Human health frames in EIA – the case of Swedish road planning

Abstract

How human health is framed provides boundaries for choices in practice and bias for certain actions in health assessments in Environmental Impact Assessment (EIA). This study examines health frames in legislation and policies of importance for EIA in Swedish road planning and their implications for practice. The emphasis is on different approaches, such as promotion of health and prevention of ill-health. Indicators of the choices in practice for which the health frames exert bias are further analysed through a review of Environmental Impact Statements (EIS) for road planning and comparison with a similar study conducted about 10 years ago. The indicators studied are: health determinants included, health impacts assessed, and aspects of the affected population concerned in the EISs. There are fundamentally different health frames in Swedish legislation and policies, but this range is not yet reflected in EISs, which mainly focus on environmental health rather than on broader health determinants and health equity. The results indicate that this situation becomes a dilemma for EIA practitioners and a challenge for the field.

Keywords:

• Human health
• EIA
• road
• traffic
• frames
• framing

Introduction

Increasing mobility and greater accessibility to places and services provide many social and practical advantages (Gudmundsson & Höjer 1996). However, transportation ‘externalities’ also contribute to ill-health (Dora 1998; Dora & Racioppi 2003; McCarthy 2006) and the negative health impacts from transport are unequally distributed (Woodcock & Aldred 2008). Some of the threats to human health (e.g. psycho-social tension and stress-related problems, obesity, asthma and allergies) identified in the Swedish national context (SNBHW 2009a; 2009b) are due to, or can be exacerbated by, road traffic and infrastructure planning not taking human health issues into consideration. In many countries road planning is subjected to statutory Environmental Impact Assessment (EIA) in order to create information about the effects of transportation and provide possibilities to enhance or prevent negative effects. In Sweden, the implementation of EIA is in fact the main mechanism for bringing information and knowledge on human health and other environmental impacts into the road planning process.

There is a growing body of international literature relating to how human health is assessed in EIAs (Kågström 2009). In the Swedish context, however, there is a general lack of academic studies and evaluations of EIA practice (Emmelin 1998; Hilding-Rydevik & Bjarnadóttir 2007). This is also the case for how health is assessed in Environmental Impact Statements (EISs) for road projects. Three existing studies, carried out 8–12 years ago, identified a number of short-comings or difficulties in the assessment of health in Swedish EIA in general (Hilding-Rydevik et al. 2005; SNBHW 2001) and for road planning in particular (Alenius 2001). However, there have been no more recent studies on how human health is managed in Swedish EIA for road planning. This issue is especially important since two of the studies of the Swedish context, that is, Alenius (2001) and Hilding-Rydevik et al. (2005), are frequently referred to in international studies, which could communicate an outdated picture of the Swedish situation. Up-to date studies are needed owing to the introduction of National Public Health Objectives in Sweden in 2003 and the growing political recognition of the need to develop assessments of health in planning the built environment, such as road planning. In parallel to the implementation of health in EIAs, through public health policy, the Swedish government has also emphasized the importance of Health Impact Assessments (HIAs). However, it was decided not to make HIA mandatory under Swedish law, whereas EIA is mandatory. This is also the situation in several other countries (Harris-Roxas et al. 2012).

A review of international research related to EIA and human health indicates that health is interpreted and understood in quite fundamentally different ways (Kågström 2009), which has provided ample room for quite different practices to develop in EIA. This was an important starting point for us in this study of the current Swedish situation in relation to implementation of human health in EIA for road planning. The aim of the study is to provide a better understanding of how health is managed in Sweden today. This is achieved first by reviewing how health is approached in current (EISs) in road planning; and second by studying, from a framing perspective, Swedish legislation and policies of importance for health assessment in EIA for road planning. Socially constructed frames provide boundaries for what are perceived as appropriate actions in certain situations, thereby creating a bias for certain decisions in practice. How health is framed in legislation and policies is thus important for the development of Swedish EIA practice for road planning and for how health is assessed in EISs. Throughout this paper, the results obtained for Sweden are compared with international findings and with results presented in earlier studies of the Swedish context. Four themes are used in the analysis. The first theme focuses on issues that show the fundamentally different approaches that can be taken in EIA, that is, whether the approach is promotion of health or prevention of ill-health. The three following themes are indicators of the choices in practice for which different health frames might create bias in terms of the health determinants included, the health impacts assessed and how this is done, and the aspects of the affected population that are covered in the EIA. The results can be used in future comparative studies in the field in other national contexts. The study is also intended to contribute to Swedish EIA practice by determining the space available for different choices in EIA practice in relation to ‘health determinants’, ‘health impacts’ and ‘affected population’ and how the different health frames in the Swedish context set different boundaries for these choices.

The paper continues with a presentation of the theoretical basis for the study, that is, concepts of frames and framing, followed by a presentation of different ways of framing health as a background to understanding the themes used for the analysis. This is followed by a presentation of materials and methods included in the study. The next section of the paper is devoted to an analysis of Swedish policy and legislation and the results from our analysis of Swedish EISs in road planning. These results are then related to findings from our review of the international literature. The paper concludes with a discussion of the content of EISs and the boundaries for choices in EIA practice, in light of the space provided by the health frames in legislation and policies of importance for health assessment in Swedish EIA for road planning.

Concepts of frames and framing

Frames and framing are used in everyday language for describing the physical (framing a picture). However, the concept of frames and framing has also proven useful in an abstract sense in a number of different disciplines (Benford & Snow 2000). This makes frames a rich concept with several different conceptualizations. In this paper, frames are perceived as social constructions (6 2005), a theoretical context in which human beings are seen as social, while reality is socially constructed and reconstructed in communication and interaction between people (Berger & Luckmann 1966; Burr 1995; Gergen 2001). Social constructions are historically specific and contingent and may thus change over time (Winther Jorgensen and Philips 2002).

frames organise experience; that is to say, they enable people to recognise what is going on, they provide boundaries, define what counts as an event or a feature; crucially, frames define what counts as relevant for attention and assessment. Secondly, they bias for action; that is to say, they represent people's worlds in ways that already call for particular styles of decision or of behavioural response. (6 2005, p. 94)

In the present paper frames are understood as biasing for certain action, but it is also emphasized that this is only a part of the picture when trying to understand how people act. Frames do not determine action, but are viewed here as important for understanding boundaries for choices made in practice. In other words, frames provide boundaries for what are perceived as appropriate actions in a certain situation. In this way, frames bias for certain decisions in practice. This is important in health assessments, since how health problems are framed determines the types of solutions proposed. For example, if health is framed in medical terms, only solutions of a medical character will tend to be proposed, while other framings can prompt people to think in different ways, for example, in terms of urban planning (WHO 2001). Important frames guiding EIA practice and exerting a bias for appropriate action are those outlined in legislation and policies related to EIA. The health frames in these thus provide boundaries for what decisions are considered appropriate when making choices about important aspects of human health in EIA practice.

Ways of framing human health

The meaning of health, and of ill-health, has evolved over time and is closely linked to personal and socio-cultural values (Tones & Green 2008). The effort to define health and what is required for human well-being has a long lineage (Medin & Alexanderson 2000; Tones & Green 2008). Different theories of health stem from a wide range of different approaches, such as sociology, philosophy, religion, psychology, ecology, natural science and medicine (Medin and Alexanderson 2000). There are thus significant ideological differences between these approaches (Tones & Green 2008). In a literature review of concepts of health and health promotion, Medin and Alexanderson (2000) categorize different health orientations into two broad approaches: the biomedical and the humanistic. Framing in the biomedical approach views the human body and ill health (which is seen as the opposite to health) as central. In the humanistic approach, health is framed as something more than the absence of disease. Tones and Green (2008, p. 1) have chosen to acknowledge the dichotomy between ‘a “positive” approach to conceptualisations of health and “disease-focused” definition’. The World Health Organization (WHO) is an important health framing actor and the frequently cited definition of health in its 1948 constitution is: ‘a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity’ (WHO 2006). The WHO has also framed health as a basic human right and has emphasized the equity dimension, that is, the need to ensure health for all (e.g. WHO 1997).

Factors in society and living conditions contribute to good or bad health (Dahlgren & Whitehead 2007; Knutsson & Linell 2010). These factors are called health determinants and are closely interlinked. Thus alteration of one determinant can alter others, all of which influence health. Barton and Grant (2006) claim that a new road has the potential to exert an impact on every sphere representing health determinants – except the inherited. In relation to social, economic and behavioural factors, some groups of people are at higher risk of experiencing ill-health (Dahlgren & Whitehead 2007). Individuals can also be affected differently by the same alteration of a determinant; for example, children, among other groups, are more sensitive to an increase in noise level (Forsberg 2004). It is therefore important to look at different prerequisites in the population affected by a change in health determinants and focus on health equity.

The core of the framing concept is to acknowledge that, given the same information, all actors will not necessarily interpret it the same way or find it of similar relevance (Schön & Rein 1994). Thus, even if inclusion of certain health determinants, health impacts and aspects of the affected population in health assessments in EIA may seem crucial to many, not everyone agrees. However, it is obvious that different health frames, such as the health or ill-health approaches, direct attention to certain kinds of health determinants and health impacts. They also include/exclude certain aspects of the affected population as relevant for assessment. These concepts are therefore important for analysing the management of human health in EIA.

Materials and methods

As background to this study, a search was made for international research literature in the area. The search covered national and international publications that had studied the practice of managing health aspects in EIA, had evaluated actual practice in this area or had described other processes focusing on the same purpose. The key search terms used were: health, well-being, impact assessment, EIA, infrastructure and transport. Full details of the results are presented in Kågström (2009).

The international literature review provided important input to the analysis of the Swedish situation and context. In the legislative and policy texts of importance for Swedish EIA for road planning, concepts of health and concepts closely related to health were sought, highlighted and compared with each other and in relation to the four themes of the present study, presented in further detail below. Two earlier reports from the Swedish National Board of Health (SNBHW 2001, 2004) on the handling of health aspects in EIA were used to further relate to the theme of ‘health’ or ‘ill-health’ approaches in the Environmental Code (SFS 1998:808) and the Planning and Building Act (SFS 1987:10). The following legislative and policy texts were used: the Environmental Code, the Planning and Building Act and policies for national objectives for Public Health (Government Bill 2002/03:35; 2007/08:110), Environmental Quality (Government Bill 1997/98:145, 2005/05:150) and Transport (Government Bill 2008/09:93).

The review of Swedish EISs includes 17 EISs on road planning issued during the period 2005–2009. We chose to study EISs for the feasibility study part of the planning phase, which occurs sufficiently early to allow the inclusion of a wide range of health issues and sufficiently late to provide an assessment of health impacts in some detail. The cases are selected by probability sampling. A full review of the documents is reported in Sjöberg (2010). The results are compared with those of an earlier study (Alenius 2001), in order to examine whether there had been any change over the intervening years. Alenius (2001) studied 28 EISs and SEA (Strategic Environmental Assessment) for road projects in Sweden produced from 1990 and about 10 years forward (the exact period is not specified), although she did not stipulate the planning process phase to which her cases were linked. In the review, both studies used checklists based on the international literature about relevant aspects, possibilities and approaches when describing and assessing the health impact from road planning.

As noted in the introduction, the analysis was based around four themes drawn from our review of ways of framing health. The first theme, which concerned issues that show the quite fundamentally different approaches that can be taken in EIA, was mainly used to give an overview of differences in framings of health in the relevant national legislation and policies. The three subsequent themes, which concern indicators of the choices in practice which different health frames might bias, were mainly used for analysing the EISs. The themes are:

	1. ‘Health and ill-health’
	Recognition of positive approaches focusing on health and health promotion and negative approaches focusing on prevention of ill-health.
	2. ‘Health determinants’
	Recognition of different kinds of health determinants and their links to health impact.
	3. ‘Health impacts’
	Recognition of different kinds of health impacts and ways of assessing these.
	4. ‘Affected population’
	Recognition of different prerequisites in the affected population and health equity (or inequalities).

It is sometimes difficult to make a clear distinction between health determinants and health impacts, since an impact on one determinant can have a direct impact on health and on another determinant, giving a secondary impact on health. However, it is still important to establish the framework and sub-divisions listed above, since they render the health choices more discernible, that is, the health determinants and impacts that are seen as relevant to include and assess.

Health in legislation and policies

The findings from the review of the international literature on health in EIA in general in relation to legislation and interpretations of health are presented below. This is followed by a presentation of the Swedish context, focusing on legislation and policies of importance for road planning EIA.

International

Legislation and policies related to EIA are important frames that bias towards the emergence of certain EIA practices. It is therefore highly interesting that a frequently mentioned key barrier to the promotion and inclusion of health in EIA in many countries is the vagueness of the legislation, that is, in not explicitly calling for health aspects to be assessed (Hilding-Rydevik et al. 2005; Obekpa Abah 2012; Steinemann 2000). However, there are also examples of the relationship to health becoming clearer, for example with legal demands for integration of EIA and HIA (Spickett et al. 2012 ).

There are differences between countries in respect of the health frames that are promoted, even when the basis is the same piece of legislation. The European Union (EU) member states often use varying interpretations of EU directives and generally use different notions of health and ‘healthiness’ in their national legislation (Hilding-Rydevik et al. 2005), providing different boundaries for choices in EIA practice. Approximately two-thirds of EU member states have a narrow approach, while the rest use a more expansive approach that is closer to the WHO definition (Hilding-Rydevik et al. 2005). In addition, there are also different opinions within a country as to whether national legislation includes health or not, or what kind of health aspects should be assessed, for example, in Canada and the USA (Bhatia & Wernham 2008; Kågström 2009; Noble & Bronson 2005; Steinemann 2000; Wernham 2007).

The way in which these health frames are interpreted differs between different stakeholders and practitioners, as do their opinions of what is a relevant action. A study by Noble and Bronson (2006) in northern Canada found consensus among interviewees that health should be considered in EIA, but lack of consensus regarding ‘what kind of health’. In a study covering 10 EU countries, the USA and Canada, Hilding-Rydevik et al. (2005) found a significant minority view that health in EIA should be understood as a state of complete physical, mental and social well-being, while others were of the opinion that EIA already covers enough in relation to health issues.

Sweden

The most important legislation providing health frames for EIA practice in Sweden is the Environmental Code (SFS 1998: 808), with its clear demand for health to be included in EIA. However, in Sweden today there are several other pieces of legislation and policies that also have to be implemented in EIA practice for road planning. Taken together, they provide several parallel, sometimes even opposing, health frames that are intended to guide assessment of health in EIA for road planning (see Figure 1). There follows a more detailed description of these different frames and of the boundaries for choices about health determinants, health impacts and affected populations for which they bias.

Figure 1 Human health approaches in Swedish legislation. Several policies become relevant when conducting an EIA for a Swedish road project. The term ‘health’ is often used here, although a number of other concepts closely linked to health are also mentioned. Taken together, they embrace several health determinants (modified from Kågström 2009, p. 22).

The Environmental Code is the main EIA legislation in Sweden and functions alongside sector laws for planning and building, such as the Planning and Building Act (SFS 1987:10) and the Road Act (SFS 1971:948). The Swedish planning system for transport and infrastructure calls only for EIA. However, since transport planning is closely linked to land use planning (Alenius 2001; Tricker 2007; Kørnøv 2009), SEAs are sometimes also undertaken for the municipal detailed plan for the same project and area. In a study by the Swedish National Board of Health and Welfare (SNBHW 2001), no distinction is made between how health is described in EIA and SEA, on the basis that the Environmental Code and the Planning and Building Act relate to each other. The Environmental Code and the Planning and Building Act lack explicit definitions of health, but each uses different concepts and formulations related to health, which builds up two fundamentally different health frames. One aim of the Environmental Code is to ‘protect against ill-health’, while the Planning and Building Act aims to ‘plan for health’ (SNBHW 2001). The health frame in the Environmental Code mainly biases for environmental health determinants, that is, noise and air quality. The Planning and Building Act takes a broader approach. In this context it could be argued that the Environmental Code appears to be more concerned with setting the lowest acceptable limit for disturbances in the environment rather than promoting a good and healthy environment (SNBHW 2004). The Road Act, which contains regulations on the planning of roads, lacks a clear definition of health or a clear health frame in respect of the planning process.

From interviews reported by The Swedish National Board of Health and Welfare (SNBHW 2001), it is apparent that different professions understand the concept of health in EIA legislation in different ways. Some interviewees argue that health should not be considered solely in relation to physical health, but should also be related to social and economic determinants. These differences in perceptions are also reflected in the findings from northern Canada by Noble and Bronson (2006) outlined above. In the Swedish study (SNBHW 2001), there was also some disagreement between professions over what a health effect actually comprises and what the final health impact could be.

In the Swedish government's National Strategy for Sustainable Development, ‘good health’ is one of the four main goals (Regeringens skrivelse 2003/04:129; 2005/06:126). The content of the strategy is somewhat clarified in the 16 Environmental Quality Objectives (Government Bill 1997/98:145; 2005/05:150) and the 11 Public Health Objectives (Government Bill 2002/03:35; 2007/08:110). The Environmental Quality Objectives describe the quality and status that should be achieved in respect of the Swedish environment and that are outlined as being ecologically sustainable. The health frame thus mainly focuses on environmental determinants such as noise, air, water, soil, forests, wetlands, climate and radiation. However, the frame is rather broad and includes both protection from ill-health and promotion of good human health approaches (SNBHW 2004), for example, good water quality, possibilities for outdoor recreation and experiences. The Objectives mainly relate to direct effects on physical health, such as exposure to air pollution, but also include mental and social impacts owing to, for example, exposure to noise.

In 2003, a new public health policy was introduced in Sweden that outlined the national Public Health Objectives. The overall objective for public health policy in Sweden is to create the social conditions for the development of good health on equal terms for the whole population (Government Bill 2007/08:110). The Public Health Objectives are based on several determinants for health, including living conditions, environments, products and lifestyles, which builds up a rather broad health frame. Some view this as a paradigm shift in public health policy in Sweden (Knutsson and Linell 2010), as health objectives had previously been based on diseases and health problems. In contrast, the new focus on health determinants shifts the majority of the responsibility for health to sectors outside medical care services (Knutsson & Linell 2010). Owing to the broad range of health determinants, the frame includes physical, mental and social impacts. Since the overall aim is to secure health on equal terms, different groups with ill-health or high risks of this are defined and highlighted as being a high priority.

The overall Transport Policy Objectives in Sweden are ‘to secure the socio-economically efficient, long-term sustainable provision of transport for citizens and the business community throughout the entire country’ (Government Bill 2008/09:93); and to direct Swedish transport planning. There are two sub-objectives; ‘accessibility’ and ‘security, environment and health’. This is the first time that health has been included in the Transport Policy Objectives (Government Bill 2008/09:93). In the security, environment and health sub-objective the health frame is rather narrow, mainly related to air pollution, noise and promotion of physical activity in relation to enhancing possibilities for walking and cycling. In both the Public Health and Transport Policy Objectives, reference is made to the Environmental Quality Objectives. However, no cross-reference is made between the Public Health and Transport Policy Objectives. The link between transport policy and public health policy thus remains rather opaque.

In sum, the main EIA legislation in Sweden, the Environmental Code, provides a frame focusing on ill-health, while the Planning and Building Act provides a broader frame focusing more on health. Consequently, the Environmental Code frame limits the choice of appropriate health determinants and impacts that can be included, while the Planning and Building Act frame also directs attention to promotion of health, a broader scope of health determinants and a stronger focus on health equity. All the Swedish national objectives include some health promotion focus, even if the avoidance of ill-health frame is most dominant in the Environmental Quality and Transport Policy Objectives, which tighten the boundaries for choices in practice. However, taken together they open up a space for quite different choices on what to include in health assessments in EIA for road planning in practice.

Human health in EIS

The section below is organized according to the ‘health determinants’, ‘health impacts’ and ‘affected population’ themes in the framework presented earlier. For each theme, there is first a presentation of findings from the international literature review of health in EISs in general (Kågström 2009), followed by a presentation of the Swedish context focusing on EISs for road planning. The latter comprises findings from a review conducted in 2001 (Alenius 2001), which included EISs and SEAs for road planning completed mainly in the late 1990s, and our study, which included EIA documents for road planning completed during the period 2005–2009 (Sjöberg 2010). In the latter study we looked for operationalized health definitions in EIAs but could only find such a definition in one EIA, that with the broadest health scope. This definition embraced following health determinants: air pollution, noise, traffic safety, outdoor recreation and barrier effects; and the following health impacts: feeling of (in)security and inconvenience, wellbeing, ill-health and death.

Health determinants

International

The major shortcomings in the EISs identified in the literature review process are that either the issue of health is either not included at all or the health scope is narrow (e.g. Hilding-Rydevik et al. 2005; Steinemann 2000). The total lack of health inclusion occurs mainly in the earliest studies, covering EIAs from the 1970s and 1980s (Steinemann 2000; WHO 1987). In most EISs, human health is generally understood as relating to the environmental risks to health, such as the risks to health posed by changes in the physical environment (e.g. noise, pollution of air, water and soil). This means that broader social determinants are either totally lacking or, at best, inadequately addressed (e.g. Cherp 2002; Noble and Bronson 2005; 2006). Previous research also shows that, even if important health determinants are included in the EIA, it is not certain that links will be made to their health impacts (e.g. Harris et al. 2009; Steinemann 2000).

Sweden

The results from the study by Alenius (2001) show quite a narrow health focus, with the primary emphasis on environmental health. Our results from 2010 confirm that this narrow health focus more or less persists. The health determinants and their frequency in EISs are presented in Figure 2. Noise is the only health determinant represented in all documents in both studies (light grey and dark grey together). Almost all of the others are represented in at least 75% of the documents. In addition to these, Alenius (2001) also studied traffic safety, which was included in 75% of the documents studied. By the time of our study, traffic safety was most often represented in other formal planning documents. In addition, we surveyed drinking water and opportunities for physical activity (since the latter are highlighted in the Transport Policy Objectives). Drinking water was included in almost 50% of the EIAs, whereas the issue of opportunities for physical activity was not given its own section in any of the EIAs, but was mentioned in the outdoor/recreation section in some.

Figure 2 Comparison between a study from 2001 (Alenius 2001; A in diagram) and one from 2010 (Sjöberg 2010, S in diagram) regarding the health determinants included, and whether potential health impacts are described, in EIAs for road planning. * Air pollution impacts were considered, but excluded as not relevant in one EIA. The same occurred for vibration impacts in two EIAs.

The results also show that some determinants for health are often included in EIAs, but that the health impact that can arise from altering these determinants is seldom clearly and effectively described (see more below). For example, our determinant of drinking water was included in almost 50% of the EISs, none of which made any connections with any health impacts.

Health impact

International

Different kinds of limit values are often used in assessing health impacts. The related problem highlighted is that the assessment does not go far enough in describing the actual consequences for human health, but stops with a description of whether the limit value will be exceeded or not (Steinemann 2000; WHO 1987). Limit values may be old and the result of compromises between health information, politics and economic considerations, and thus do not always secure a healthy environment (SNBHW 2001; WHO 1987). Thus Steinemann (2000, p. 637) argues that: ‘EIAs tend to address the health impacts that can be measured, not necessarily the health impacts that are most significant’. In addition, a focus on the cumulative effects on health and inter-generational health risks is often lacking (Kørnøv 2009; SNBHW 2001; Steinemann 2000).

The health consequences identified here primarily relate to physical illness. Mental, cultural, social or wellbeing-related issues are thus once again either completely ignored or, at best, ineffectively handled. However, considerable variations exist both between and within countries. A study dealing with health aspects in relation to EIA in 10 European countries plus the USA and Canada showed that psychosocial issues (in relation to recreational areas, the state of the local economy and the availability of job opportunities) are commonly featured, while, for example, social capital and cohesion, mental disorders and issues relating to worry and anxiety are rarely assessed (Hilding-Rydevik et al. 2005).

Sweden

In both the study by Alenius (2001) and our review (Sjöberg 2010), comparisons against limit values and norms, as a way of assessing the health impact, are made in all but one document. In our study, it emerged that this is primarily the case when discussing noise and air pollution. In the 2001 study, probability calculations are used to assess the risk of health impacts in 20% of the documents reviewed, while this approach was used in more than 40% of EISs in our study. In the context of our study, it emerged that this is mainly due to the transport of hazardous goods and greater consideration of the impact of vibrations. The study from 2001 emphasized that it is important to look at cumulative health effects, although it did not address how this can be done in the context of EIA. In our study, the discussion of cumulative effects took place only in relation to how vibrations and noise could alter the experiences of noise. Both of these findings agree with the results from another Swedish study, which indicated that the consideration of cumulative effects is often lacking in EIAs (Wärnbäck & Hilding-Rydevik 2009).

Figure 2 demonstrates whether the health impacts of the determinants are described in a clear manner (dark grey). As our study indicates, the health impact is now more often referred to in the context of noise, vibrations and air pollution, with the latter holding the top position, than was previously the case. However, for noise and vibration this is still only the case in 50–75% of all EISs. Of the remaining determinants included in both studies, the health impact seems to be even less well documented and assessed than was the case 10 years ago. With regard to the transport of hazardous goods, our study suggests that less discussion centres on death and injuries as a result of accidents with hazardous goods nowadays, while the characteristics of the goods themselves, that is, their potentially poisonous and explosive nature, has attracted much greater attention, although further directions on how this could affect health are not evident. Outdoor life/recreation and barrier effects can be seen as examples of more social determinants of health. In our study the health impact of outdoor life/recreation was not described in any of the EISs referred to except in relation to specific issues such as noise and barriers. This generated poorer results than in the previous study, where terms such as rest, relaxation and experiencing nature were used in 20% of the EISs surveyed. Similar results can be seen for the determinant barrier effects (Figure 2). In the 2001 study, health impacts were assessed by viewing social effects as changes in security/safety and patterns of contact/communication links. This had not been done in any of the EISs included in our 2010 study.

International findings indicate that the issue of anxiety and worry is more likely to be considered in the context of a Swedish EIA than in those of many other European countries, Canada or the USA, although the reason for this is often not effectively explained (Hilding-Rydevik et al. 2005). However, we did discover the implications of this in the context of our study.

Affected population

International

Baseline data concerning the population are often deficient or lacking in some way in EISs and SEAs, while the distribution of health impacts in the population with emphasis on vulnerable groups such as children, the elderly, socio-economically weak groups, etc., is seldom assessed (Cherp 2002; Douglas et al. 2011; Kørnøv 2009). Such an inadequate focus on vulnerable groups may lead to health inequality. Again, this is something that is seldom properly assessed (Hilding-Rydevik et al. 2005).

Sweden

In the study by Alenius (2001) the population is not even described in 50% of the EISs/SEAs reviewed and in our study from 2010 it was even less common. The same applies for description of the characteristics of the population and representation of groups that are more sensitive to health effects, with the difference here being that mention of these is rather more frequent in the later study. When sensitive groups are discussed, the main concerns are children's traffic safety and the accessibility issues faced by the physically challenged in using infrastructure.

Discussion

The results from our study on how human health is approached in EISs for road planning coincided for the most part with the findings of the previous study by Alenius (2001). The health scope in Swedish EISs is still narrow and mainly focuses on environmental determinants. Comparison against limit values is the most common way of assessing health impacts and there are seldom other descriptions of health impacts for the predicted levels. Baseline data and identification and concern about groups that are more vulnerable to health effects are often lacking. All of these findings are also in line with international findings. However, our review revealed fuller management of health in some EISs, mainly in relation to the approaches taken in the areas of noise and air pollution. For these environmental factors the different calculated levels have been clearly related to human health, thus showing the exposure and health effects such as asthma and death. There is also a slight possibility that some progress has been made but was not detected owing due to the different sampling approaches used in our study and in that by Alenius (2001). Our study used probability sampling of EISs, whereas in the 2001 study developers were asked to choose projects themselves once they had been informed about the topics to be studied. Thus those cases were potentially positively biased towards projects that included health management.

Use of the framing perspective and the ‘health and ill-health’, ‘health determinants’, ‘health impact’ and ‘affected population’ themes makes it possible to distinguish relationships between the framing in the legislation and policies and the choices made in the EISs reviewed. It is clear from the review presented here of Swedish health policy and legislation regarding what should be included in EIA in road planning that several quite different frames are available. However, only some of these are used. In Sweden, the Environmental Code provides a ‘protection against ill-health’ frame, while the Planning and Building Act provides a broader health frame with an aim to ‘plan for health’. The existence of a dividing line between health framed strictly in relation to ill-health or as something more than the absence of disease was also apparent in our review of the international literature (Kågström 2009), where the concept of health in EIA is defined quite differently between countries. However, our interpretation from reading the literature is that the dividing line is seldom expressed in such detail as in the present paper, that is, by identifying different frames in legislation and policies and their implications for specific health choices in EIAs and their correlation with the choices being made in EISs.

The findings from the review of Swedish EISs are most closely connected to the frame provided by the Environmental Code, which is not surprising since this is the main legislation governing EIAs. It is also apparent that the health promotion focus in the Planning and Building Act and the Environmental Quality Objectives is seldom used in EIAs. In relation to other Swedish legislation and policies of relevance for EIAs, the Environmental Code frames health in a way that limits the choice of parameters studied, which can lead to, for example, the exclusion of social health determinants and impacts on the assessment. It also means that differences in the affected population, such as unequal risks of developing ill-health owing to socio-economic and behavioural factors, fall outside the frame. The frame provided by the Planning and Building Act directs attention in another direction, that is, the promotion of health. It also includes a broader scope of health determinants and places some emphasis on health equity.

It is obvious that the introduction of a new health frame through the Public Health Objectives and the promotion of health through physical activity in the Transport Policy Objectives have not had a great impact on the EISs reviewed here. The introduction of the new health frame has increased the space available for EIA practitioners to use different frames when taking decisions about ‘health determinants’, ‘health impact’ and ‘affected population’. The promotion of voluntary HIAs also provides a new frame for health assessment. This might increase the space for using different frames in EIAs, but might also decrease the space. The result may be that certain health frames are only used in HIA, thereby limiting even further the health frames used in EIAs.

Health frames have shifted and will probably continue to shift over time, as is the nature of social constructions. Internationally, there have been parallel developments in framing in the fields of EIA and public health. Over time and in relation to EIA, there has been a broadening of the concept of the environment, from a narrow biophysical focus in the 1970s to, in recent times, a much more expansive view, including cultural, social, economic and health impacts (Steinemann 2000; Glasson et al. 2005). This follows developments in the public health arena, where the term ‘health’ has been expanded from simply including physical health to having a more societal focus, and the development of HIA (first closely linked to EIA) from focusing on health effects from the environment and risk assessments to focusing on broader determinants and health equity (Steinemann 2000; Harris et al. 2009; Harris-Roxas et al. 2012). In the Swedish context, some of the legislation and policies of relevance for EIA have developed in a similar way. However, as our review of Swedish EISs shows, this move towards a broadening of the health concept and emphasis on health equity does not seem to be reflected in EISs, at least not yet.

Internationally and in Sweden, the most common health frames in EIAs still have a rather narrow environmental scope. One major change in EIAs since the situation reported by WHO (1987) is that health issues are generally included to a larger extent than was previously the case. Reviews of health in EIAs have consistently found health to be insufficiently managed in terms of coverage of health determinants and impacts or in relation to the affected population. Reviews of SEAs are slightly more positive, but there are still calls for improvement (Douglas et al. 2011; Kågström 2009). A framing perspective sheds light on how different frames bias for certain expectations, meaning that these kinds of negative results partly depend on the frame used for evaluating EIA practice. It appears as though many studies and evaluations are carried out with the expectation that EIAs should include a rather broad health frame with the focus on health equity. This is partly the case in the present paper, in which this kind of frame forms part of the background used for the analysis. This is not unusual, since these aspects are well recognized in the field of public health. However, in cases where this frame is not found in EIA legislation or EISs, there are often calls for improvement. It should be mentioned that framing in ‘health’ or ‘ill-health’ and their relationship to the other themes discussed in this paper does not answer all previous criticisms mentioned in the literature (such as shortage of methods, skills, money and status). However, using a framing perspective highlights one of the core issues for managing human health in EIA, that is, what health frame to be guided by in order for the EIA to be considered sufficient.

Conclusions

The existence of fundamentally different health frames in Sweden means that EIA practitioners have a large space for interpretation and implementation of them. This implies that important decisions are taken at the micro-level in terms of the health frames that should guide the development of road planning in Sweden. The chosen perspective permeates the whole planning process and provides the basis for decisions taken. Choosing solutions to plan for the promotion of good health is clearly fundamentally different from choosing solutions to plan to protect against ill-health. The fact that decisions in EIA and planning practice are taken by different EIA practitioners at the micro-level is not an anomaly or a new insight, but a crucial issue in relation to the management of human health in EIAs and in planning processes more generally is that human health (however it is framed) is part of what we as humans view as important for a meaningful life. Having good health is an essential quality of life. Promoting good human health also has an economic dimension from a community point of view. The impacts of new roads can affect several health determinants, for example, road traffic and infrastructure might have impacts on many of the public health problems existing in Sweden today. There are also indications that impacts from transportation are generally unequally distributed. Overall, managing human health in a road planning and EIA process is a highly ethical and highly political activity of significant democratic importance. It is thus also a sensitive issue for developers, EIA consultants or decision-makers to deal with in relation to consultation and dialogue with the general public.

This study of legislation and policies of importance for health assessment in EIAs for road planning shows that there has been a broadening of health frames in Sweden to include more than the absence of disease. This is mainly because of the introduction of the Public Health Objectives, but is also partly due to the goals to promote health through physical activity in the Transport Policy Objectives. However, these broader goals seem not to have had a great impact on the EISs reviewed. The very different health frames provided by policy and legislation in Sweden create poor conditions for guiding EIA practice and might become a dilemma for EIA practitioners. The wide space for EIA practice to frame human health also indicates a challenge for the field. Especially if EIA is to function as the main mechanism for providing information on human health impacts in road planning processes.

Acknowledgements

The PhD project on which this paper is based is a co-operative venture between the Department of Urban and Rural Development at the Swedish University of Agricultural Sciences and the consulting firm Tyréns. Tyréns is acknowledged in particular for providing primary funding for the PhD project. Antonis Georgellis, Karolinska Institute, Marianne Klint (formerly Tyréns), Ulla Myhr, Swedish University of Agricultural Sciences and Ulf Wiklund, Tyréns, also deserve special recognition for their wholehearted support of the overall PhD work. In addition, the members of the Spira research group and two anonymous reviewers of an earlier version of this paper are warmly thanked for their useful comments. The views expressed in this article are though those of the authors.