Environmental pollution and the global

burden of disease

David Briggs
Small Area Health Statistics Unit, Department of Epidemiology and Public Health, Imperial College,
London, UK

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

Exposures to environmental pollution remain a major source of health risk
throughout the world, though risks are generally higher in developing
countries, where poverty, lack of investment in modern technology and weak
environmental legislation combine to cause high pollution levels. Associations
between environmental pollution and health outcome are, however, complex
and often poorly characterized. Levels of exposure, for example, are often
uncertain or unknown as a result of the lack of detailed monitoring and
inevitable variations within any population group. Exposures may occur
via a range of pathways and exposure processes. Individual pollutants may be
implicated in a wide range of health effects, whereas few diseases are directly
attributable to single pollutants. Long latency times, the effects of cumulative
exposures, and multiple exposures to different pollutants which might act
synergistically all create difficulties in unravelling associations between
environmental pollution and health. Nevertheless, in recent years, several
attempts have been made to assess the global burden of disease as a result of
environmental pollution, either in terms of mortality or disability-adjusted life
years (DALYs). About 8–9% of the total disease burden may be attributed to
pollution, but considerably more in developing countries. Unsafe water, poor
sanitation and poor hygiene are seen to be the major sources of exposure,
along with indoor air pollution.

Introduction
Despite the major efforts that have been made over recent years to clean
up the environment, pollution remains a major problem and poses con-
Correspondence to:
David Briggs, Small
tinuing risks to health. The problems are undoubtedly greatest in the
Area Health Statistics developing world, where traditional sources of pollution such as industrial
Unit, Department of emissions, poor sanitation, inadequate waste management, contaminated
Epidemiology and Public
water supplies and exposures to indoor air pollution from biomass fuels
Health, Imperial College,
London, UK. E-mail: affect large numbers of people. Even in developed countries, however,
[email protected] environmental pollution persists, most especially amongst poorer sectors

British Medical Bulletin 2003; 68: 1–24 British Medical Bulletin, Vol. 68 © The British Council 2003; all rights reserved
DOI: 10.1093/bmb/ldg019
Impact of environmental pollution on health: balancing risk

of society1,2. In recent decades, too, a wide range of modern pollutants

have emerged—not least, those associated with road traffic and the use
of modern chemicals in the home, in food, for water treatment and for
pest control. Most of these pollutants are rarely present in excessively
large concentrations, so effects on health are usually far from immediate
or obvious. As Taubes3 has noted, few of the problems of environmental
exposure that concern us today imply large relative risks. Detecting
small effects against a background of variability in exposure and human
susceptibility, and measurement error, poses severe scientific challenges.
The progressively larger number of people exposed to environmental

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

pollution (if only as a result of growing population numbers and increasing
urbanization) nevertheless means that even small increases in relative
risk can add up to major public health concerns. The emergence of new
sources of exposure and new risk factors, some of them—such as endocrine
disruptors—with the capacity to have lifelong implications for health,
also means that there is a continuing need for both vigilance and action.
As the impact of human activities and issues of environmental health
become increasingly global in scale and extent, the need to recognize and
to address the health risks associated with environmental pollution
becomes even more urgent. Effective action, however, requires an under-
standing not only of the magnitude of the problem, but also its causes and
underlying processes, for only then can intervention be targeted at where
it is most needed and likely to have greatest effect. As background to the
other chapters in this volume, therefore, this chapter discusses the nature
of the link between environmental pollution and health and considers the
contribution of environmental pollution to the global burden of disease.

Links between environmental pollution and health

Environmental pollution can be simply, if somewhat generally, defined
as the presence in the environment of an agent which is potentially
damaging to either the environment or human health. As such, pollutants
take many forms. They include not only chemicals, but also organisms
and biological materials, as well as energy in its various forms (e.g. noise,
radiation, heat). The number of potential pollutants is therefore essentially
countless. There are, for example, some 30,000 chemicals in common use
today, any one of which may be released into the environment during
processing or use. Fewer than 1% of these have been subject to a detailed
assessment in terms of their toxicity and health risks4. The number of
biological pollutants is truly unquantifiable. They include not only living
and viable organisms, such as bacteria, but also the vast array of endotoxins
that can be released from the protoplasm of organisms after death. There
is, therefore, no shortage of potential environmental risks to health. What

2 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

is lacking, for the most part, is an understanding of the nature and mecha-
nisms of these risks.

The source–effect chain

The link between pollution and health is both a complex and contingent
process. For pollutants to have an effect on health, susceptible individuals
must receive doses of the pollutant, or its decomposition products, sufficient
to trigger detectable symptoms. For this to occur, these individuals must

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

have been exposed to the pollutant, often over relatively long periods of
time or on repeated occasions. Such exposures require that the susceptible
individuals and pollutants shared the same environments at the same
time. For this to happen, the pollutants must not only be released into
the environment, but then be dispersed through it in media used by, or
accessible to, humans. Health consequences of environmental pollution
are thus far from inevitable, even for pollutants that are inherently toxic;
they depend on the coincidence of both the emission and dispersion
processes that determine where and when the pollutant occurs in the
environment, and the human behaviours that determine where and
when they occupy those same locations.
The whole process can simply be represented as a causal chain, from
source to effect (Fig. 1). As this indicates, most pollutants are of human
origin. They derive from human activities such as industry, energy
production and use, transport, domestic activities, waste disposal, agri-
culture and recreation. In some cases, however, natural sources of pollution
may also be significant. Radon, released through the decay of radioactive
materials in the Earth’s crust, arsenic released into groundwaters from
natural rock sources, heavy metals accumulating in soils and sediments
derived from ore-bearing rocks, and particulates and sulphur dioxides
released by wildfires or volcanic activity are all examples.
Release from these various sources occurs in a wide range of ways, and
to a range of different environmental media, including the atmosphere,
surface waters, groundwaters and soil (Fig. 2). Estimates of emission by
source and environmental medium are inevitably only approximate, for
they can rarely be measured directly. Instead, most emissions inventories
derive from some form of modelling, either based on emission factors
for different processes or source activities5 or on input–output models
(i.e. by calculating the difference between quantities of the material
input into the process and quantities contained in the final product).

Atmospheric emissions
Emissions to the atmosphere tend to be more closely modelled and
measured, and more generally reported, than those to other media, partly

British Medical Bulletin 2003;68 3

Impact of environmental pollution on health: balancing risk

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

Fig. 1 The source–effect chain.

because of their greater importance for environmental pollution and

health (emissions to the atmosphere tend to be more readily discernible and
to spread more widely through the environment), and partly because of
the existence of better established policy and regulation. Figure 3 shows
the main sources of emissions of selected pollutants in the European Union.
As this shows, combustion represents one of the most important emission
processes for many pollutants, not only from industrial sources, but also

4 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

Fig. 2 Sources and pathways of emission into environmental media. Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

from low-level sources such as motorized vehicles and domestic chimneys,

as well as indoor sources such as heating and cooking in the home or
workplace. Emissions from industrial combustion or waste incineration
tend to be released from relatively tall stacks, and often at high temperature,
with the result that they are dispersed widely within the atmosphere.
Emissions from low-level sources such as road vehicles and low-temperature
combustion sources such as domestic heating, in contrast, tend to be
much less widely dispersed. As a result, they contribute to local pollution

British Medical Bulletin 2003;68 5

Impact of environmental pollution on health: balancing risk

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

Particulates Ozone Acidifying
precursors substances

Fig. 3 Atmospheric emissions by source in the European Union (EU15). Note: Ozone precursors
include methane, NMVOCs, NOx and CO; acidifying substances include NO2, SO2, NH3.
Source: European Environment Agency.

hotspots and create steep pollution gradients in the environment. In urban

environments, for example, traffic-related pollutants such as nitrogen
dioxide and carbon monoxide typically show order-of-magnitude variations
in concentration over length-scales of tens to a few hundred metres6.
Evaporation and leakage are also important emission processes contrib-
uting to local variations in environmental pollution. In the UK, releases
from filling stations account for ca. 1.8% of benzene emissions; leakages
from gas pipelines contribute ca. 13.7% of methane emissions to the
atmosphere; evaporation and leakage of solvents during processing and
use produce ca. 40% of atmospheric emissions of non-methane volatile
organic compounds (NMVOCs)7. In addition, abrasion, corrasion and
corrosion release significant quantities of emissions to the atmosphere.
Wear and tear of catalytic converters during operation is a major source
of platinum emissions8, for example, whereas tyre wear (corrasion) and
road wear (abrasion) account for about 16% of particulate emissions
from road transport and almost 97% of zinc emissions from road
transport7—and perhaps more where studded tyres are used9.
These fugitive and local emissions are often overlooked in epidemio-
logical and other studies that use modelling techniques to estimate
exposures, but they can be extremely important, both because they are
frequently responsible for the highest concentrations of environmental

6 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

pollution, and because—unlike high-level emissions—they remain close to

source and show marked dilution gradients with distance from source.
In many cases, they may therefore be the real sources of variation being
considered when distance is used as a proxy for exposure around point
industrial sources in epidemiological studies.

Emissions to surface water, groundwater and soil

Releases to other media, such as surface waters, groundwaters and soil,
also occur through a range of processes. Deliberate discharge, spillage

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

(e.g. from storage, during transport, or during processing and usage),
leakage and runoff (e.g. of agricultural chemicals) are all important in
terms of aqueous pollutants. Legal limits for discharges to streams are
set for many industries, aimed at keeping levels of contamination within
accepted limits. Illegal discharges, or accidental spillage, however, some-
times occur and accounted for the majority of reported surface water
pollution incidents in the UK in 2001, for which the cause is known10.
Dumping (both legally in landfill sites and illegally) represents a major
source of emission of solid wastes, though final release into the wider
environment may only occur when these materials decompose or break
up. Landfill sites may thus be responsible for emissions of a wide range
of pollutants, via different pathways, especially when these sites are
inadequately sealed or poorly maintained11. The contribution of informal
and illegal dumping to environmental pollution is, inevitably, only poorly
known.

Environmental fate
Once released into the environment, pollutants may be transported via
many different processes and pathways, often moving from one medium
to another, and undergoing a wide range of modifications in the process.
Chemical reactions, physical abrasion, sorting by size or mass and
deposition all change the composition of the pollutants and alter the pol-
lution mix. Dilution occurs as pollutants spread outward into a wider
volume of space; concentration may occur as pollutants accumulate in
local ‘sinks’ or in the bodies of organisms, as they pass along the food
chain.
In general, these processes tend to result in some degree of distance-
decay in environmental concentrations, if only because the opportunity
for dilution, decomposition and deposition increases with increasing
distance of transport. It is largely on this basis that distance is often used
as a surrogate for exposure in many epidemiological studies. The realities
of environmental patterns of pollution are, however, often much more
complex than these simple distance-based models imply. They also vary
greatly between different pollutants and environmental media, because

British Medical Bulletin 2003;68 7

Impact of environmental pollution on health: balancing risk

of the different transportational behaviours that are involved. In addition,

dispersion processes and resulting pollution concentration fields may vary
substantially depending on the prevailing (e.g. meteorological) condi-
tions at the time. Patterns of atmospheric dispersion, for example, differ
not only in relation to windspeed and direction but also atmospheric
stability (e.g. between stable and unstable weather conditions, or when
there is a temperature inversion)6. Movement of many pollutants through
soils occurs mainly as mass flow in water passing through larger pore
spaces and fissures: the irregular distribution of these within highly
structured soils means that dispersion often follows highly discrete

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

pathways12,13. Gaseous pollutants may follow similar preferred pathways.
Releases from landfill sites may thus travel relatively long distances in
the soil or bedrock, before emerging at the surface, where they can cause
local hazards including explosions14. Radon shows the same discrete
and complex pattern, such that concentrations may vary by orders of
magnitude from one home to another in the same district15,16. Modelling
these locally variable pathways poses severe challenges.
To a large extent, the increased opportunity for mixing means that dis-
persion of pollutants in surface and groundwaters is more regular, leading
to more uniform patterns of contamination, at regional scales. In developed
countries, also, considerable water mixing often occurs during treatment
and distribution, so that water quality is relatively uniform across large
areas and populations. Local variations may occur, however, because of
contamination within the distribution system or differences in the length
of the network, and thus in the time available for contamination and
decomposition of the disinfectants incorporated at treatment17. In devel-
oping countries, especially, considerable variations may also occur between
waters in shallow wells, particularly where these are affected by local
pollution sources, such as badly sited latrines or agricultural activities.
Again, this makes exposure assessment difficult, without the ability to
collect data on water quality for individual wells.
Similar difficulties occur in tracking and modelling transport of pollutants
in the food chain. Whilst the general pathways followed by pollutants
are often clear in natural (and some farmed) food chains, in that persistent
compounds tend to accumulate as they pass from one trophic level to
another, the detailed patterns of contamination are often far more complex.
Many animals have very restricted feeding behaviours: even in areas of
open grazing land, for example, sheep tend to focus on distinct home
ranges from which they rarely stray18. As a result, marked variations
may occur in contaminant uptake by livestock, even over short distances,
as illustrated by patterns of contamination from the Chernobyl incident
in the UK19. Significant accumulation of these contaminants in humans
likewise tends to occur only where small groups of individuals rely on
local food sources. On the other hand, in many modern food supply

8 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

systems, industrial-scale processing and distribution operations mean that

foodstuffs often travel large distances before consumption and are drawn
from far-flung sources. In the UK, as in most developed countries, therefore,
the average distance travelled by foodstuffs before consumption has
increased markedly, from an average of about 82 km in 1978 to 346 km
in 199820. In the light of these changes, several attempts have been made
in recent years to calculate the distance travelled by ingredients in com-
mon food products or meals (so called ‘food-miles’). In Iowa, USA, for
example, ingredients for a standard meal of stir-fry and salad were esti-
mated to have been transported 20,000 km21; in the UK, Sustain, a pres-

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

sure group on food and agriculture, estimated that ingredients for a
traditional turkey dinner had been transported some 38,620 km22! Apart
from implications for increased energy consumption and environmental
pollution, such extended distribution networks clearly mean that it can
be difficult to track and control potential contamination between source
and consumption.
By whatever pathways and processes pollutants pass through the envi-
ronment, four related factors are especially important in determining the
potential for exposure and health effects: their persistence, their mobility,
their decomposition products and their toxicity. The problems associated
with the release of persistent pollutants into the environment were high-
lighted many years ago with recognition of the global extent of contami-
nation, and wide-ranging environmental and health effects, caused by
DDT and other organochlorine pesticides23. The story is in many ways now
being repeated in relation to chlorofluorocarbons and other atmospheric
pollutants that act as greenhouse gases or scavengers of stratospheric
ozone24, and perhaps also in relation to endocrine disruptors25. Persistence,
however, is not necessarily the most important issue, for where they
persist in inert yet inaccessible forms, pollutants may pose relatively limited
risks. Thus, whereas inorganic mercury is persistent, it is less toxic and
less readily bioavailable than methyl mercury, to which it is naturally
converted through chemical reactions and the action of soil and aquatic
microorganisms26,27. Equally, many solid wastes represent little risk to
health so long as they remain in their original form. The problems in
these cases often come when decomposition occurs, either because the
decomposition products are inherently more toxic or because they are
more mobile, and thus are more likely to result in human exposure.

Exposure and dose

Whilst the potential for health impairment initially depends upon the
existence and concentrations of pollutants in the environment, for health
effects to occur exposures must take place that lead to a dose sufficient to
have adverse health consequences. Exposure in this context is defined as
the contact between a hazardous agent (in this case a pollutant) and an

British Medical Bulletin 2003;68 9

Impact of environmental pollution on health: balancing risk

organism. Dose refers to the quantity of the substance in the body. The
absorbed dose refers to the amount of the substance entering the body as
a whole; target organ dose refers to the amount reaching the specific
organs that are affected.
Exposure can take place in many different ways. Three main forms of
exposure are generally recognized: dermal contact, inhalation and ingestion.
In some cases, however, it may also be useful to recognize a fourth—
injection—for example, when pollutants are transmitted by animal bites
or by deliberate injection. In each of these cases, exposure may occur in
a range of different environments. Whilst some exposures occur in the

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

outdoor (ambient) environment, most people spend the majority of their
time indoors, either at home or at their place of work or learning. Indoor
exposures therefore often make up a major proportion of total exposure28—
although they tend to be rather neglected in many epidemiological studies.
Food and drinking water are likewise important routes of exposure for
many pollutants.
What determines levels of exposure is consequently not just the distri-
bution of pollution within the environment, but also human behaviours
and lifestyles, and thus the sorts of exposure environments in which
people spend their time. By the same token, exposure is not only an
environmental process, it is also a social, demographic and economic
one. Indeed, because of the myriad ways in which socio-economic and
demographic factors influence and interact with environmental conditions,
exposures, human susceptibility and health outcome, they may often
appear to outweigh the effects of the environment per se in associations
between pollution and health. One expression of these complex interactions
is the patterns of what is often called environmental injustice that are
seen throughout the world: namely, the tendency for environmental
pollution and poverty or other forms of disadvantage to be strongly
correlated, such that poorer people tend to live in more polluted
environments29,30. Whilst the reasons for this association are not fully
understood, and may be more subtle than often assumed, the double
jeopardy that it represents seems to be generally reflected in terms of
health inequalities as well. The problem, however, comes in trying to
separate the contributions to these adverse health outcomes from socio-
economic and environmental factors—and thus to quantify the attributable
effects of pollution.
It also has to be recognized that pollutants rarely occur in isolation;
more typically they exist in combination. Exposures are therefore not
singular. Instead we are usually exposed to mixes of pollutants, often
derived from different sources, some of which may have additive or
synergistic effects. Unravelling the effects of individual pollutants from
this mix is a challenging problem that has yet to be adequately resolved
in many areas of epidemiology.

10 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

Health effects: dose–response relationships, latency and attributable risk

One of the underlying tenets of environmental epidemiology is that, for
the health effects of interest, a relationship exists between the level of
exposure (or dose) and the degree of effect. Effects can, in fact, be repre-
sented in two different ways: by the type of effect or by its severity or the
probability of its occurrence (often termed the ‘response’). In either case,
these associations are generally assumed to be broadly linear, such that
the effect or response increases with each increment of exposure to a
pollutant (Fig. 4A). For many pollutants and many health effects, this
assumption seems to hold true at least over a wide range of exposures and

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

responses. Some, however, appear to be characterized by more complex
associations. Thresholds may exist, for example, below which no detectable
health effects occur (Fig. 4B). At high levels of exposure, responses may
weaken, so that the dose–response relationship is essentially curvilinear-
convex (Fig. 4C) or S-shaped (Fig. 4D). In a few cases, there is some
evidence that ∩- or, more rarely, U-shaped relationships may exist—for
example, in relation to solar radiation or vitamin intake. One of the
main purposes of epidemiology is to demonstrate and, if possible, quantify
these relationships, where they exist.
Just as exposures may be long- or short-term, so health effects can
be short-lived (acute) or prolonged (chronic). Health effects may also be
delayed to a greater or lesser extent after initial exposure, either because

Fig. 4 Common forms of exposure–response relationships.

British Medical Bulletin 2003;68 11

Impact of environmental pollution on health: balancing risk

it takes time for exposures to reach a critical level, or because the disease
itself takes time to develop and become apparent (latency). Many acute
effects are almost immediate, and have latencies typically of no more
than a few minutes to a few days. Many chronic effects, on the other
hand, can have latencies of several years—up to 20 years or more, for
example, in the case of some cancers and diseases such as asbestosis.
Dealing with these latencies is problematic in epidemiological studies,
both because they often imply the need for information on past (in some
cases long-past) exposures, and because the degree of latency may vary
from one individual to another, depending on factors such as the level of

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

exposure, the age at which exposure occurred and pre-existing health
status. In some cases, too, a so-called harvesting effect may occur, such
that, following a brief increase in disease rates as more vulnerable people
are affected, rates of illness fall because there are fewer vulnerable people
left (Fig. 5). This, too, can complicate epidemiological studies, for the
scale (and even the direction) of the dose–response relationship may
vary depending on the length of latency allowed for in the analysis.
Long-term legacies for health may also occur as a result of sensitization
and predisposition of individuals to the effects of exposure in early life.
Sensitization to house dust mite and other allergens both in the diet and
the indoor environment during the first few months of life, for example,
appears to increase risks of allergic airway disease later in childhood31.
Similarly, inverse associations have been found between birth weight and
the incidence of a range of diseases including hypertension, type-2 diabetes

Fig. 5 Harvesting effects.

12 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

and cardiovascular disease in adults32–34. Environmental exposures, such

as air pollution, that contribute to these predisposing conditions may thus
have long-term (and in some cases lifelong) implications for health.
Health outcomes may also be more or less specific to exposures to
particular pollutants. Very few diseases are, in practice, pollutant-specific.
Amongst these are asbestosis and mesothelioma (due to exposures to
asbestos) and bagassosis, through exposure to organic dusts (the most
common example, of which, is ‘Farmer’s lung’). Far more commonly,
individual health effects may arise as a result of exposures to a number
of different risk factors, either individually or in combination, whereas

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

individual exposures can give rise to a range of different health effects.
Environmental health is thus characterized by many-to-many relationships;
understanding these is, again, a major challenge for epidemiology. Partly
for this reason, it is often extremely difficult to assess the health burden
attributable to an individual pollutant. Over-estimation may occur due to
double-counting (or multiple attribution) of health effects; under-estimation
may arise due to the failure to recognize some contributions to the disease
burden as a result of masking by other risk factors.
In addition, of course, all epidemiological studies—and other studies
that contribute to the establishment of dose–response relationships, such
as laboratory experiments and clinical trials—are subject to error and
uncertainty. These arise for many different reasons: because of errors in
exposure assessment or classification, because of errors in diagnosis or
reporting of health outcome, because of inadequate sample size, because
of inadequate adjustment for confounding or effect modification by
other factors, because of biases in sampling and statistical analysis, and
because of the underlying indeterminacy of some of the associations of
interest. As a result, dose–response reported by different studies often
shows substantial differences, and many separate studies may be needed
before a clear pattern of association emerges. Even then, problems may
be encountered in deriving reliable dose–response relationships (e.g. through
some form of meta-analysis), because of inconsistencies in study design
(e.g. in methods of exposure classification, target populations or speci-
fication of health outcome). Most dose–response relationships are thus
accompanied by a relatively large degree of uncertainty.

Models of pollutant pathways

As the discussion above has indicated, the relationships between pollution
and health are both complex and often indirect. Considerable difficulties
are thus encountered in quantifying the associations involved. It is
largely for this reason that many of the health effects of environmental
pollution are still uncertain, and that problems arise in attempting to
attribute health outcome to environmental causes—for example, when
trying to confirm or explain apparent spatial clusters in health.

British Medical Bulletin 2003;68 13

Impact of environmental pollution on health: balancing risk

These subtleties and complexities highlight the importance of examining

critically any hypothesis about a relationship between a pollutant and an
apparent health effect, and of setting such hypotheses within a wider envir-
onmental context. Assumptions about simple, singular cause–effect rela-
tionships often need to be eschewed; in their place we need to recognize
the possibility for multifactorial effects in which single health outcomes
are attributable to a wide variety of (possibly inter-related) environmental
and other risk factors; and in which individual exposures may contribute
to a range of different health effects. The contingent and historical nature
of many of these associations also needs to be appreciated: health effects

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

seen now in many cases owe their existence to exposures, sensitization or
some process of predisposition far in the past. Because environmental
conditions, and even the very nature of the risk factors involved, may
change quite considerably over time, uniformitarianist principles may not
hold true, i.e. the present cannot always be seen as the key to the past.
Against this background, the use of models to conceptualize the possible
interplay of different risk factors and exposure pathways, and how they
might have evolved over time, represents an important tool for attempts
to understand associations between pollution and health. One example
is illustrated in Figure 6, which shows possible sources and pathways of
exposure of environmental pollution associated with landfill sites. Several
important lessons can be drawn from this example. First, it is evident
that the pathways of exposure are highly varied and complex. Which is
the most important may well differ from one situation to another. The
possibility of contributions from each and all of them needs to be
allowed. Second, it is evident that landfill sites leave a legacy which may
persist long after they are no longer operational. Present-day land use
and activity may therefore not account for current exposures. Third,
related to this, sources and pathways of exposure change markedly over
time—and, indeed, many of the risks associated with what are now
landfill sites may well predate the sites themselves (e.g. from prior land
use). Perhaps it is for this reason that several studies of health risks
around landfill sites have found that raised levels of risk existed before
the landfill sites were opened35,36.

The contribution of environmental pollution to the global

burden of disease
Estimating the global burden of disease
For all the reasons outlined above, estimating the contribution of environ-
mental pollution to the burden of disease is far from easy. In general,
too little is known either about the causal links between environmental

14 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

Fig. 6 A model of emission processes and exposure pathways from landfill sites.

pollution and health, or about the levels of exposure across the population,
to make reliable assessments of the proportion of disease or mortality
attributable to pollution. These difficulties are severe in developed coun-
tries, where disease surveillance, reporting of mortality, environmental
monitoring and population data are all relatively well established. In most
developing countries they become all but insurmountable, because of the
generally impoverished state of routine monitoring and reporting. Given
that controls on emissions and exposures in the developing world are often
limited, it is in these countries that risks from environmental pollution are
likely to be greatest. Such uncertainties thus render any attempt to quantify
the environmental burden of disease highly approximate at best.
Assessments of the disease burden attributable to different forms and
sources of pollution are nevertheless worth the effort. They are needed,
for example, to raise awareness about some of the risks associated with
environmental pollution, and as a basis for advocacy—to ensure that
those most in need have a voice. They are needed to help motivate and
prioritize action to protect human health, and to evaluate and monitor
the success of interventions. They provide the foundation, therefore, for
extremely powerful indicators for policy support, and a means of pricking
the global conscience about inequalities in health.

British Medical Bulletin 2003;68 15

Impact of environmental pollution on health: balancing risk

Over recent years, therefore, many attempts have been made to assess
the health status of the population, both nationally and globally, and to
deduce the contribution made by pollution and other environmental
factors. In Europe, for example, more than 50 national environmental
health action plans have been developed, following the Helsinki Conference
in June 1994, setting out strategies to tackle problems of environmental
health37. Although these differ substantially in terms of their content and
scope, many have involved attempts to make formal assessments of the
disease burden attributable to different environmental hazards, and to
rank these in terms of their public health significance38,39. Various methods

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

were used for this purpose, though most relied on some form of expert
judgement, informed where available by quantitative data on mortality
or disease rates. Whatever the weaknesses of these assessments, their
practical importance is evident, for they have contributed directly to policy
prioritization and development in the countries concerned.
The same need has arisen to support the development of environmental
health indicators. Since the early 1990s, largely motivated by WHO,
increasing attention has been given to constructing indicators on envi-
ronmental health at all levels from the local to the global scale40–43, and
a number of indicator sets have been created (and to a lesser extent
used)44–46. Environmental pollution is, inevitably, a major focus of concern
in these indicator sets. By definition, also, environmental health indicators
provide measures that link environmental hazards and health effect41.
As such they depend upon an understanding of the association between
pollution and health, either in the form of what have been called ‘exposure-
side indicators’, which use information on exposures to imply degrees of
health risk, or ‘health-side indicators’, which use information on health
outcome to suggest attributable effects41,43. In both contexts, the ability
to make at least semi-quantitative interpretations of the link between
pollution and health, and thus to assess the contribution to the burden
of disease, is assumed.

Mortality
The most explicit attempts to quantify these links, however, have come
in recent years through work to estimate the global (and to some extent
regional) burden of disease. Earlier efforts in this direction were targeted
specifically at making broad-scale enumerations of the total disease burden
across the world47,48. The traditional measure used for such assessments
was mortality, both because data on deaths tended to be more reliable
and widely available, and because mortality is directly comparable in
terms of health outcome, unlike morbidity which implies differences
in severity of effect. Even so, results from the various efforts differed
somewhat, largely because of the ways in which gaps and uncertainties
in the available data were dealt with49. Overall, however, cardiovascular

16 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

diseases were seen to be the major cause of mortality, accounting for

between 19% (based on the World Health Survey estimate) and 28%
(based on Murray and Lopez’s Global burden of disease study44) of total
deaths worldwide. Cancer (an estimated 12% in each case), acute respi-
ratory diseases (8.1, 8.7%, respectively), unintentional injuries (5.7, 6.4%),
diarrhoeal diseases (ca. 5.8%), chronic respiratory diseases (ca. 5.7%)
and perinatal conditions (6.2, 4.8%) were other major killers.

Years of life lost and disability adjusted life years

Crude estimates of the number and proportion of deaths due to different

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

diseases, of this nature, obviously give only a distorted picture of the
true burden of disease, for they take no account of the age of death or
the duration of any preceding illness and disability, nor the amount of
suffering involved. In an attempt to redress this, Murray and Lopez48
also computed estimates of the ‘years of life lost’ (YLL) and ‘disability
adjusted life years’ (DALYs). Years of life lost are estimated as the dif-
ference between age at death and the life expectancy in the absence of
the disease, based on an advanced developed country (82.5 years for
women and 80 years for men at that time). DALYs also incorporate an
allowance for the number of years lived with a disability due to disease
or injury, weighted according to its severity (based on expert assessments
of the relative impact of some 500 different conditions and disease
sequelae). The years of disability or life lost are also discounted according
to the age of onset (since it is assumed that future years of life lost
contribute less to the burden of disease than current ones).
Results of these calculations are summarized and discussed by
WHO49. Estimates of YYL and DALYs provide a somewhat different
ranking of disease compared to crude mortality, since they give additional
weight to early-onset diseases and chronic illness. Cardiovascular diseases
are thus seen as somewhat less important (making up ca. 13% of YYL
and 9.7% of DALYs). Acute respiratory diseases (12% and 8.5%,
respectively), diarrhoeal diseases (10% and 7.2%) and unintentional
injuries (9.3% and 11%) all become proportionally more significant.
Mental health conditions also figure as a major source of ill health in terms
of DALYs, contributing 11% of the total burden of disease worldwide.

Variations in the global burden of disease

By whichever method they are computed, marked variations are evident
in the burden of disease between different sectors of the population.
Children are seen to be especially at risk—and young children most of all.
More than 30% of all deaths for all diseases in the Global burden of disease
study occurred to children under 15 years of age; in the case of diarrhoeal
diseases they accounted for 88% of deaths, and for acute respiratory
illness 67% of deaths. Malaria also struck children disproportionately

British Medical Bulletin 2003;68 17

Impact of environmental pollution on health: balancing risk

(82% of deaths), whilst mortality as a result of perinatal diseases and

vaccine-preventable diseases was inevitably almost wholly of children.
When measured in terms of DALYs, the overwhelming burden of all
these diseases falls on children49.
Similar inequalities occur both socially and geographically. The World
Bank, for example, compared mortality rates and DALYs between poor
and rich nations in the world50. Clear differences were shown. Whereas
ischaemic heart disease, for example, was responsible for 23.4% of
deaths in the rich countries, it accounted for only 7.3% in the poor
countries; malignant neoplasms were responsible for 22.6% of deaths

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

among rich countries, but only 5.6% among the poor. Conversely, respi-
ratory infections and diarrhoeal diseases accounted for 13.4% and
11.3% of deaths, respectively in poor nations, compared with 4.0% and
0.3% in rich countries; for childhood cluster diseases, the proportions
were 7.8% and 0.1%, respectively. Similar disparities have been shown
in a comparison of ‘less developed’ and ‘more developed’ countries by
Smith et al51. As these examples show, generalizations about the burden
of disease thus need to be interpreted with care. Beneath the often stark
global figures lie even starker indications of health inequalities that cry
out to be addressed.

The environmental burden of disease

Whilst the original estimates of the global burden of disease made by
Murray and Lopez and WHO during the mid 1990s were a major step
forward in terms of providing comparable data on health status across
the world, they gave information only on health outcomes and did not
for the most part attempt to attribute these outcomes to specific causes.
Smith et al51 did, however, make an attempt to assess the environmental
contribution to the global burden of disease, using Murray and Lopez’s
data. This suggested that environmental factors accounted for between
25% and 33% of the total burden of disease, but with a disproportionate
share of this falling on children under 5 years of age. Diarrhoeal diseases
(for which some 90% of DALYs were attributed to the environment),
malaria (ca. 88%) and acute respiratory illness (60%) were seen as out-
comes for which the environment was especially influential. Murray and
Lopez52 also made preliminary assessments of the relative importance of
different risk factors for the global burden of disease, based on their
1990 data. Malnutrition stood out as the most important factor considered,
accounting for ca. 11.7% of deaths and 15.9% of DALYs worldwide.
Poor water and sanitation was estimated to be responsible for ca. 5.3%
of deaths and 6.8% of DALYs, whereas air pollution contributed 1.1%
and 0.5%, respectively. Subsequently, a more specific attempt was made

18 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

by Prüss et al53 to assess the effects of water, sanitation and hygiene,

using a combination of exposure-based risk assessment and out-
come-based disease attribution54. Based on available information,
they estimated that these environmental factors were responsible for
ca. 4% of global mortality and 5.7% of total DALYs. These esti-
mates are somewhat lower than those implied by the original Global
burden of disease study52, partly perhaps because of differences in
methodology and partly because of a decline in mortality in the
intervening years.
All these attempts to partition the global burden of disease by causa-

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

tive risk factor have faced, and admitted, a number of major difficulties.
These relate not only to uncertainties in the available data on health
outcome, but also to problems of how to attribute any single death to a
single cause or risk factor. Two main approaches have been proposed
for disease attribution55. Categorical attribution assigns each death to a
specific disease or risk factor, according to a defined set of rules (e.g. the
ICD system). The advantage of this approach is that it is relatively
straightforward and consistent, and avoids double-counting; the dis-
advantage is that it ignores the multi-factorial nature of many diseases
and still leaves unresolved the problem of how to define appropriate
rules. Counterfactual attribution involves comparing the current level of
disease or mortality with that which might be expected to occur in the
absence of the risk factor (or at some other reference level). One of the
main difficulties with this approach is how to define this reference level.
Several possibilities exist: for example, the complete absence of the risk
factor, the level of risk in some reference population or area, or the
achievable level of risk with current technologies. Each will tend to give
a rather different measure of the attributable burden of disease. In this
context, another difficulty also arises, i.e. how to assess the likely change
in disease burden under the selected scenario, in the absence of empirical
data.
Notwithstanding these difficulties, a revised assessment of the global
burden of disease has recently been carried out, involving explicit attempts
at attribution by risk factor or hazard56. A counterfactual approach was
used, with the reference level for each disease being defined as that
which would occur under conditions of a minimum theoretical exposure
distribution (i.e. that which would achieve the lowest population risk,
irrespective of whether this is achievable in practice). Assessments were
carried out by a series of expert groups, who first undertook a detailed
review of relevant literature, and derived estimates of the exposures and
relative risks for specific age and gender groups, for each of 14 sub-
regions. Based on these data, estimates were then made of the population
impact fraction of the disease or death in each region, for each risk
factor:

British Medical Bulletin 2003;68 19

Impact of environmental pollution on health: balancing risk

m m

∫ RR ( x )P ( x ) – ∫ RR ( x )P′ ( x )
x=0 x=0
PIF = -------------------------------------------------------------------------------------------
-
m

∫ RR ( x )P ( x )
x=0

where PIF is the population impact fraction, RR(x) is the relative risk at
exposure level x, P(x) is the population distribution of exposure, P′(x) is
the counterfactual distribution of exposure and m is the maximum

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

exposure level.
Results from this assessment for a number of environmental risk factors
are summarized in Table 1. More detail on the methods and results are
available on the WHO website for a number of the disease groups,
including chronic obstructive pulmonary disease, malnutrition and
injuries57, and others will be published as available. The sources of envi-
ronmental and occupational pollution listed in Table 1 account for 8–9%
of the total global burden of disease, measured either in terms of mortality
or DALYs. Amongst these risk factors, water, sanitation and hygiene
and indoor air pollution are seen to be the most important; health effects
of outdoor air pollution are comparatively small, although to some
extent this may reflect differences in methodology between this and the
other expert groups. It is also evident that a range of other sources of
pollution, not included in this table, might be implicated in the global
burden of disease, such as exposures to ionizing and non-ionizing radiation,
food contamination, pesticides, household hazardous chemicals, wastes
and other forms of indoor air pollution. The overall burden of disease
attributable to pollution, therefore, cannot yet be assessed.
As with previous estimates of the burden of disease, marked variations
can also be recognized between different parts of the world. As is to be
Table 1 Global burden of disease (thousand and percent) attributable to selected sources
of environmental and occupational pollution

Risk factor Deaths DALYs

Number % Number %

Total (all risk factors) 55,861 1,455,473

Water, sanitation and hygiene 1730 3.1 54,158 3.7
Urban outdoor air pollution 799 1.4 6404 0.4
Indoor smoke from solid fuels 1619 2.9 38,539 2.6
Lead 234 0.4 12,926 0.9
Occupational carcinogens 118 0.2 1183 0.1
Occupational airborne 356 0.6 5354 0.4
particulates
Occupational noise 0 0.0 4151 0.3
Total (pollution-related) 4856 8.7 122,715 8.4

Source: based on Ezzati et al 56.

20 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

expected, the developing countries bear the major proportion of the burden.
Problems of unsafe water, sanitation and hygiene, for example, account
for an estimated 6.6% of DALYs in Africa, and 4.7% in south-east Asia,
compared with 0.5% in Europe. Indoor air pollution accounts for 4.4%
of DALYs in Africa and 3.6% in south-east Asia, compared to 0.4% in
Europe. In absolute terms the differences are even more stark. The total
number of DALYs per head of population attributable to these two risk
factors in Africa are 29.1 per thousand for unsafe water, sanitation and
hygiene and 19.3 per thousand for indoor air pollution; in south-east
Asia they are 12.8 and 9.9 per thousand, respectively; in Europe they are

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

0.8 and 0.6 per thousand, respectively. Risks attributable to environ-
mental pollution in the developing world are thus 15–35 or more times
greater than in developed countries.

Conclusions
The complexities involved in the link between environmental pollution and
health, and the uncertainties inherent in the available data on mortality
and morbidity, in existing knowledge about the aetiology of diseases,
and in environmental information and estimates of exposure, all mean
that any attempt to assess the environmental contribution to the global
burden of disease is fraught with difficulties. The estimates produced to
date must therefore be regarded as no more than order-of-magnitude
estimates. Despite these limitations, however, several conclusions seem
beyond refute.
The first is that environmental pollution plays a significant role in a
number of health outcomes, and in several cases this adds up to a serious
public health concern. Water pollution, sanitation and hygiene, indoor
air pollution, and to a lesser extent outdoor air pollution and exposures
to chemicals in both the indoor and outdoor environment are all important
risk factors in this respect. Ionizing and non-ionizing radiation and noise
are also causes for concern in many cases.
Secondly, it is clear that the distribution of risks from these factors is
not equal across the world. The global burden of disease may be difficult
to quantify, but stark contrasts in that burden are evident between the
developed and the developing world, between rich and poor, and often
between children and adults. The developed world is not risk-free, and
development is no panacea for all environmental health ills. On occasions,
in fact, the opposite is true: developments, such as increased reliance on
road transport, increased use of chemicals in agriculture, and increased
proportions of time spent in modern, hermitically sealed buildings
surrounded by chemically-based fabrics and furnishings may actually
increase exposures and exacerbate health risks. But overall the developing

British Medical Bulletin 2003;68 21

Impact of environmental pollution on health: balancing risk

world is far more severely affected by pollution, and in many instances

becoming more so, as pressures from development add to traditional
sources of exposure and risk.
Thirdly, and perhaps most importantly, many of these risks and health
effects are readily avoidable. Rarely does the solution lie in advanced
technologies or even expensive drugs. Instead, the need is for preventive
action to reduce the emission of pollutants into the environment in the
first place—and that is largely achievable with existing know-how.
Indeed, in many cases it has already been implemented in many of the
richer countries. Science, therefore, certainly has a role to play in

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

addressing these issues. More research is undoubtedly needed on a range
of emerging environmental health issues. But the deficit of action that
has allowed environmental pollution still to take its toll on health
derives not so much from failures in science or technology as from the
lack of political will and economic empowerment. It is from that direction
that salvation needs ultimately to come for those at the mercy of envir-
onmental pollution.

References

1 Samet JM, Dearry A, Eggleston PA et al. Urban air pollution and health inequities: A workshop
report. Environ Health Perspect 2001; 109 (Suppl. 3): 357–74
2 Sexton K, Adgate JL. Looking at environmental justice from an environmental health perspective,
J Expos Anal Environ Epidemiol 2000; 9: 3–8
3 Taubes G. Epidemiology faces its limits. Science 1995; 269: 164–9
4 Royal Commission on Environmental Pollution. Chemicals in products: safeguarding the envi-
ronment and human health. 24th Report of the Royal Commission on Environmental Pollution.
London: Royal Commission on Environmental Pollution, 2003
5 Lindley SJ, Longhurst JWS, Watson AFR, Conlan DE. Procedures for the estimation of
regional scale atmospheric emissions—an example from the NW region of England. Atmos
Environ 1996; 30: 3079–91
6 Colvile R, Briggs DJ. Dispersion modelling. In: Elliott P, Wakefield JC, Best NG, Briggs DJ
(eds) Spatial Epidemiology. Methods and Applications. Oxford: Oxford University Press,
2000; 375–92
7 National Atmospheric Emissions Inventory http://www.naei.org.uk/emissions/emissions_2000.
php?action=data (date last accessed 29 June 2003)
8 Jarvis KE, Parry SJ, Piper JM. Temporal and spatial studies of autocatalyst-derived platinum,
rhodium, palladium and selected vehicle-derived trace element in the environment. Environ Sci
Technol 2001; 35: 1031–6
9 Björklund S. Emissioner av partiklar från dieselfordon och vedförbränning. In: Partiklar och
hälsa-ett angeläget problem att undersöka. Skandias miljökommision, rapport number 5.
Försäkringsaktiebolaget Skandia, 1996, S-103 50
10 Environment Agency http://www.environment-agency.gov.uk/yourenv/eff/pollution/296030/
296054/?lang=_e&region= (date last accessed 29 June 2003)
11 Rushton L. Health hazards and waste management. Br Med Bull 2003; 68: 183–198
12 Department of Food, Environment and Rural Affairs. The Government’s Strategic Review of
Diffuse Water Pollution from Agriculture in England. Agriculture and Water: A Diffuse Pollution
Review. London: DEFRA, 2001
13 Beven K, Germann P. Macropores and water flow in soils. Water Resour Res 1982; 18: 1311–25

22 British Medical Bulletin 2003;68

 Environmental pollution and the global burden of disease

14 ATSDR. Landfill Gas Primer: An overview for Environmental Health Professionals. Washington:
Agency for Toxic Substances and Disease Registry, 2001
15 WHO. Air Quality Guidelines for Europe, 2nd edn. Copenhagen: WHO Regional Office for
Europe, 2001
16 Henshaw DL. Radon exposure in the home: its occurrence and possible health effects.
Contemp Phys 1993; 34: 31–48
17 Fawell J, Nieuwenhuijsen MJ. Contaminants in drinking water. Br Med Bull 2003; 68: 199–208
18 Fisher A, Matthews L. The social behaviour of sheep. In: Keeling LJ, Gonyou HW (eds) Social
Behaviour in Farm Animals. Wallingford: CAB International, 2001
19 Beresford NA, Barnett CL, Crout NMJ, Morris CC. Radiocaesium variability within sheep
flocks: relationships between the 137Cs activity concentrations of individual ewes within a
flock and between ewes and their progeny. Sci Total Environ 1996; 177: 85–96
20 UK Sustainable Development Commission. State of Sustainable Development in the UK.

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

Preparatory Paper. London: UK Sustainable Development Commission, 2001
21 Pirog R, van Pelt T, Enshayan K, Cook E. Food, Fuel and Freeways. An Iowa Perspective on
how Far Food Travels, Fuel Usage and Greenhouse Gas Emissions. Ames, IA: Leopold Centre,
Iowa State University, 2001
22 Jones A. Eating Oil: Food Supply in a Changing Climate. London: Sustain and Elm Farm
Research Centre, 2001
23 Bro-Rasmussen F. Contamination by persistent chemicals in food chain and human health. Sci
Total Environ 1996; 188: S45–S60
24 McFarland M, Kaye J. Chlorofluorocarbons and ozone. Photochem Photobiol 1992; 55: 911–29
25 Joffe M. Infertility and environmental pollutants. Br Med Bull 2003; 68: 47–70
26 WHO. Mercury, Inorganic. Environmental Health Criteria, vol. 118. Geneva: World Health
Organization, 1991
27 WHO. Methylmercury. Environmental Health Criteria, vol. 101. Geneva: World Health
Organization, 1990
28 Zhang J, Smith KR. Indoor air pollution: A global health concern. Br Med Bull 2003; 68:
209–225
29 Morello-Frosch R, Paster M, Porras C, Sadd J. Environmental justice and regional inequality in
Southern California: Implications for future research. Environ Health Perspect 2002; 110: 149–53
30 Waller LA, Louis TA, Carlin BP. Environmental justice and statistical summaries of differences
in exposure distributions. J Expos Anal Environ Epidemiol 1999; 9: 56–65
31 Halken S. Early sensitisation and development of allergic airway disease—risk factors and
predictors. Paediatr Respir Rev 2003; 4: 128–34
32 Leon DA, Lithell HO, Vagero D, Koupilova I, Mohsen R, Berglund L, Lithell UB, McKeigue PM.
Reduced fetal growth rate and increased risk of death from ischaemic heart disease: cohort
study of 15,000 Swedish men and women born 1915–29. BMJ 1998; 317: 241–5
33 Rich-Edwards JW, Stampfer MJ, Manson JE, Rosner B, Hankinson SE, Colditz GA, Willett WC,
Hennekens CH. Birth weight and risk of cardiovascular disease in a cohort of women followed
up since 1976. BMJ 1997; 315: 396–400
34 Suzuki T, Minami J, Ohrui M, Ishimitsu T, Matsuoka H. Relationship between birth weight
and cardiovascular risk factors in Japanese young adults. Am J Hypertens 2000: 13: 907–13
35 Elliott P, Briggs DJ, Morris S, de Hoogh C, Hurt C, Kold Jensen T, Maitland I, Richardson S,
Wakefield J, Jarup L. Risk of adverse birth outcomes in populations living near landfill sites.
BMJ 2001; 323: 363–8
36 Fielder HM, Poon-King CM, Palmer SR, Moss N, Coleman G. Assessment of impact on health
of residents living near the Nant-y-Gwyddon landfill site: retrospective analysis. BMJ 2000;
320: 19–22
37 Kello D, Haralanova M, Stern RM, Briggs DJ. National environmental health action plans:
background and process. In: Briggs DJ, Stern RM, Tinker TL (eds) Environmental Health for
All. Risk Assessment and Risk Communication for National Environmental Health Action
Plans. Dordrecht: Kluwer Academic Publishers, 1999; 3–15
38 Cizcova H, Kazmarova H, Dumitrescu A, Janikowski R. The NEHAP experience in the Czech
Republic, Romania and Poland. In: Briggs DJ, Stern RM, Tinker TL (eds) Environmental
Health for All. Risk Assessment and Risk Communication for National Environmental Health
Action Plans. Dordrecht: Kluwer Academic Publishers, 1999; 17–34

British Medical Bulletin 2003;68 23

Impact of environmental pollution on health: balancing risk

39 Victorin K, Hogstedt C, Kyrklund T, Eriksson M. Setting priorities for environmental health

risks in Sweden. In: Briggs DJ, Stern RM, Tinker TL (eds) Environmental Health for All. Risk
Assessment and Risk Communication for National Environmental Health Action Plans.
Dordrecht: Kluwer Academic Publishers, 1999; 35–51
40 Corvalán CF, Kjellström T, Smith KR. Health, environment and sustainable development.
Identifying links and indicators to promote action. Epidemiology 1999; 10: 656–60
41 Corvalán C, Briggs D, Zielhuis G (eds) Decision-making in Environmental Health: From
Evidence to Action. London: E & F.N. Spon, 2000
42 Kjellström T, Corvalán C. Framework for the development of environmental health indicators.
World Health Stat Q 1995; 48: 144–54
43 Briggs DJ, Wills J. Presenting decision-makers with their choices: environmental health indicators
for NEHAPs. In: Briggs DJ, Stern RM, Tinker TL (eds) Environmental Health for All. Risk
Assessment and Risk Communication for National Environmental Health Action Plans.

Downloaded from http://bmb.oxfordjournals.org/ at Ondokuz MayisUniversity on May 22, 2014

Dordrecht: Kluwer Academic Publishers, 1999; 187–201
44 WHO. Environmental Health Indicators: Framework and Methodology. 1999, http://
www.who.int/environmental_information/Information_resources/on_line_general.htm (date last
accessed 30 June 2003)
45 von Shirnding H. Health in Sustainable Development Planning: The Role of Indicators.
Geneva: WHO, 2003
46 Briggs DJ. Making A Difference: Indicators For Children’s Environmental Health. Geneva:
WHO; In press
47 WHO. World Health Report. Geneva: WHO, 1995
48 Murray CJL, Lopez AD (eds) The Global Burden of Disease. Cambridge: Harvard School of
Public Health (on behalf of the World Health Organization and World Bank), 1996
49 WHO. Health and Environment in Sustainable Development. Five Years after the Earth
Summit. Geneva: WHO, 1997
50 Gwatkin DR, Guillot M. The Burden of Disease Among the Global Poor. Current Situation,
Future Trends, and Implications for Strategy. Washington, DC, USA: The World Bank, 1999
51 Smith KR, Corvalán C, Kjellström T. How much global ill health is attributable to environ-
mental factors? Epidemiology 1999; 10: 573–84
52 Murray CJL, Lopez AD. On the comparable quantification of health risks: lessons from the
global burden of disease study. Epidemiology 1999; 10: 594–605
53 Prüss A, Kay D, Fetrell L, Bartram J. Estimating the burden of disease from water, sanitation
and hygiene at a global level. Environ Health Perspect 2002; 110: 537–42
54 Prüss A, Corvalán C, Pastides H, de Hollander AEM. Methodologic considerations in
estimating burden of disease from environmental risk factors at national and global levels. Int
J Occup Environ Health 2002; 7: 58–67
55 Mathers CD, Ezzati M, Lopez AD, Murray CJI, Rodgers AD. Causal decomposition of
summary measures of population health. In: Murray CJL, Saloman J, Mathers CD, Lopez AD
(eds) Summary Measures of Population Health. Geneva: WHO, 2002
56 Ezzati M, Lopez AD, Rodgers A, Vander Hoorn S, Murray CJL, the Comparative Risk Assessment
Collaborating Group. Selected major risk factors and global and regional burden of disease.
Lancet 2002; 360: 1347–60
57 WHO. http://www.who.int/peh/burden/globalestim.htm (date last accessed 30 June 2003)

24 British Medical Bulletin 2003;68