ORIGINAL ARTICLES How Much Global Ill Health Is Attributable to Environmental Factors? Kirk R. Smith,1 Carlos F. Corvalán2, and Tord Kjellström3 Over the years, estimates have been made of the portions of attributed to environmental risk factors. Children under 5 human mortality and morbidity that can be attributed to years of age seem to bear the largest environmental burden, environmental factors. Frustratingly, however, even for a and the portion of disease due to environmental risks seems single category of disease such as cancer, these estimates to decrease with economic development. A summary of have often varied widely. Here we attempt to explain why these estimates first appeared in the 1997 report, "Health such efforts have come to such different results in the past and Environment in Sustainable Development," which was and to provide guidance for doing such estimates more the World Health Organization's contribution to the 5-year consistently in the future to avoid the most important pitfalls. anniversary of the Rio Earth Summit. A full explanation of We do so by carefully defining what we mean by the terms how these estimates were made is first presented here. We "environmental," "ill health," and "attributable." Finally, end with a call for a program of "strategic epidemiology," based on these recommendations, we attempt our own which would be designed to fill important gaps in the estimate, appropriately qualified according to the many understanding of major environmental health risks in remaining uncertainties. Our estimate is that 25-33% of the important population groups worldwide. (Epidemiology global burden of disease can be 1999;10:573-584) Keywords: environmental exposures, morbidity, mortality, children, environmental health risks, populations, global factors, public health. The question posed in the title is difficult to answer to a What is "Environment" in a Health Context? large extent because of a lack of basic knowledge about The strict medical definition of environmental causes of disease etiology. This is likely to be resolved only by diseases would be all those that are not genetic. This is the additional laboratory, clinical, environmental, and epi- classic dichotomy between "nature" and "nurture," in which demiological research. It is also a moving target, in that environmental factors include all those that affect the some environmental changes and their health impacts have organism after conception regardless of whether they are long latency periods. There is a class of difficulties mediated by social conditions and individual choice or associated with answering this question, however, that can through environmental media. It could be argued even be usefully addressed with existing knowledge by careful further, however, that genetic factors are actually also exposition of the concepts involved and the assumptions that environmental, but merely on a different time scale. Thus, are needed. Here we address these conceptual issues by mutation, natural selection, and other mechanisms of carefully and coherently defining exactly what we mean by evolution have changed the genetic composition of humanity the terms "environment," "ill health," and "attributable." By according to environmental conditions existing in the past. In doing so, many of the difficulties in answering the overall this context, ie, in which current genes are seen as the question are reduced or eliminated. In addition, the outcome of previous environments, all diseases are entirely remaining knowledge uncertainties become more carefully environmental.1 delineated, thus helping focus further research. Neither of these definitions (“diseases are entirely environmental" or "all nongenetic causes are environ- mental") is sufficiently useful for most purposes, and both fly somewhat in the face of common everyday understanding of From the 1Center for Occupational and Environmental Health, University of what constitutes an environmental factor. In terms of health, 2 California, Berkeley, CA; Department of Protection of the Human Environment, 3 the most important difference from common perception of the World Health Organization, Geneva, Switzerland; and Department of 2 Community Health, University of Auckland, New Zealand. term is the inclusion of "behavioral" or "life-style" factors. In particular, diet, including alcohol, and smoking are Address correspondence to: Kirk R. Smith, Professor of Environmental Health extremely significant risk factors for a range of important Sciences, University of California, Warren Hall, MC-7360, Berkeley, CA 94720. diseases and thus for total health status in many parts of the The views expressed in this article are those of the authors and do not world. Yet, their inclusion as environmental risk factors necessarily reflect the position of the World Health Organization. would tend to ©1999 by Epidemiology Resources Inc. 573 574 Smith et al Epidemiology September 1999, Vol. 10 No. 5 TABLE 1. Traditional "Natural" Hazards in the Environment* Another way to define the scope of Constant search for sufficient food and water while avoiding natural toxins environmental health is in terms of Infections and parasites that spread from person to person or animal to person through measurable stressors that penetrate air, food, water, or insects from the outer environment across the Dust, damp, woodsmoke, pollen, and other airborne hazards body barriers to affect bodily health. Injuries from falls, fires, and animal attacks Heat, cold, rain, snow, wind, natural disasters, and other adverse conditions The usual pathways mentioned are in- gestion, inhalation, and dermal/placen- * Expanded from Ref. 4. tal absorption/penetration.' Clearly the eyes and ears ought to be added to this conventional list, however. Ultraviolet overwhelm the other, more conventionally understood 3 light, laser light, and noise, for example, are environmental factors. Furthermore, the most effective conventionally measured stressors that can cause ill interventions for mitigation tend to be somewhat differ- health without involving the lungs, gastrointestinal tract, ent from those for more conventional environmental or skin. The point of ambiguity is reached, however, factors, such as pollution. when it is not toxic stressors in the usual chemical or There at least two other major risk categories that physical sense, but rather toxic information content that introduce ambiguity. Some of the important social risk reaches the sense organs to cause ill health, for example, factors for disease, including crime, stress, and war, being in the form of verbal child abuse, racism, or television external to the body, are sometimes classified as envi- violence. ronmental. In addition, natural hazards, such as earth- Some economists take the approach that environmental quakes and inclement weather, although clearly "envi- issues are those in which the effects are externalities, ie, ronmenta" in some contexts, are sometimes not that they occur to a group other than those making the considered environmental because their frequency and decisions about the activities that cause them. When this scale are not generally affected by human actions. Thus, happens, it is argued, there is justification for social there is sometimes an unstated presumption that envi- (government) action to either rearrange the marketplace ronmental health deals only with those aspects of the such that the externalities are internalized (for example, environment that are affected measurably by human pollution taxes) or to regulate. Thus, urban outdoor air activities and not those due to nature in the raw. Indeed, pollution, which is clearly an externality, would be the term "natural" has come to imply clean and safe to included, but household indoor air pollution caused by many people. This view could only develop in recent sources within the household would not. There are severe history in rich countries, however, because most of hu- problems with this argument in practice, however. For manity has spent most of history protecting itself from a example, it assumes perfect information, ie, that range of far-from-benign natural environmental condi- householders understand the risks, whereas individuals .4 tions, some of which are listed in Table 1 are clearly not able to detect even rather large Figure 1 shows one way to represent the relation environmental risks on their own (or discriminate among these major categories of risk factors. It may not between minuscule and significant risks). It also assumes be worthwhile to attempt to be too precise, because the that the household is a rational decisionmaking unit, demarcations between categories are not sharp. For ex, when it is clear in many societies that there are significant ample, even diet, alcohol, and smoking, while certainly differences in power and risk among household members. being functions of individual behavior, are also functions Finally, even if making decisions rationally and of the social environment. understanding the risks, the capability of households to enact solutions individually may be sharply limited. (How, for example, can individual households change national oil policy to make cleaner household fuels available?) Perhaps most importantly, dismissing household u, any other risk factors because in some theoretical sense they are not strict externalities flies in the face of public health tradition. This tradition is to find and reduce all threats to health no matter where they occur. Thus, for example, all