released in the environment, and the magnitude of exposures
and their effects, singly and in combination, are only partially
understood. The need for precautionary action has, therefore,
increased. In recognizing this imperative, the European
Commission  Communication on the PP states that
precautionary decisions are indicated when preliminary
objective scientific evaluation suggests reasonable grounds
in relation to more proximate exposures. In regard to global
complex mix of determinants that act to degrade environ-
mental carrying capacities and endanger sustainable develop-
its various life-supporting ecosystems, is essential to human
catastrophic impacts on health related to environmental
degradation, the greater the need for precaution. Because the
consequences for human health and well-being from changes
more proximate hazards, precautionary strategies and actions
are even more critical for minimizing harm.
LESSONS FROM APPLYING PRECAUTION
In public health, removing one risk may produce others.
Removing one risk may incur other unanticipated hazards.
to be convincing, other additives were sought to raise octane
levels. From a precautionary viewpoint, enough was known
gen in gasoline; instead, methyl-tert-butylether (MTBE), a
less studied octane booster, was employed. This substance
has now produced health risks from severewater contamina-
tion, which was not easily anticipated despite the highly
persistent nature of MTBE. Another example relates to the
microbiological safety of drinking water in developing
countries, where precaution suggested that well-water from
with infectious agents than surface water. However, the
release of arsenic and fluoride from naturally occurring sub-
terranean sources have now caused mass intoxication in
Bangladesh and India.
In these examples, the preventive approach failed to
anticipate the risk from the substitutes. Especially when the
evaluate the intervention becomes even more important.
Current risk assessment and standard-setting are gen-
erally based on limited information about single hazards.
Uncertainty has been taken into account when hypothesized
legal limits. However, the history of declining exposure
limits for many substances testifies to the failures of this
approach. Early warnings were often misinterpreted or
ignored [European Environment Agency, 2001].
SCIENCE FOR PRECAUTION
The tools and principles of science, when applied to
public-policy, have often implicitly worked against pre-
caution. In particular, the absence of universally accepted
evidence demonstrating damage or harm has often been
misinterpreted as evidence of safety.
Implementing a precautionary approach in decision-
making must be linked to established scientific principles.
The PP, therefore, has important implications for scientific
strategies, methods, interfaces with policy-making, and risk
communication. These considerations present science with
Precautionary Principle in Research and Policy-Making383
opportunities and challenges, and relate to the way scientific
The interrelationships, for instance between science and
policy, should reflect the need for new knowledge generated
from using precaution in public-health decisions. Although
this prevention strategy should stimulate the development
public confidence in applying the PP for decision-making.
of results or observations. Still, by insisting on confirmatory
evidence, science may penalize false positives more than
to reject an hypothesis of no effect, but this tradition places
an unreasonable preference on the absence of an effect.
Under the PP, incomplete, but credible, scientific documen-
tation may be deemed sufficient for decision-making.
A mere replication or simple extension of the scientific
knowledge base beyond a minimal requirement may no
longer constitute a priority. While current research has often
hazards, one by one, an expanded and refocused research
agenda should rather emphasize the sources of variability
and uncertainty, including individual susceptibility, impacts
of mixedand variable exposures, susceptible life-stages, and
Statistical analysis of research data has aimed at testing
whether an effect could be considered a possible result of
natural variability. This strategy has been considered a
necessary part of scientific rigor, although it is often coupled
with demands for the demonstration of causality. It is also
necessary to consider both the severity of potential adverse
outcomes and their likelihood. Although a prerequisite for
implementing the PP, such information has not usually been
considered a standard product of the scientific endeavor.
In general, statistical analysis of research data tends
to bias the conclusions toward the null hypothesis. For
example, the standard application of a limit of 5% for
associations, simply because the study was too small and
thus lacked statistical power, or because some imprecision
or limited sensitivity of the parameters precluded a more
definitive observation. Table I shows examples that current
research practices are biased by producing many more
instances of false negatives than false positives.
Decisions should not be based on an assumed certainty
of scientific information, but rather on an estimate of the
information to facilitate consideration of the risk of both
false negatives and false positives. Type II errors, i.e., over-
looking a true hazard, become an essential issue in the
Uncertainty has always been recognized as a key
component of our incomplete understanding of human
disease and the environment. An important goal now is to
achieve a better characterization of these uncertainties and
TABLE I. Potential Sources of Bias in Environmental Research Using Standard Methodological Approaches in
384Grandjean et al.
their implications. Appropriate statistical methods should,
therefore, be developed and applied to determine the like-
lihood of adverse or extreme outcomes and how they may be
affected by precautionary actions.
Science should continue to update existing knowledge
and to improve communication of risks as well as their
associated uncertainties. In analyzing and communicating
such information, science should help frame the debate and
to apply the PP.
Research should emphasize studies to determine the
societal impacts and benefits of preventive interventions.
was successful and to modify actions that may have resulted
in new risks.
Scientific uncertainty has too often been the excuse for
limiting and postponing preventive efforts. In addition,
debates over details in risk assessments have resulted in
delayed action to protect the public health and the environ-
ment. New research is needed to expand current insight into
disease causation, to elucidate the full scope of potential
implications for health and the biosphere resulting from
the prevention of risks at the source. Recognizing the nature
of cumulative, complex, and synergistic effects on whole
precautionary action, and necessary to avert the serious
consequences of ecological disintegrity.
Current scientific research agendas, funding priorities,
science education, risk communication, and science policy
need to be re-examined in light of the PP. The need for better
links between science and public policy in the spirit of the
PP are clearly warranted as reflected in the statement from
the Collegium Ramazzini published in the present issue of
We recommend that researchers, in collaboration with
research funding agencies, regulatory bodies, and stake-
holders, revise current scientific paradigms and traditions
to ensure that scientific information is used in decision-
making within a framework that facilitates the application of
precaution to supplement prevention.
This article was developed under the auspices of the
Collegium Ramazzini and its Committee on the Precau-
tionary Principle: Substantial input was received at the
Collegium’s International Scientific Conference on The
Precautionary Principle: Implications for Research and
Prevention in Environmental and Occupational Health,
Bologna, 23–24 October, 2002 [Grandjean et al., 2003].
The proceedings from this conference are available at www.
the precautionary principle (COM(2000) 1). Brussels. URL: http://
(accessed, 30 November, 2003).
European Environment Agency. 2001. Late lessons from early
warnings: The precautionary principle 1896–2000. Environmental
Grandjean P, Sofritti M, Minardi F, Brazier J, editors. 2003. The
Precautionary Principle. Implications for research and prevention in
environmental and occupational health. Eur J Oncol Library 2:1–245.
URL: http://www.collegiumramazzini.org (accessed, 30 November,
Precautionary Principle in Research and Policy-Making 385