Article Text


Urban air pollution, health, and equity
  1. J M Samet,
  2. R H White
  1. Johns Hopkins Bloomberg School of Public Health, Department of Epidemiology, the Risk Sciences and Public Policy Institute, Baltimore, USA
  1. Correspondence to:
 Professor J M. Samet
 Johns Hopkins Bloomberg School of Public Health, 615 N Wolfe Baltimore, MD, 21205, USA;

Statistics from

Air pollution and equity

Over the past decade, an ever increasing number of epidemiological studies have linked urban air pollution, particularly particulate matter, to increased risk for morbidity and mortality.1,2 These new findings have led to revised air pollution standards for the United States and they will probably have similar consequences in other countries around the world. This new evidence on adverse health effects of air pollution has also motivated research to identify those groups within the population who may be at increased risk from exposure, for example: infants, persons with chronic heart and lung disease, and the elderly population.3 This issue of the journal includes three papers that address socioeconomic status and vulnerability to air pollution.

This is not a new topic for scientific investigation or for public health concern. The environmental justice movement began more than two decades ago in the United States, originally related to the locating of toxic waste landfills in minority communities.4 More recently, urban air pollution has surfaced as a significant international environmental justice concern because of the large concentration of minority and low income residents living in urban environments with unhealthful air quality.5 These persons often have unhealthy housing and significant exposures to indoor air pollution as well.

Adding to the public health concern regarding the disproportionate exposure of minority and low income populations to high levels of urban air pollution is the recognition that these groups often have higher prevalence rates of diseases such as asthma that are adversely affected by air pollution. Recognition of this disparity in exposures to environmental contaminants, and the need to tackle the potential public health consequences of these disproportionate exposures, was embodied in US national policy through a 1994 Presidential Executive Order6 and in Europe in 2001 through the World Health Organisation.7

Understanding the role of socioeconomic status as a component of susceptibility to the adverse health effects of air pollution is essential to the process of setting ambient air quality standards and implementing programmes to achieve these standards. In the United States, ambient air quality standards are required to be set under the Clean Air Act at a level sufficient to protect the health of “sensitive groups.” Internationally, the WHO Regional Office for Europe has developed air quality guidelines that explicitly recognise the need to consider that subpopulations may be at considerably increased risk of suffering adverse health effects8 and therefore must be taken into account in the risk management process. Historically, sensitive groups have been identified on the basis of preexisting health status (for example, people with asthma), physiological development (for example, children), or level of response to pollution (for example, ozone “responders”). In this context, susceptibility can be defined by host factors such as an increased responsiveness to a given dose of air pollution or the prevalence of underlying disease, as well as by exposures to other environmental factors increasing risk for the same outcomes, for example, indoor air pollutants.9

The papers in this issue of the journal illustrate some approaches taken by epidemiological researchers to assessing vulnerability to air pollution. In addressing the question of whether persons having lower socioeconomic status are at greater risk from air pollution, epidemiologists test whether the risk estimated for air pollution (or a specific pollutant) varies across strata of socioeconomic status; such variation is referred to as effect modification. Gaining insight into modification of the effect of air pollution on health by socioeconomic status poses several methodological challenges. Firstly, socioeconomic status indicators are only surrogates for more proximal factors that determine health status and potential vulnerability to air pollution. These factors might include nutritional status and prevalence rates of chronic heart and lung diseases, for example. The finding of effect modification by socioeconomic status should trigger further research to better understand the intervening factors. Secondly, some correlates of socioeconomic status may be confounding the relation between air pollution and health. Disentangling complex causal pathways may not be possible, depending on the richness of the data available on relevant correlates of socioeconomic status. Thirdly, estimates of the extent of effect modification are notoriously imprecise, so that sample size may prove a barrier to gaining a picture of variation of the effect of air pollution by socioeconomic status.

Two of the papers in this issue assess socioeconomic status as a modifier by exploring variation of the effect of air pollution across regions within two cities: the city of Hamilton, Canada,10 and São Paulo, Brazil.11 Both investigative groups followed a similar approach: stratifying the urban region into areas defined by proximity to monitoring stations, developing ecological measures of socioeconomic status for the zones, and testing for variation in the effect of air pollution measures among the zones. Both locations had sufficient spatial variation of socioeconomic status and air pollution to test for effect modification. Despite the substantial differences between these locations, the findings of the two studies were similar in showing greater risk in areas having a predominantly lower socioeconomic status population.

The third paper addresses ambient air pollution and birth weight in São Paulo.12 In this analysis, air pollution exposures during each trimester were estimated and their associations with birth weight examined in multivariable models that took several factors, including maternal age, maternal education, and number of prenatal visits into account. A reduction of birth weight with estimated first trimester exposures to particulate matter and carbon monoxide was found. This finding adds to a growing literature on reproductive outcomes and urban air pollution.13,14 Notably, in this study, maternal education, a socioeconomic status measure, was treated as a potential confounding factor and included in the multivariable model. Effect modification was not explored.

What have we learned from these new studies? Firstly, they confirm a number of previous reports with similar findings in both time series studies of acute events2 and in longer term cohort studies of mortality.15 Secondly, the authors’ thoughtful discussions re-emphasise the need for cautious interpretation of findings on effect modification, given the range of methodological considerations affecting the results. Thirdly, research on socioeconomic status and the effect of air pollution might be improved by harmonisation of methods and pooled analyses so that differences among studies might be better understood. Clearly, socioeconomic measures have differing correlates across populations and the development of data on the most relevant correlates would be informative. A pooled analysis of mortality data from North America and Europe will soon be underway that will provide an opportunity to assess the role of effect modification across a broad range of cities.

The findings of these and other studies are beginning to provide a coherent and not surprising picture: persons having lesser socioeconomic status seem to be at increased risk from urban air pollution. Further research on this topic is warranted but studies need to extend beyond empiric exploration of effect modification to explore the underlying causal pathways. Hierarchical designs will be needed that explore the relevant individual level correlates of socioeconomic status; personal exposure assessments for key air pollutants should also be incorporated to better characterise exposure by socioeconomic status. Relevant examples include Diez Roux,16 and research methods are available for this purpose.17

Air pollution and equity

View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.