Summary of Meeting Paper

The 1996 Annual Meeting of the Society for Risk Analysis-Europe

Combining Air Pollution Epidemiology and Risk Perception and Communication in the CESAR Study. Erik Lebret, National Institute of Public Health and the Environment, Bilthoven, The Netherlands; Tony Fletcher, London School of Hygiene and Tropical Medicine, London, UK; Bert Brunekreef, Wageningen Agricultural University, Wageningen, The Netherlands; and Alan Pinter, National Institute of Hygiene, Budapest, Hungary

BACKGROUND

In 1991 the European Commission PHARE programme proposed a research programme on the effects of air pollution on human health in six Central European countries (Bulgaria, Czech Republic, Hungary, Poland, Romania, Slovak Republic), with the overall aim to:

• identify population groups at risk from exposure to air pollution;

• determine the impact on health of such exposures;

• provide sound scientific basis for reducing health risks in the most cost-effective way;

• utilise the data obtained to improve the understanding of air pollution and health;

• facilitate the exchange of data and expertise between regions with similar air pollution problems and research institutes in Europe.

Based on results of a feasibility study (Lebret et al. 1993¹), the CESAR (Central European Study of Air Quality and Respiratory Health) project was developed. The project has two complementary objectives:

1. To establish comparable base-line data on respiratory health of children and on current exposure to air pollutants in polluted regions in the six participating countries and to investigate the relationship between respiratory health and air pollution;

2. To strengthen the research capacity in the participating countries to undertake high quality environmental epidemiology studies.

OUTLINE OF CESAR COMPONENTS

The CESAR study was developed by a consortium of National Institute of Public Health and the Environment (The Netherlands), the London School of Hygiene and Tropical Medicine and Wageningen Agricultural University in conjunction with the National Research Teams (NRT) in each country. The CESAR study consists of three parts.

1. a Study on Air Pollution and Respiratory Disease in Children.
   This study involves the protocol development, data collection, analysis and reporting of an investigation into the relationship of S0₂ and Suspended Particulate Matter (PM10 and PM2.5) levels and pulmonary function, and respiratory symptoms in children aged 7-11 years. In each country, four study areas were selected with different levels of these pollutants.

2. a Quality Assurance activity.
   This project is fully integrated into project 1 to optimise the quality of data , but also to assure a good infrastructure for forthcoming projects.

3. a Risk Perception and Risk Communication Study.
   This part forms the background to developing a strategy for the communication of the results of project 1. The project principally explores how risk is perceived in the study areas of project 1, what information is received and what kind of information is needed, with emphasis on the outcomes of project 1. The study consists of two parts: 1) a qualitative survey on risk perception by interviewing selected stakeholders in the local communities; 2) a quantitative questionnaire survey on risk perception in a random sample of the local communities under study. This part of the CESAR project is more extensively described elsewhere in these proceedings (Fletcher et al. 1996²).

DEFINITIONS OF RISK

Within the context and overall aim of the CESAR project, at least three groups of people are dealing with and communicating about environmental health risks. These are 1) the scientists in the air pollution epidemiology study on respiratory health in children; 2) the general public in the study area who are in part enrolled in the study; and 3) the environmental health policy makers in these areas. It is expected that these different groups (and other stakeholders) will use different sets of definitions for risk.

In most dictionaries 'risk' is defined as 'the danger of damage, harm or loss'. In the context of environmental health, risk can be circumscribed as the probability of damage to public health, in combination with its nature and magnitude. Risk may be associated with human activities, natural events, or a combination of the two. An example of the latter are episodes of summer smog that are caused by a combination of high levels of antropogenic precursor pollutants and stagnant, sunny meteorological conditions. Both dimensions, probability and (uncertainty over) the occurrence, timing and magnitude of consequences are involved in the evaluation of acceptability of a risk generating activity or situation. In the literature on risk assessment various definitions of the concept of risk are being used. On the one extreme, some authors conceptualise it merely as the probability of an adverse outcome or a set of outcomes; on the other extreme, in particular among social scientists, risk is defined solely in sociological terms, such as social equity (who runs the risk, who benefits), voluntariness of exposure, familiarity with or perceived control over risk generating activities, trust in risk management, or social group interests. In general, different attributes of risk encompassed by these extremes may be addressed in risk management, depending on the nature of the risk problem.

In the epidemiological context, risk is often defied as: "The probability that an event will occur, e.g. than an individual will become ill or die within a stated period of time or age" (Last 1995³) Such measures of absolute risk are often contrasted to measures of relative or attributable risk in epidemiology. The relative risk is the ratio of risk of disease or death among the exposed to the risk of the unexposed. This is typically estimated in analytical studies such as. the one on respiratory health in children in the CESAR project, where the risk is determined in relation to exposure to air pollution. The attributable risk is the rate of disease in exposed individuals that can be attributed to (is caused by) the exposure. This measure of risk can be important to gauge the (potential) health gain when exposure is reduced, e.g. due to policy measures.

The experience in the Feasibility study indicated that an apparent approach used to identify "risk areas" by policy makers and scientists in the Central European countries involved in the project were based on:

• either unfavourable health status indicators (short life-expectancy, high specific mortality, morbidity, IgE levels) compared to national average and/or a control area; or
  

• unfavourable pollutant levels compared to national levels or to control area, or to air quality standard.

RISK CHARACTERISATION IN THE CESAR STUDY

The primary perspective of the risk characterisation in the CESAR study is an epidemiological one. The output of the project on respiratory health in children will be figures of the absolute risk of having CNSLD/COPD broken down by sex and age for the different study areas. Similarly, distribution of pulmonary function level in the children will be reported. In terms of risks due to exposure to (environmental) determinants of respiratory health, relative risks for air pollutants levels will be reported, as well as relative risk associated with a number of confounder variables, e.g. family history of respiratory disease, environmental tobacco smoke, home dampness, etcetera. It is generally acknowledged in epidemiological circles that results from a single study should be interpreted with extreme cautions, 'm particular when relative risks smaller than 3 to 5 are observed (Sauter 1995⁴) . This means that results of the CESAR study need to be integrated with results reported in the literature. A qualitative comparison is quite common, of course, but this usually does not allow the quantitative estimation of for instance, attributive risk in the population. To do this, a more formal meta-analysis and quantitative risk estimation would be required.

RISK COMMUNICATION IN THE CESAR STUDY

The risk communication of the CESAR study takes places at several levels. Firstly, the results are fed back to the subjects who participated in the study, i.e. the parents of the children in the respiratory health study, and respondents in the risk perception study. Secondly, the results of the study are communicated to the environmental health policy makers and stakeholders at the local and national level. Thirdly, the results will be published in scientific journals. The latter target group is usually well understood and concepts and terms used for risk characterisation well described. For the other two target groups this is far less true. The risk perception part of the projects is partly aimed at the elucidation of this, i.e. to assess the risk perception and knowledge about environmental health risks in the population, and to evaluate how information on environmental health could be best disseminated.

Given the different notions and interpretations given to risks by the various target groups, it is anticipated that the common set of terms to characterise risk is probably limited, as depicted in the diagram below.

The risk communication process is expected to develop in three phases. The NRT and consortium must first agree on the validity and interpretation of various risks estimates obtained in the project in the scientific context. Then, these scientific risk characteristics must be translated into risk terms and dimensions as used by the public and by policy makers. Next, the proper channels for dissemination of information need to be selected and used. It is envisaged that this will involve presentations at public hearings, following some prior training of members of the NRT in communications skills.

DISCUSSION

The CESAR study has, as one of its important features, the combination of an environmental health study with a risk perception and communication study. In the ideal world, the risk perception part of the study would have been the initial step in the definition phase of the CESAR project. As it was, only limited consultation of, among others, NGO's could be made as part of the Feasibility study that preceded the definition of CESAR. The integration of the scientists', policy makers' and general public's perspectives on risk and risk characterisation is therefore far from complete yet, and has to be develop along the way of the project. It is clear, however, that expansion of a common set of definitions and dimensions of risk that are shared by public, policy makers and scientists is needed to properly address questions about environmental health risks in this and other situations.

¹Lebret E, Wolters N, Elliott P, Fletcher T. Feasibility Study for a Plan of Action to Investigate the Effects of Air Pollution on Health in PHARE Countries. Part I. Report 263510001. National Institute of Public Health and Environmental Protection, Bilthoven, 1993.

²Fletcher T, Jones K, Ball D, Woudenberg F, Lebret E. Results of a Study of Risk Perception in six Central European Countries. Proceedings of the Annual SRA-Europe Meeting Risk in modern society: Lessons from Europe. June 3-5, 1996, Guildford, UK.

³Last JM (ed). A dictionary of Epidemiology. Oxford University Press, Third Edition. New York, 1995.

⁴Taubes G. Epidemiology faces its limits. Special news report. Science 269;1995:164-169.