Social disadvantage (broadly construed) is both a cause and consequence of mental illness (see Reference Horwitz and SchiedHorwitz & Schied, 1999); that is, mental illness – in all its forms – is intrinsically social. In recent years social exclusion has emerged as a prominent concept in discussions about disadvantage. Indeed, as an explanatory concept, it has become almost ubiquitous. This is reflected in current efforts to promote social inclusion among those with mental health problems, on the basis that people with long-term mental illness are among the most excluded in society (Social Exclusion Unit, 2004). However, social exclusion is a contested concept, with multiple meanings. If it is to be of value in understanding the social experiences of those with mental health problems, there is a need for fundamental conceptual and methodological work. This paper aims to provide an overview of the current conceptual and methodological literature on social exclusion, focusing in particular on the origins and definitions of the concept, and approaches to its measurement.
Conceptual and methodological literatures are less easily subjected to fully systematic reviews than discrete bodies of research evidence (Reference Lilford, Richardson and StevensLilford et al, 2001). Such work is often disparate and it is questionable whether it needs to be exhaustively reviewed in order to identify all the available definitions and approaches. There are several practical problems, including problems of specifying, in advance, a search strategy; difficulties in extracting relevant material from papers in a consistent and unbiased manner; and issues of how to synthesise material (Reference Lilford, Richardson and StevensLilford et al, 2001). Lilford et al (Reference Lilford, Richardson and Stevens2001) made a number of recommendations to limit these problems, including: not attempting to review all literature, as in a Cochrane-style review, but searching widely (e.g. disparate databases and sources); building in safe-guards to reduce potential biases (e.g. establishing a steering group; using multidisciplinary teams); and allowing some overlap in the various stages of the review process (searching, analysis and writing up), so that the precise nature and scope of the review can be clarified. These recommendations have guided our approach to this review.
We adopted two strategies to provide a ‘way in’ to the potentially vast and disparate literature on social exclusion. First, we utilised expertise within the study team to identify major texts and reviews on social exclusion and related topics. Second, we searched the relevant major bibliographic databases for literature on the topic of social exclusion and mental health, for articles that discussed either the concept and definition of social exclusion, or its measurement. The databases searched were Medline, EMBASE, Web of Science, Psyc-INFO, the International Bibliography of the Social Sciences, the Health Management Information Consortium, Sociological Abstracts and the Institute of Scientific Information (ISI) Proceedings (Social Science and Humanities). For each database we applied the following purposely broad search terms: (social inclusion AND mental) OR (social exclusion AND mental). We also carried out a hand search of the journals Open Mind and Mental Health Today. The titles and abstracts of all identified papers were then reviewed to assess their relevance, and on the basis of this all potentially relevant papers were retrieved. Each paper was then reviewed, and information from all those meeting our broad criteria was extracted using a specially designed form.
For each paper or text identified, we systematically scrutinised reference lists to find new potential references. For each of these, the abstract and if necessary the full paper were read to determine its relevance. We continued until the point at which new papers were not producing novel information, a point akin to that of theoretical saturation in qualitative research. Subsequently, an expert group was asked to comment on the initial draft of this review and to identify any further relevant literature not included.
Data extraction and synthesis
Using a tailored form, we extracted information on the concept and definition of social exclusion, and on approaches to its measurement, from each paper, chapter or report. We adopted a non-quantitative approach to synthesising the findings (Reference Mays, Roberts, Popay, Fulop, Allen and ClarkeMays et al, 2001). Members of the multidisciplinary research team met monthly to assist in determining which literature to include and how to focus the review.
The literature on social exclusion is vast and many of the papers and other texts we scrutinised in our review inform our discussion. In our more systematic search for literature on mental health and social exclusion, 232 references were identified. Of these, 50 were deemed potentially relevant following title and abstract sifts and were subsequently read, with 12 meeting our fairly broad inclusion criteria (Reference DunnDunn, 1999; Reference BatesBates, 2002; Reference Bonner, Barr and HoskinsBonner et al, 2002; Reference NashNash, 2002; Reference Targosz, Bebbington and LewisTargosz et al, 2003; Reference Hjern, Wicks and DalmanHjern et al, 2004; Reference Parr, Philo and BurnsParr et al, 2004; Social Exclusion Unit, 2004; Reference Todd, Green and HarrisonTodd et al, 2004; Reference Webber and HuxleyWebber & Huxley, 2004; Reference Fakhoury and PriebeFakhoury & Priebe, 2006; Reference Payne, Pantazis, Gordon and LevitasPayne, 2006) (Fig. 1). Key information from these references is summarised in Data Supplement 1 to the online version of this paper.
Increase in citing of social exclusion
As usage of the concept of social exclusion has increased, so all forms of social differentiation studied have tended to be adapted to it. That is, studies of specific social variables (unemployment, housing, income, education and so on) are increasingly reported as studies of social exclusion; studies of interventions designed to improve these aspects of people's lives are styled as programmes to promote social inclusion. Despite this, social exclusion is rarely defined and other dimensions of these concepts are usually not considered. Consequently, our broad search strategy identified many studies, the majority of which did not directly investigate social exclusion (as a multidimensional concept) and mental health.
There are large bodies of research on specific variables that are currently conceptualised as indicating social exclusion. For example, the research on all aspects of employment and all forms of mental distress is extensive, and many of the findings are of central relevance to ongoing efforts to understand the relationship between labour market involvement and social inclusion. The Social Exclusion Unit (2004) report on social exclusion and mental illness summarised key findings relevant to each of the dimensions of social exclusion identified within it, including stigma and discrimination, employment, education and housing. There are also several reviews of specific aspects of social exclusion and mental illness, such as unemployment (e.g. Reference WarnerWarner, 2003; Reference Leff and WarnerLeff & Warner, 2006) and stigma (Reference Leff and WarnerLeff & Warner, 2006; Reference ThornicroftThornicroft, 2006). The extensive literature in each of the domains is not explored here.
Our purpose was to review literature relating to the concept and measurement of social exclusion, both in general and in relation to mental health. We have organised the findings from our review into three parts: origins, definitions and measurement of social exclusion.
Origins of social exclusion
Within the context of European social policy, most commentators locate the origins of the modern conception of social exclusion in the work of Lenoir (Reference Lenoir1974) (e.g. Reference Levitas, Pantazis, Gordon and LevitasLevitas, 2006; Reference Silver and MillerSilver & Miller, 2003). Lenoir used the term ‘les exclus’ to refer to those who, in the 1970s, fell through the social insurance system safety net, for example lone parents and the uninsured unemployed. Within France, the term expanded to encompass more groups of people on the margins of society, and came to denote a ‘rupture of the social bond’ considered central to the social contract between the state and its citizens (Reference Silver and MillerSilver & Miller, 2003). Concern with social exclusion, and strategies to promote social inclusion, were key components of French social policy through the 1980s and, under the presidency of Jacques Delors, began to influence European Commission policy.
In the UK the concept originated in the critical social policy of the 1980s (Reference Levitas, Pantazis, Gordon and LevitasLevitas, 2006), particularly in the work of Peter Townsend. In the ongoing debate about how poverty should be defined and measured, Townsend (Reference Townsend1979) developed a broad definition of poverty as relative deprivation. This went beyond material deprivation and incorporated the notion of participation in the customary activities of society:
‘Individuals, families and groups in the population can be said to be be in poverty when… [their] resources are so seriously below those commanded by the average individual or family that they are, in effect, excluded from the ordinary living patterns, customs and activities’ (Reference TownsendTownsend, 1979: p. 32; our italics).
Social exclusion was increasingly used to capture the consequences of material deprivation in terms of restricted opportunities to participate in wider social and cultural activities (Reference Levitas, Pantazis, Gordon and LevitasLevitas, 2006). Paradoxically, many commentators (e.g. Reference ByrneByrne, 2005) have argued that the notion of social exclusion gained currency in UK government and policy circles during the 1980s and 1990s because it allowed the less politically acceptable language of poverty to be removed from policy debates. The precise meaning of social exclusion was not clarified in this context, allowing the development of different ideas about its core features.
Concepts and definitions of social exclusion: general
There are many definitions of social exclusion in the broader literature with clear differences in emphasis (Reference Vranken and GeldofVranken & Geldof, 1992: p. 19; Reference DuffyDuffy, 1995; Reference RoomRoom, 1995; Reference Walker and RoomWalker, 1995: p. 103; Social Exclusion Unit, 1997; Reference Walker and WalkerWalker & Walker, 1997: p. 8; Reference GiddensGiddens, 1998: p. 104; Reference Mandanipour, Cars and AllenMandanipour et al, 1998: p. 22; Reference Berman and PhillipsBerman & Phillips, 2000: p. 330; Reference SayceSayce, 2001: p. 122; Reference Vleminckx, Berghman, Mayes, Berghman and SalaisVleminckx & Berghman, 2001: p. 46; Reference Barry, Hills, Le Grand and PiachaudBarry, 2002: pp. 14–15; Reference Burchardt, Le Grand, Piachaud, Hills, Le Grand and PiachaudBurchardt et al, 2002: pp. 30, 32; Reference Vranken, Anderson and van KempenVranken, 2001: p. 86; Reference Dewilde and De KeulenaerDewilde & De Keulenaer, 2003; Reference EstivilEstivil, 2003: p. 19; Reference Repper and PerkinsRepper & Perkins, 2003; Reference CollinsCollins, 2004: p. 727; Reference Parr, Philo and BurnsParr et al, 2004: p. 405; Reference ByrneByrne, 2005: p. 81; these definitions are quoted in Data Supplement 2 to the online version of this paper). Nevertheless, Burchardt (Reference Burchardt and Davies2000: p. 320) has argued that lack of participation in mainstream social, cultural, economic and political activities is the primary element at the core of most definitions, and most share an emphasis on multiple dimensions of exclusion (e.g. low income, poor housing and isolation) on the dynamic nature of exclusion (i.e. people's level of participation will vary over time) and on the multilevel causes of exclusion (i.e. at the level of individual, household, community and institutions).
However, a number of questions remain. Despite general agreement that social exclusion is multidimensional, there is no consensus on which dimensions are relevant, which if any are the most important, and whether multiple and cumulative disadvantage is necessary or whether one of a range of deprivations is sufficient. In other words, the formulation lacks precision. It also remains unclear just what it is to be socially excluded. Is it an objective state or a subjectively felt experience? Some definitions suggest an objective state:
‘an individual is socially excluded if he or she does not participate’ (Reference Burchardt, Le Grand, Piachaud, Hills, Le Grand and PiachaudBurchardt et al, 2002: p. 30).
Others focus on subjective experience:
‘Inclusion denotes relations and practices that people with mental health problems perceive to signify their positive involvement in and “mattering” to a local setting’ (Reference Parr, Philo and BurnsParr et al, 2004: p. 405).
There are other issues. Some people choose not to participate in wider society, but are they socially excluded? Barry suggests:
‘individuals or groups are socially excluded if they are denied the opportunity of participation, whether they actually desire to participate or not’ (Reference Barry, Hills, Le Grand and PiachaudBarry, 2002: p. 16).
The point is that apparent voluntary exclusion may be a function of restricted opportunities to participate or a response to the experience of discrimination and this points to a definition of social exclusion as ‘enforced lack of participation’.
Social exclusion clearly overlaps with other concepts. For example, some definitions of poverty appear indistinguishable from those of social exclusion. The 1995 Copenhagen World Summit on Social Development adopted a definition of poverty that included ‘unsafe environments and social discrimination and exclusion’ (United Nations, 1995: p. 57). Others appear to treat poverty and social exclusion as synonymous:
‘Poverty and social exclusion are concerned with a lack of possessions, or an inability to do things that are considered normal by society’ (Reference Howarth, Kenway and PalmerHowarth et al, 1998: p. 18).
‘This broad concept of poverty coincides with the emerging concept of social exclusion’ (Reference Howarth, Kenway and PalmerHowarth et al, 1998, p. 13).
Silver & Miller have argued that social exclusion offers a broader, more holistic understanding of deprivation, in contrast to poverty, which they see as ‘exclusively economic, material, or resource-based’ (Reference Silver and MillerSilver & Miller, 2003: p. 8). This is particularly relevant to mental illness. The loss of roles, meaningful relationships and discrimination that both precede and accompany mental illness do not necessarily stem from a lack of material resources. Negative societal attitudes and responses towards those with a mental illness powerfully affect their social experiences and often underpin social rejection and isolation (Reference Link and PhelanLink & Phelan, 2004).
In most definitions of social exclusion, social relationships and networks are a central component, a key requirement for a fully participative and inclusive life. There is overlap here with other concepts in which social relationships are integral, most noticeably social capital (also a complex and contested concept, with multiple definitions). At the core of all concepts of social capital is the idea that networks of social relationships are a potentially valuable resource that people can draw on, and as such constitute a form of capital (Reference FieldField, 2003). Putnam's concept is the most influential current formulation of social capital in relation to mental health (see Reference McKenzie and HarphamMcKenzie & Harpham, 2006). Putnam has defined social capital as:
‘features of social life – networks, norms and trust – that enable participants to act together more effectively to pursue shared objectives’ (Putnam, (Reference PutnamPutnam, 1996: p. 56)
‘connections among individuals – social networks and the norms of reciprocity and trust-worthiness that arise from them’ (Reference PutnamPutnam, 2000: p. 19).
These connections, and associated trust and reciprocity, that form social capital are viewed as the ‘glue’ that holds societies together. There are clear overlaps between the concepts of social capital and social exclusion in that both focus, to varying degrees, on participation in social networks. Social exclusion is a broader concept. In simple terms, access to social capital is about access to valuable contacts within a social network, and, as such, social capital potentially promotes social inclusion by tying people to others in the wider community. Lack of social capital may contribute to social exclusion.
Social exclusion, social capital and, to a lesser degree, poverty also overlap with other concepts that have been used in attempts to study social processes, e.g. social networks, social integration and fragmentation. It is beyond the scope of this review to unpick these different concepts. However, the range of overlapping notions begs the question of which are most useful in making sense of the social experiences of those at the lowest end of the current social order. A question that must continue to hang over the concept of social exclusion is whether it is an advance on previous concepts, and whether its heuristic potential is greater than what has gone before.
Concepts and definitions of social exclusion: mental health
The lack of clarity in the general literature about the precise nature of social exclusion, how it is defined and how it relates to other key concepts is replicated in the mental health literature. The majority of studies of mental health identified in this review either do not provide any definition of social exclusion or rely, largely uncritically, on one or more of a range of previous definitions.
Despite purporting to measure social exclusion, none of the following studies provides a definition: Shimitras et al (Reference Shimitras, Fossey and Harvey2003); Targosz et al (Reference Targosz, Bebbington and Lewis2003); Hjern et al (Reference Hjern, Wicks and Dalman2004); Todd et al (Reference Todd, Green and Harrison2004); Fakhoury & Priebe (Reference Fakhoury and Priebe2006). This leads to confusion about what precisely is being measured. Both Todd et al (Reference Todd, Green and Harrison2004) and Targosz et al (Reference Targosz, Bebbington and Lewis2003) used a series of markers of social disadvantage and differentiation (e.g. social class, employment status, contact with the legal system) without consideration of whether these variables do in fact reflect components of social exclusion. Hjern et al (Reference Hjern, Wicks and Dalman2004) used the terms ‘social adversity’ and ‘social exclusion’ interchangeably, the implication being that the two are synonymous. The remaining studies relied on one or more existing definitions, most commonly that provided by the Social Exclusion Unit (1997), for example Nash (Reference Nash2002), Byrne et al (Reference Byrne, Jones and Williams2004).
A small number of papers demonstrated a more critical approach to conceptualising social exclusion, particularly in terms of its relevance to understanding the social experiences of people with mental health problems. Sayce (Reference Sayce2001) drew from previous definitions, but extended these to relate more directly to people with mental health problems, focusing attention specifically on the impact of both impairment and societal responses. She argued that for mental illness social exclusion has more explanatory power than poverty or related concepts, in that it focuses attention on the non-material disadvantages that result from the discriminatory responses of others and institutions. Sayce (Reference Sayce2000, Reference Sayce2001) linked this with the social model of disability, arguing that many of the apparent social impairments experienced by those with mental health problems are a function of societal responses. From this perspective, the social inclusion of people with mental health problems can be achieved only when society changes. Here the focus is on those doing the excluding rather than on the excluded, a perspective further evident in the work of Repper & Perkins (Reference Repper and Perkins2003), who considered social reintegration to be a key component of recovery from mental health problems.
Sayce (Reference Sayce2000) and Repper & Perkins (Reference Repper and Perkins2003) emphasised two key aspects of social exclusion: process and agency. Parr et al (Reference Parr, Philo and Burns2004) emphasised similar themes in their account of the social exclusion and inclusion of people with mental health problems in the Highlands of Scotland, defining the concepts of inclusion and exclusion as follows:
“‘Inclusion” denotes relations and practices that people with mental health problems perceive to signify their positive involvement in and “mattering” to a local setting… By contrast, “exclusion” denotes more negative eventualities that involve rejection, avoidance and distancing from other community members, such that individuals are “made different” through more or less deliberate social actions reinforcing their problematic mental health status’ (Reference Parr, Philo and BurnsParr et al, 2004: p. 405).
This perspective conceptualises exclusion and inclusion as subjective states of belonging and involvement in local communities determined by the actions of others in the immediate social milieu. Individuals, in this account, feel more or less included at different times; being excluded is not a static fixed state that can be objectively measured, but a fluid process:
‘The lines between inclusion and exclusion turn out to be quite blurred, particularly in that superficially inclusionary moments cannot be taken as evidence of a deep-seated inclusionary tendency’ (Reference Parr, Philo and BurnsParr et al, 2004: p. 405).
This implies a very different approach to studying social exclusion from that used in most studies of social exclusion and mental health to date. Further, this approach shares much with ethnographic studies that have documented the lived experience of mental illness (e.g. Reference EstroffEstroff, 1981; Reference Jenkins and BarrettJenkins & Barrett, 2003). Although such work has not been conducted within the framework of social exclusion, it reveals much about how societal responses to long-term mental illness contribute to excluding people with such disorders from social activities and public spaces. It is further notable in this context that innovative research in the USA designed to refocus attention on social reintegration (a concept with clear overlaps with social exclusion) is being driven by medical anthropologists (e.g. Reference Ware, Hopper and TugenbergWare et al, 2007).
Measuring social exclusion
The most common approach to measuring social exclusion in general is the use of lists of multiple indicators of exclusion, with data usually drawn from pre-existing data-sets. This is how the initiatives to promote social inclusion among those with mental health problems are currently measured, as set out in the 2004 Social Exclusion Unit report (Social Exclusion Unit, 2004). There are problems with this approach. The New Policy Institute Report Monitoring Poverty and Social Exclusion: Labour's Inheritance (Reference Howarth, Kenway and PalmerHowarth et al, 1998) specifies 50 indicators and Opportunity for All: Tackling Poverty and Social Exclusion (Department of Social Security, 1999) specifies 32. However, both of these conflate poverty and social exclusion, using the terms interchangeably. It is consequently not clear which are indicators of poverty primarily and which of social exclusion. Such conceptual uncertainty makes it difficult to interpret these indicators. Levitas (Reference Levitas, Pantazis, Gordon and Levitas2006) further argues that most indicator lists lack the very social dimension that is unique to social exclusion and distinguishes it from poverty. Few indicators relate directly to participation in social and cultural life, making it unclear how these lists differ from measures of multiple deprivation and poverty.
A further important issue is the lack of a distinction between direct and indirect indicators and risk factors. Is unemployment to be considered a risk factor for social exclusion, an indirect indicator, or a direct measure? What about low income? Such distinctions are required if social exclusion is to be more precisely defined, its prevalence documented and its causes understood. In the Social Exclusion Unit report, stigma is identified as a core domain of exclusion, and improvement in public attitudes is used as an indicator of social inclusion (Social Exclusion Unit, 2004). It might be more logical, however, to consider stigma and discrimination as causes of exclusion rather than as core defining features.
The Centre for the Analysis of Social Exclusion at the London School of Economics adopts a dimensional approach. They identify four dimensions of social exclusion, lack of participation in each one being seen as sufficient in its own right to constitute social exclusion (Reference Burchardt, Le Grand, Piachaud, Hills, Le Grand and PiachaudBurchardt et al, 2002: p. 31):
(a) consumption: the capacity to purchase goods and services;
(b) production: participation in economically or socially valuable activities;
(c) political engagement: involvement in local or national decision-making;
(d) social interaction: integration with family, friends and community.
These activities are operationalised in terms of data available from the British Household Panel Survey, which has the advantage of allowing indicators to be tracked over time, capturing the dynamic nature of social exclusion. This approach has been criticised partly because of a lack of clarity about what constitutes ‘key activities’ (Reference Levitas, Pantazis, Gordon and LevitasLevitas, 2006).
Some response to these criticisms is found in the consensual approach used in the Poverty and Social Exclusion (PSE) Survey, conducted in 1999 (Reference Pantazis, Gordon and LevitasPantazis et al, 2006). This survey sought to measure social exclusion directly across four dimensions: impoverishment or exclusion from adequate resources or income, labour market exclusion, service exclusion and exclusion from social relations. Pantazis et al argue: ‘Uniquely, the PSE Survey treats exclusion from social relations as a constitutive aspect of social exclusion’ (p. 8). Members of the public were asked to define the items and activities they considered necessary for everyone in Britain to reach a minimum standard of living and inclusion. Items that 50% or more of the population considered necessary were included (Reference Pantazis, Gordon and LevitasPantazis et al, 2006: p. 9).
Mental health literature
Eight of the studies of mental health and social exclusion we identified attempted to quantify social exclusion using indicators across a number of domains or dimensions (Reference Bonner, Barr and HoskinsBonner et al, 2002; Reference Shimitras, Fossey and HarveyShimitras et al, 2003; Reference Targosz, Bebbington and LewisTargosz et al, 2003; Reference Hjern, Wicks and DalmanHjern et al, 2004; Reference Todd, Green and HarrisonTodd et al, 2004; Reference Webber and HuxleyWebber & Huxley, 2004; Reference Fakhoury and PriebeFakhoury & Priebe, 2006; Reference Payne, Pantazis, Gordon and LevitasPayne, 2006). The study by Payne (Reference Payne, Pantazis, Gordon and Levitas2006) was part of the PSE Survey. None of the other studies used a questionnaire designed specifically to measure social exclusion. Data on indicators such as employment and housing) were obtained from case records or registers using data extraction forms designed for the study (e.g. Reference Webber and HuxleyWebber & Huxley, 2004); data on specific domains of exclusion were collated from questionnaires initially designed to measure another concept, such as needs or life satisfaction (e.g. Reference Bonner, Barr and HoskinsBonner et al, 2002). The domains and indicators considered were chosen on the basis of what data were available. For example, Webber & Huxley (Reference Webber and Huxley2004); their study of social exclusion and pathways to mental healthcare, acknowledged that their choice of indicators (housing, education, income, employment, social support and neighbourhood deprivation) was mainly determined by what could be reliably extracted from case-notes, and important potential dimensions such as stigma, discrimination and access to services and goods were not included. Similarly, the Swedish population-register study by Hjern et al (Reference Hjern, Wicks and Dalman2004) of social adversity and psychosis among migrant groups was limited to those indicators already available (i.e. area of residence, housing, social class, single adult household, employment, receipt of welfare benefits and immigrant status). These studies covered a mix of indicators, many of which are indistinguishable from measures of socio-economic adversity. Todd et al (Reference Todd, Green and Harrison2004) and Targosz et al (Reference Targosz, Bebbington and Lewis2003) used social exclusion and social adversity/disadvantage interchangeably. These studies did, however, attempt to capture the multidimensional component of social exclusion by using a number of markers across different domains.
Dunn (Reference Dunn1999) and Parr et al (Reference Parr, Philo and Burns2004) used qualitative methods to explore the relationship between mental health problems and social exclusion. In the most influential study, Parr et al argued that there is a need to look beyond indicators to the
‘experiential processes… leading particular individuals and groupings to be excluded from the norms of everyday social life, activity and participation’ (Reference Parr, Philo and BurnsParr et al, 2004: p. 47).
They conducted in-depth interviews with over 100 users of mental health services in rural Scotland to explore experiences of inclusion and exclusion in everyday life. What emerges is a complex picture in which feelings of being included or excluded are subject to constant change depending on ongoing interactions with others in both public and private spheres. Subtle perceptions of being ignored or befriended compound or mitigate feelings of exclusion in a continual process, such that inclusion and exclusion can co-occur; individuals can experience elements of both simultaneously. In this account, social exclusion is subjectively experienced, relative and changeable; it is not a state that can be measured. It further emphasises the spatial element of social exclusion. The actions of others in particular places can serve to exclude people with mental health problems from public spaces.
This review set out to provide an overview of the origins and definitions of the concept of social exclusion, and approaches to its measurement. It was not our intention to attempt a complete and systematic review of the literature on social exclusion and inclusion; it is not obvious that this is either possible or necessary. Rather, we sought a comprehensive review of relevant theoretical and conceptual frameworks, and approaches to the measurement of social exclusion.
Many of the inevitable limitations to our search strategy are well documented (Reference Lilford, Richardson and StevensLilford et al, 2001). The extensive initial reliance on the expertise and interests of the research group may bias the review's focus. As a consequence of this, important publications (particularly institutional reports) may be missed if the research group are not aware of them or they are not referenced in bibliographic databases. Deciding what information is relevant is largely a matter for the individual conducting the review (with some checks from the wider research group), limiting the replicability of the review. Another research team following the same procedures might produce different findings. These limitations should temper any conclusions drawn from this review.
It is widely acknowledged that ‘social exclusion’ is a contested term that defies easy definition. There have been numerous attempts to define social exclusion, each one having a slightly different emphasis, and each one underpinned by slightly different philosophical perspectives. This conceptual confusion stems from social exclusion being primarily a political term, originating in the social policy discussions of the European Commission in the 1980s and 1990s. It served, according to some (Reference Burchardt and DaviesBurchardt, 2000), as a more neutral term than ‘poverty’ for discussing the problems posed and faced by those at the bottom of the social order. It may be precisely this feature of social exclusion (its vagueness, with its multiple meanings and connotations) that makes it so useful in the world of politics. However, to document and understand its role in social experience much greater precision is needed.
Is social exclusion in any way distinct from other related concepts, such as poverty or social capital? There are clear overlaps, and the inclusion by Townsend (Reference Townsend1979) of lack of participation in certain activities in his definition of relative deprivation blurs the boundary between poverty and social exclusion. Many subsequent researchers have more or less conflated the two concepts. However, exclusion from participation in particular aspects of society does not always stem from material disadvantage. This is particularly clear for people with mental health problems, whose exclusion is frequently a result societal stigma (independent of material wealth). Social exclusion is a broader concept than poverty, and, as Sayce (Reference Sayce2001) among others argues, potentially of more use in understanding the social experiences of people with mental health problems. The concept of social capital focuses attention on the value that can derive from social contacts and networks, and the trust and reciprocity that inhere within them. As such it overlaps with a central component of social exclusion: participation in social networks. Once again, social exclusion is arguably a broader concept, with limited social capital being one potential cause of such exclusion.
Measurement: future challenges
Lack of conceptual clarity poses problems in measuring social exclusion. In the wider literature this remains an issue, and attempts to develop direct measures of social exclusion are in their infancy. If social exclusion is a useful concept for understanding the social experiences of those with mental health problems, the development of valid and reliable measures is urgently needed. On the basis of the literature reviewed here, we favour a definition of social exclusion as ‘enforced lack of participation’ in key social, cultural and political activities, as proposed by Burchardt (Reference Burchardt and Davies2000). As a precursor to measurement, this has a number of advantages over other definitions. First, ‘participation’ can be measured directly and quantified in terms of frequency and duration. Second, these components of ‘participation’ can be mapped over time, thereby capturing the dynamic nature of exclusion. Third, there is flexibility to incorporate more subjective aspects of ‘participation’, such as the perceived quality of social relationships arising from involvement in, say, leisure activities. Fourth, this more precise definition allows for greater clarity in distinguishing direct and indirect indicators of exclusion, and risk factors. For example, in this definition, stigma is a risk factor for social exclusion in that it can create a barrier to participation. This definition, therefore, provides a solid conceptual basis from which to develop a measure of social exclusion in which social, cultural and political participation is central and in which both objective indicators and subjective experiences are included. Such a measure, if brief, could be used in intervention studies and in evaluating routine clinical outcomes. These observations point to further challenges for future work. For example, there is a need to establish just what the core activities are for a fully participative life, and clear distinctions need to be made between direct and indirect indicators of exclusion and risk factors.
The definition and measurement of social exclusion and inclusion are not simply academic matters. As Sayce (Reference Sayce2000) and Repper & Perkins (Reference Repper and Perkins2003) argue, social exclusion has considerable potential value in improving our understanding of the lived experiences of people with mental health problems. Kingdon et al (Reference Kingdon, Ramon and Perkins2005) have provided a useful outline of practical steps that services and clinicians can take to promote social inclusion. None the less, while the meaning of social exclusion remains vague and contested, the potential utility of the concept will be undermined. Interventions designed to promote social inclusion need clear guiding definitions and goals if they are to be evaluated and subsequently integrated into routine clinical care. In other words, if policy rhetoric is to be transformed into meaningful and effective interventions, there is a pressing need for further conceptual and methodological work. Without this, clinicians may well be left wondering whether social exclusion, as a framework for understanding the social needs of patients and guiding interventions, is of any more value than what has gone before.
We thank participants of an expert group who commented on an initial draft of this review: George Smith, James Nazroo, Trudy Harpham, Mary De Silva, Jane Boydell and Tania Fisher. We would also like to thank Naomi Lewis for help retrieving the literature. This review was funded by a grant from the National Health Service Research and Development Methodology Programme.