Overview of the EBoDE-project

Introduction

Exposures to many environmental stressors are known to endanger human health. Negative impacts on health can range from mild psychological effects (e.g. noise annoyance), to effects on morbidity (such as asthma caused by exposure to air pollution), and to increased mortality (such as lung cancer provoked by radon exposure). Properly targeted and followed-up environmental health policies, such as the coal burning ban in Dublin (1990) and the smoking ban in public places in Rome (2005) have demonstrated significant and immediate population level reductions in deaths and diseases. In order to develop effective policy measures, quantitative information about the extent of health impacts of different environmental stressors is needed.

As demonstrated by the examples above, health effects of environmental factors often vary considerably with regard to their severity, duration and magnitude. This makes it difficult to compare different (environmental) health effects and to set priorities in health policies or research programs. Public health policies generally aim to allocate resources effectively for maximum health benefits while avoiding undue interference with other societal functions and human activities. In order to develop such policies, it is necessary to know what ‘maximum health benefits’ are. Decades ago, such decisions tended to be made based on mortality statistics: which (environmental) factor causes most deaths? However, nowadays, most people get relatively old, and priority has shifted from quantity to quality of life. This has lead to the need to incorporate morbidity effects into public health decisions, and therefore to find a way of comparing dissimilar health effects.

Such comparison and prioritisation of environmental health effects is made possible by expressing the diverging health effects in one unit: the environmental burden of disease (EBD). Environmental burden of disease figures express both mortality and morbidity effects in a population in one number. They quantify and summarize (environmental) health effects and can be used for:

• Comparative evaluation of environmental burden of disease (“how bad is it?”)
• Evaluation of the effectiveness of environmental policies (largest reduction of disease burden)
• Estimation of the accumulation of exposures to environmental factors (for example in urban areas)
• Communication of health risks

An example of an integrated health measure that can be used to express the environmental burden of disease is the DALY (Disability Adjusted Life Years). DALYs combine information on quality and quantity of life. They give an indication of the (potential) number of healthy life years lost in a population due to premature mortality or morbidity, the latter being weighted for the severity of the disorder. The concept was first introduced by Murray and Lopez (1996) as part of the Global Burden of Disease study, which was launched by the World Bank. Since then, the World Health Organization (WHO) has endorsed the procedure, and the DALY approach has been used in various studies on a global, national and regional level.

WHO collects a vast set of data on the global burden of disease. The first study quantified the health effects of more than 100 diseases for eight regions of the world in 1990 (Murray and Lopez, 1996). It generated comprehensive and internally consistent estimates of mortality and morbidity by age, gender and region. In a former WHO study, it was shown that almost a quarter of all disease worldwide was caused by environmental exposure (Prüss-Üstün and Corvalán, 2006). In industrial sub-regions this estimate was about 16% (15–18%). These fractions, however, are dependent on the conclusiveness of the included environmental factors and health effects. The WHO programme on quantifying environmental health impacts has addressed more than a dozen stressors . In order to support further applications of the environmental burden of disease (EBD) assessments, a methodological guidance has been published by WHO (Prüss-Üstün et al., 2003) and was followed here too.

In Europe, national environmental burden of disease (EBD) assessments are on-going in several countries. The work by RIVM was one of the first systematic European works in this area that utilized disability-adjusted life years (DALY) as a measure to compare the burden of different health outcomes related to the exposure of the population to environmental stressors (Hollander et al., 1999). The results highlighted that (i) a number of environmental stressors may cause chronic or acute diseases or death, (ii) a few top ranking stressors cause over 90% of the national EBD, and (iii) these top ranking stressors are not necessarily those that have drawn the most concern, regulatory action and/or preventive investment.

Objectives

The EBoDE-project was set up in order to guide environmental health policy making in the six participating countries (Belgium, Finland, France, Germany, Italy and the Netherlands) and potentially beyond. From a policy perspective, these insights from the EBoDE-project can be useful to evaluate past policies and to gain insight in setting the policy priorities for the future. We have calculated the total EBD associated with the nine environmental stressors. The total EBD is not identical to the avoidable burden of disease, because some exposures are not realistically reducible to zero (e.g. fine particles). Also, our estimates do not take into account the costs of reducing the EBD. Thus, the results are only one input into the full process of developing cost-effective policies to achieve better environmental health.

The objectives of the project were to update the available previous assessments, to focus on stressors relevant for the European region, to provide harmonized EBD assessments for participating countries, and to develop and make available the methodologies for further development and other countries. The specific objectives are to: • Provide harmonized environmental burden of disease (EBD) estimates for selected environmental stressors in the participating six countries; • Test the methodologies in a harmonized way across the countries. • Assess the comparability of the quantifications and ranking of the EBD • between countries • within countries • between environmental stressors; • Qualitative assessments of variation and uncertainty in the input parameters and results.

Environmental burden of disease estimates have been calculated for: • nine environmental stressors: benzene, dioxins (including furans and dioxin-like PCBs), second-hand smoke, formaldehyde, lead, noise, ozone, particulate matter (PM) and radon; • six European countries: Belgium, Finland, France, Germany, Italy and the Netherlands; • the year 2004 (and some trend estimates for the year 2010). As outlined above, the EBoDE study was carried out in order to test the environmental burden of disease methodology in various countries. The results of the studies are intended to allow comparison of the disease burden between different environmental stressors and between countries. Consequently, the study does not to identify the ‘reduction potential’. Our estimates should therefore not be interpreted as the ‘avoidable burden of disease’: most risks cannot realistically be completely removed by any policy measures. For some exposures, however, the numbers may nonetheless be interpretable as reduction potential, eg for dioxins, formaldehyde, benzene, etc, as these exposures could potentially be completely eliminated.

Outline of this report

This report describes the methods, data and results of the EBoDE-project. Chapter 2 presents the methodology. The environmental stressors are introduced in Chapter 3, which also presents the data used (selected health endpoints, exposure data, exposure response functions). In Chapter 4, the results are presented and discussed. Chapter 5 gives information about uncertainties in the approach, and provides some alternative calculations using different input values. In Chapter 6 conclusions are drawn. The report ends with the references and two appendices: Appendix A presents country-specific results and Appendix B some considerations for using a life-table approach in EBD modelling.

Uncertainties and limitations

Assessment of uncertainties is essential in a comparison of quantitative estimates that are based on data from heterogeneous sources and slightly varying methods. Due to the wide range of data sources and models and the limited resources within the EBoDE project, systematic analysis of all uncertainties was not possible. However, we were able to assess a number of specific sources of uncertainties in more detail as part of the work, yielding some insights into the reliability of the overall assessment. The studied health impacts span approximately four orders of magnitude in size from few DALYs per million to almost 10 000 DALYs per million. The overall ranking of the environmental stressors seems to be rather robust against the relatively large uncertainties in individual estimates or methodological choices like discounting and age-weighing. However, some of the estimated ranges are overlapping. This concerns especially second hand smoke, radon and transportation noise that compete for the questionable honour of being the second most important environmental stressor in the participating countries. Among these stressors the differences are smaller than the corresponding uncertainties of the estimates. The health state of an individual person is the result of a complex mixture of genetic, environmental and behavioural factors. In a typical case of death, numerous factors play together. This means, for example, that a single death caused by a cardiovascular disease could be avoided by either reducing air pollution, or a better diet, or more physical activity. Therefore, if the individual attributable fractions are summed over a number of risk factors, a value over 100% may sometimes be found. For this and other reasons, it has been argued that death counts are not suitable for quantification of the impacts (Brunekreef et al., 2007). Therefore the authors recommend to mainly use aggregate population measures of health like DALYs, YLLs and YLDs. This chapter presents the quantitative results for selected sources of uncertainties and discusses the project limitations and author judgment of the reliability of the ranking.

Uncertainties per stressor and comparison with other studies

A list of the most important sources of uncertainty for each stressor in the EBoDE calculations is provided in Table 5-1. Some of these are further explained below. In addition, we will compare our estimates to results of a selection of similar studies. Comparison of different studies on environmental burden of disease helps to understand the role of various methodological and strategic selections made in each study, like the selection of stressors or health endpoints.

SHS

Our burden of disease calculation for SHS was based on a WHO model (Öberg et al., 2010). The exposure estimates were updated against available national and international data sources for the target year 2004, but otherwise the results are comparable with the WHO assessment. Other recent estimates of burden of disease for SHS were also available for Germany (Heidrich et al. 2007; Keil et al. 2005), which provided similar results as the current estimates.

Transportation noise

Burden of disease estimation for transportation noise is currently under active development. The estimates presented here were based on the only available international exposure data source, the first stage version of the European Noise Directive database (2007), which is not conclusive yet. Therefore it is clear that most of the exposures for transportation noise are underestimated. In some studies annoyance and cognitive impairment have been used as an additional health end-points for environmental noise. However, due to the selected more limited definition of ‘health’ as ICD-classified health states used in our assessment, annoyance and cognitive impairment were not included here. Only road, rail and air traffic exposures were included; many other sources also contribute to the noise exposures. Low exposures below the END data collection limits (50 and 55 dB) were not included. For these reasons it can be expected that when these limitations are solved, the impact estimates will increase.

PM and ozone

The methodology developed in Clean Air for Europe -project (CAFE) (Hurley et al., 2005) was applied using updated exposure estimates. The updated exposures are based on ambient air quality monitoring data that contain, besides the anthropogenic components that CAFE focused on, also natural sources of PM2.5. The spatial resolution of the updated model is 25 times higher (grid size 10x 10 km² instead of 50x50 km²). Compared to the CAFE estimates the current work adds estimation of the impacts in DALYs. The WHO Environmental Burden of Disease programme uses a non-linear exposure-response function (Ostro, 2004) that at higher exposures yields lower impacts than the linear CAFE model. WHO also sets a threshold level at 7.5 μg m-3.