Understanding and representing ‘place’ in health research: A relational approach
Introduction
The idea that ‘context’ matters for individual health is not new and has its roots in the holistic, Hippocratic tradition of medicine (Macintyre & Ellaway, 2003; Meade & Earickson, 2002). However, it is particularly since the early 1990s that we have seen a considerable expansion of theoretical and empirical work investigating the role of contextual factors in the production and maintenance of health variation. Geographers and sociologists have long argued that place is relevant for health variation because it constitutes as well as contains social relations and physical resources (Jones & Moon, 1993; Kearns, 1993; Kearns & Joseph, 1993; Macintyre, McIver, & Sooman, 1993). This re-engagement with the idea that ‘place’, as well as the characteristics of individuals, contributes to health variation appeared to arise simultaneously in geography, sociology and epidemiology and researchers within each of these disciplines have since made significant contributions to the development of knowledge in this field (see also Curtis & Rees-Jones, 1998; Diez-Roux (1998), Diez-Roux (2000); Macintyre, Ellaway, & Cummins, 2002 for further reviews). In this paper we aim to highlight selected examples of empirical research investigating associations between place and health that, implicitly or explicitly, incorporate ‘relational’ views of context and space. Relational conceptions of space and place have recently emerged in theoretical discussions in geography and other disciplines and here we argue that further development of these theoretical approaches could help to improve the empirical evidence about how ‘place’ matters for health.
A significant amount of empirical research on health and place has applied extensive, quantitative methods and techniques for statistical modelling of a general ‘contextual’ effect (usually of deprivation) on the health of populations (reviewed, for example in: Chaix, Merlo, & Chauvin, 2005; Chaix, Rosvall, Lynch, & Merlo, in press; Duncan & Jones, 1993; Diez-Roux (1998), Diez-Roux (2004); Raudenbush & Sampson, 1999; Sampson, Morenoff, & Gannon-Rowley, 2002; Subramanian, Jones, & Duncan, 2003). However recent advances in ‘place-based’ health research have focused on the articulation and development of plausible conceptual models of the causal pathways by which ‘place’ (especially place of residence) may influence health and there has been a growing emphasis on the importance of establishing empirical evidence to substantiate these theories, partly in response to the drive to make public health policy more ‘evidence based’. Some recent studies have used qualitative methods to research individuals’ experience and perceptions of place and what these mean for health (for example, Airey, 2003; Berkman, Glass, Brissette, & Seeman, 2000; Frohlich, Corin, & Potvin, 2001; Frohlich, Potvin, Chabot, & Corin, 2002; Popay, Williams, Thomas, & Gatrell, 1998; Popay, Thomas, Williams, Bennett, Gatrell, & Bostock, 2003). These qualitative studies are valuable because they provide insights that show us how conditions in particular places are thought to influence health and health related behaviour, and they are powerfully suggestive of causal pathways relating environmental factors to individual health. Quantitative studies are also now beginning to take such an approach by testing hypotheses regarding how specific features of places (e.g. features of the built or social environments) are related to relevant health-related outcomes (for example Addy et al., 2004; Giles-Corti & Donovan, 2003). This research is valuable because in order to design policies that improve public health we need to be able to estimate the magnitude of such relationships and understand how far relationships between health and places are generalizable (or variable) across whole populations. In addition, understanding the specific mechanisms through which places affect health, as well as quantifying their impact, is important not only for strengthening causal inferences but also for identifying potential avenues for intervention.
Much of the early research concentrated on exploring whether places do indeed ‘matter’ for health variation and the extent to which they produce significant health inequalities. In retrospect, this analytical focus has had the unintended consequence of constructing places and people (or ‘context’ and ‘composition’) as mutually exclusive and competing explanations for health inequality (Macintyre et al., 2002). Although there are a number of notable exceptions, which we discuss below, we tend to concur with Smith and Easterlow (2005) that empirical research has often been limited because it has tended to rely too much on rather conventional representations of space and place. These aspects may explain why, for example, the majority of quantitative studies find only a small proportion of health variation attributable to ‘context’ when compared to conventional, individual level risk factors (Pickett & Pearl, 2001) and why, despite significant advances in research on health and place, the empirical evidence on what specific aspects of context matter for which health outcomes remains relatively weak. It may be argued (for example Carpiano, 2006) that better theoretical frameworks in health research need to be tested through empirical research, before robust ‘contextual’ interventions to improve health can be designed and implemented. We argue below that research which employs ideas about ‘relational geographies’ may help to provide us with this evidence. Our discussion here elaborates first on the key theoretical aspects of a ‘relational’ geographical perspective and then considers how this might be ‘operationalized’ through research methodologies, particularly in the field of quantitative research, where, we argue, a ‘relational’ perspective might bring the most benefit.
Section snippets
Key dimensions of relational geographies
This paper does not aim to propose major new theories about how we should understand and represent relational aspects of place and space; we are more concerned with discussing how these ideas have been, and could be, further applied in research on health inequalities. Fig. 1 briefly summarises some key differences between ‘relational’ and ‘conventional’ views of place, and we accept that, for illustrative purposes, it may over-emphasize the ‘extremes’ of these different perspectives in order to
Moving beyond context and composition: reconnecting people and place
In an earlier paper, two of the current authors asked whether it was still useful to establish the relative importance of context over composition for health (Macintyre et al., 2002). This earlier paper chiefly urged researchers not to inadvertently control for, or overlook the intervening variables which might mediate the causal pathways between place and health and highlighted the lack of pre-specified a priori theories of contextual causal mechanisms in epidemiological models. In addition,
More than ‘distance to resources’: locating context in time and space
We have argued above that conventional arguments about the relationship between distance and environmental effects on populations are beginning to be revised in the light of a relational perspective on place. This implies that we have to be much more careful in thinking how ‘context’ should be measured and move towards a more flexible and fluid approach to exposure assessment.
Matthews, Detwiler, and Burton's (2005) geo-ethnographic studies illustrate that access to resources for the maintenance
Spatial scale and the production of unhealthy places
A relational perspective allows place to be defined as a result of endogenous and exogenous processes operating at a variety of spatial scales. We can conceive, for example, that trends in regional economies, national and regional environmental pollution, national or supra-national regulatory policies and the action of trans-national organizations and entities can all define the ‘local’ and other contexts in differing ways and this in turn contributes to the spatial distribution of health
Development of relational geographies: the example of food consumption and diet
In this part of the paper we consider how the perspectives that we have discussed above can fit together to enhance our understanding of the ways that places relate to health. One illustration is provided by studies of place effects on food consumption and diet. Contextual effects on diet based on the ‘food desert’ model have been chiefly conceived as the product of two related pathways concerned with neighbourhood physical accessibility to food: access to foods for home consumption from
Conclusion
The recent resurgence of interest in place and health has been mainly based on a traditional, Euclidian conception of space and place and this may be one reason why the relative magnitude of risk ascribed to ‘place’ is limited when compared to individual-level factors. In this paper we have proposed an alternative view, utilizing a ‘relational’ perspective, which might offer a deeper understanding of how ‘place’ affects population health.
In particular, we have suggested that if we are to
Acknowledgements
This paper was written while Steven Cummins was a Visiting Scholar at the Centre for Social Epidemiology & Population Health at the University of Michigan. He would like to thank them for hosting his visit, and allowing him to attend lectures, meetings and classes funded by the Robert Wood Johnson Foundation Health & Society Scholars Program. Steven Cummins is supported by a UK Medical Research Council Special Fellowship in Health of the Public. Sally Macintyre is also funded by the MRC. Ana V.
References (100)
Nae as nice a scheme as it used to be: lay accounts of neighbourhood incivilities and well-being
Health & Place
(2003)- et al.
From social integration to health: Durkheim in the new millennium
Social Science & Medicine
(2000) - et al.
The effect on morbidity of variability in deprivation and population stability in England and Wales: An investigation at small area level
Social Science & Medicine
(1999) Towards a neighbourhood resource based theory of social capital for health; can Bourdieu and sociology help?
Social Science & Medicine
(2006)Differential geographies: Place, indigenous rights and ‘local’ resources
Political Geography
(2004)Landscape, care and the relational self: Therapeutic encounters in rural England
Health & Place
(2005)- et al.
A methodology for measuring neighbourhood social and material context: Generation and interpretation using routine and non-routine data
Health & Place
(2005) - et al.
McDonald's restaurants and neighbourhood deprivation in Scotland & England
American Journal of Preventive Medicine
(2005) - et al.
The ecological relationship between deprivation, social isolation and rates of hospital admission for acute psychiatric care: A comparison of London and New York City
Health and Place
(2006) - et al.
Area effects on health variation over the life-course: Analysis of the longitudinal study sample in England using new data on area of residence in childhood
Social Science & Medicine
(2004)