Theories of practice and public health: understanding (un)healthy practices

Psychological understandings and individualistic theories of human behaviour and behaviour change have dominated both academic research and interventions at the ‘coalface’ of public health. Meanwhile, efforts to understand persistent inequalities in health point to structural factors, but fail to show exactly how these translate into the daily lives (and hence health) of different sectors of the population. In this paper, we suggest that social theories of practice provide an alternative paradigm to both approaches, informing significantly new ways of conceptualising and responding to some of the most pressing contemporary challenges in public health. We introduce and discuss the relevance of such an approach with reference to tobacco smoking, focusing on the life course of smoking as a practice, rather than on the characteristics of individual smokers or on broad social determinants of health. This move forces us to consider the material and symbolic elements of which smoking is comprised, and to follow the ways in which these elements have changed over time. Some of these developments have to do with the relation between smoking and other practices such as drinking alcohol, relaxing and socialising. We suggest that intervening in the future of smoking depends, in part, on understanding the nature of these alliances, and how sets of practices co-evolve. We conclude by reflecting on the implications of taking social practices as the central focus of public health policy, commenting on the benefits of such a paradigmatic turn, and on the challenges that this presents for established methods, policies and programmes.

Psychological understandings and individualistic theories of human behaviour and behaviour change have dominated both academic research and interventions at the 'coalface' of public health. Meanwhile, efforts to understand persistent inequalities in health point to structural factors, but fail to show exactly how these translate into the daily lives (and hence health) of different sectors of the population. In this paper, we suggest that social theories of practice provide an alternative paradigm to both approaches, informing significantly new ways of conceptualising and responding to some of the most pressing contemporary challenges in public health. We introduce and discuss the relevance of such an approach with reference to tobacco smoking, focusing on the life course of smoking as a practice, rather than on the characteristics of individual smokers or on broad social determinants of health. This move forces us to consider the material and symbolic elements of which smoking is comprised, and to follow the ways in which these elements have changed over time. Some of these developments have to do with the relation between smoking and other practices such as drinking alcohol, relaxing and socialising. We suggest that intervening in the future of smoking depends, in part, on understanding the nature of these alliances, and how sets of practices co-evolve. We conclude by reflecting on the implications of taking social practices as the central focus of public health policy, commenting on the benefits of such a paradigmatic turn, and on the challenges that this presents for established methods, policies and programmes.
Keywords: social practice; public health; behaviour change; smoking Introduction: behaviour change paradigms in public health The epidemics of non-communicable disease (NCD) resulting from smoking, alcohol consumption, low levels of physical activity and obesity and the concurrent high rates of type 2 diabetes, respiratory and cardiovascular diseases and cancer present considerable challenges to public health systems (Beaglehole et al., 2012;Horton, 2013;Hosseinpoor et al., 2012;Lozano et al., 2012;United Nations General Assembly, 2011;World Health Organization, 2012). Medicine can very precisely define the mechanisms of the bio-pathogenesis involved in these diseases and the associated risk factors of tobacco smoking, overconsumption of food and alcohol, and lack of physical activity.
However, medicine's ability to effect the actions that underpin the prevalence of these contemporary diseases remains limited (Fineberg et al., 2012). Further, the associated mortality and morbidity is strongly linked to social disadvantage: the patterns of health inequalities seen in all high-income societies are driven by NCD regardless of welfare systems (Buck & Frosini, 2012;Hosseinpoor et al., 2012;Mackenbach, 2012).
Over the last few decades, much research and work in public health has focused on persuading individuals to change their behaviour in an effort to reduce their propensity to develop these 'lifestyle' diseases. Since the aetiology of the NCD epidemics is to an extent rooted in the details of daily life and in what people do, this does not seem, on the face of it, to be an inappropriate response. However, while the distal behaviours are easily described, routines, habits and accepted ways of living are not so easily shifted.
The challenge of inducing behaviour change has led to an outpouring of arguments, methods and tools and a host of experiments and evaluations. Despite important points of difference, most behaviour change models presume that individuals are capable of making 'better' choices for themselves on the basis of information received, and that their well-being is in part an outcome of the decisions they make. This interpretation has been critiqued by Ioannou (2005), by Thompson and Kumar (2011) and most recently by nudge theory which reminds us that humans are only partly rational calculating assessors of information and often respond automatically to their immediate environments (Marteau, Ogilvie, Roland, Suhrcke, & Kelly, 2011). The evidence as to whether nudge may be applied with any effect to health-related behaviour change is, as yet, inconclusive (Hollands et al., 2013).
Not all theories and responses rest on models of more or less unfettered individual choice. For example, explanations that focus on 'wider determinants' highlight the importance of social contexts, and the systemic impact these have on peoples' health (e.g. Marmot et al., 2010). Changing such conditions has been suggested as a way of dealing with the epidemics of NCD especially since the diseases which follow in the wake of particular patterns of repeated behaviours show such a strong social-class gradient. However, whilst contextual/structural approaches argue for different forms of intervention (focusing on conditions and contexts and not on individual motivation), many reproduce a similarly individualised methodological and ontological approach that conceptualises macro-social structures as straightforwardly limiting, restricting or simply determining, the health choices or behaviours of individuals.
In calling for a theoretical reorientation within the field, and for an analytical and practical shift from health behaviours to health practices, Cohn (2014) highlights this paradox, arguing that structural or cultural approaches do not necessarily challenge the conceptual primacy of the individual and of his or her choices. In his words: … although discussion of context may ostensibly resemble adoption of a more sociological perspective, by assuming the delineated characteristics of health behaviour and pre-empting a focus for causal explanation, its inclusion frequently serves simply to maintain, rather than revise, conceptualisations of health behaviour. (Cohn, 2014, p. 159) A further problem for those seeking to explain the impact of context is that it is frequently difficult to know exactly how 'contextual factors' have effect. One consequence is that the practical actions which spring from such an analysis are often far from clear and exact examples of how to deliver effective change quite rare (Millward, Kelly, & Nutbeam, 2003). This may reflect the complexities of evaluation rather than ineffectiveness per se, but there is no simple prescription for action through which to affect the 'wider determinants' of health.
We suggest that neither approach provides much insight into the patterned, routine and habitual ways in which people live their lives. Although both approaches are well established and have been used to inform practical measures and programmes of intervention, in each case the potential for effecting change reaches its limits quite quickly: neither tradition has been exceptionally successful in dealing with the problems of disease that follow (Cohn, 2014;Kelly, 2010).
We highlight the potential and the practical relevance of an alternative socialtheoretical tradition: one which views the patterning of daily lives (and their implications for health) as outcomes of the coordination and synchronisation of social practices which persist over time and space, and which are reproduced and transformed by those who 'carry' them. We contend that public health policy would do better to focus on the 'lives' of social practices, treating social practices as topics of analysis and as sites of intervention in their own right.
Our basic proposition is that patterns of health and wellbeing are influenced by the practices people enactbearing in mind that practice is not a synonym for individual behaviour. This is not in itself a novel observation. Within Critical Public Health, for example, Evans (2011), Milne (2011 and Meah (2014) have provided empirical studies of food waste and food safety practices, respectively, using these to critique behaviourbased approaches in which associated 'risks' are treated as matters of individual consumer responsibility. Others have made similar points, highlighting the salience of Bourdieu's concepts of habitus and practice for health and well-being, or invoking Giddens' structuration theory as a means of overcoming the divide between agency and structure (see Maller, in press). Our distinctive contribution is to emphasise the relevance of these social theoretical traditions (Reckwitz, 2002;Schatzki, 2002;Shove, Pantzar, & Watson, 2012) for understanding processes of change in variously (un) healthy activities. As we show, focusing on the dynamics of social practices highlights processes and relationships that are obscured by explanations of change couched either in terms of macro phenomena (e.g. economic trends and structures of inequality), or with reference to individual choices and behaviours.
In making the case for social practice-oriented public health policy, we areof necessitycalling for a radical overhaul of the social theoretical foundation of much current research, and much of what goes on within the field. This is an ambitious project, but we are unlikely to make much progress without a paradigm change of this scale.
Beyond behaviour: considering the practices of smoking tobacco We start from the premise that drinking alcohol, eating, taking exercise and smokingto give a few examplesare not single behaviours in the way that public health conventionally defines them. They are broad domains of human activities that are reproduced and transformed through the re-enactment and performance of specific social practices, coordinated and synchronised across space and time. For each broad domain of practice, there exists an array of different possibilities, and a variety of ways in which practices can be performed. For instance, whilst the practice of consuming alcohol involves the action of ingesting ethanol, what is drunk, how much is drunk, how it is drunk, what is said, varies considerably from a dinner at high table in an Oxford College, a group of teenagers planning a Friday evening out, a gang of workmen from a building site relaxing after a hard day's work, drinking 'on the street', or someone drinking alone at home; they constitute significantly different variants of 'the' practice. Efforts to change 'drinking behaviour' as if this was a single entity are doomed to failure because drinking is manifestly not like that. To understand these broad domains of human activity, we need a framework that is able to explain (a) how practices are constituted and enacted and (b) how they relate to other practices across space and time. Both features are important if we are to account for patterns of distribution and persistence, and if we are to suggest interventions capable of changing the lives of practices and of thereby stemming the tide of NCD associated with them.
We outline some of the key characteristics of an approach that meets these criteria using tobacco smoking as an example. There is already a detailed literature within Critical Public Health that considers tobacco smoking and the various performances and contexts through which it is enacted. In a special edition of this journal, Bell, Salmon and McNaughton argue that to address health inequalities disguised by rhetorics of 'individual choice', those seeking to intervene in the consumption of alcohol, tobacco and fat need 'to recognise the social, cultural and political context in which public health policy is conceived and carried out ' (2011, p. 5). To this end, Dennis (2011) has detailed the lived experiences, and the meanings and narratives that smokers use to maintain their resilient habits in the face of anti-smoking policies, whilst Bell (2011) unpicks the 'discursive formation' of second-hand smoke to show that both popular and public health responses to health concerns are formed more by 'subjectively' experienced discomfort than by 'objectively' demonstrated harms. In concluding their review of the anthropological literature on tobacco, Kohrman and Benson (2011) call for more of this kind of detailed investigation into the subjective experiences and narratives of smokers.
Our aim is not to add detail to an already well-established body of work on smoking, the experiences of those who smoke or on discourses and narrative responses about smoking and anti-smoking policy. Whilst it is important to recognise the social, cultural and political context both of public health policy and of individual narratives, as well as the interplay between the two, we go further. Referring to smoking as an exemplary case through which to develop our argument, we suggest a paradigmatic shift in the way that public health conceptualises the reproduction of the (un)healthy activities that people do.
In what follows we consider smoking as a practice. There are different ways of delimiting 'a' practice and not all practice theorists would take this approach. For example, some might consider smoking as part of other more encompassing practices such as 'working', or 'going out' or 'taking a break'. Others might treat each of the actions of which smoking is made (for example, rolling, lighting and inhaling) as separate practices, consequently viewing smoking as a complex or bundle of practices. Different routes make sense depending on the purpose of the enquiry and the analytic strategy that follows. In our case, we take smoking itself as the central unit of enquiry on the grounds that this method allows us to examine its reproduction and transformation over time and in relation to other practices and practice bundles.
There are a number of reasons for choosing tobacco smoking as an example. First, it represents an activity that contributes to current cancer and respiratory and cardiovascular disease epidemics. As such, it has received vast amounts of attention from medical practitioners and the media and has been a focus of public health guidance and a variety of interventions including behaviour change policies, and specific clinical advice.
Second, the case of smoking allows us to illustrate the historical development of various (un)healthy practices. At one time a socially acceptable and even putatively healthy activity that distinguished an elite social class, smoking is now often associated with the oppositemembership of disadvantaged social groups and poor health. Over this same career, all kinds of smoking paraphernalia and kit have emerged and disappeared: including cigarette cases and holders, tobacco pouches, filters, humidifiers, lighters, papers and so on. Focusing on how these symbolic and material elements have evolved and been reconfigured allows us to address more fundamental questions about how social practices change and how these changes might be steered and shaped.
Third, smoking is evidently something that is closely related to other practices, like taking tea breaks at work, going out for the night, relaxing at home and so on. Understanding how these relationships are forged, and how smoking is coordinated and synchronised with other aspects of daily life is particularly important for the broader project of understanding how people are recruited to, and how they come to defect from specific social practices.
On all three counts, a discussion of tobacco smoking allows us to reveal some of the processes involved in the transformation and reproduction of social practice. At the same time, it is important to recognise that smoking has certain distinctive physiological features. Tobacco smoke delivers a powerful shot of the relatively harmless but highly addictive drug nicotine. We argue that even though addiction is an important part of smoking this makes it no less a 'social' practice, or any less amenable to analysis in these terms. It has been well documented (Murphy, Taylor, & Elliott, 2012) that addiction is multiplex, and that different social, psychological and physiological factors produce a range of addictive responses in different people. A focus on smoking as a social practice draws attention to the multiple ways in which addiction is reproduced (MacAndrew & Edgerton, 1969, take a similar approach to alcohol consumption).
Smoking is readily combined with other activities such as drinking, talking and working because unlike the intoxication that comes with alcohol or other drugs, it does not seriously interfere with speech, memory, concentration or motor skills. Bell and Keane (2012) have noted that whilst successful treatments for alcohol and drug addictions are generally believed to require extensive work on the many social/cultural associations of dependency (identity, relationships, lifestyles, etc.), treatments for tobacco continue to be focused more straightforwardly on breaking physical dependence.
In our analysis we bring these features to the fore, but do so without equating the 'social' aspect of addiction with relationships, identities or lifestyles. Rather than viewing addiction as a personal characteristic, we take it to be an outcome of the reproduction of a particular form of social practice organised through its relation to other practices in space and time.
Partly because of its addictive qualities, smoking has strong associations with habit. Since we use smoking as an example, some might conclude that our arguments are only or especially relevant to unreflexive aspects of daily life. This is not the case. Theories of practice are not inherently better suited to the analysis of unreflexive actions as opposed to those that call for more conscious and reflexive thought. In our view, social theories of practice are of value precisely because they take us beyond distinctions between the automatic and the rational, the conscious and the unconscious, and beyond interpretations of social action that are, at heart, centred on the individual and his/her state of mind (Reckwitz, 2002). More pragmatically, our central project is not that of understanding the regularity with which individuals smoke, but of understanding how the social practice of smoking itself changes.
We now turn to the literature on practice theory to briefly introduce a handful of key concepts that help in specifying and analysing practices like smoking, and in explaining how they develop and change.
Theories of practice Theories of practice have their roots in the works of Giddens, Bourdieu, Foucault, Heidegger, Wittgenstein and Marx, amongst others. Despite important differences of orientation and ambition, one common theme is that social practices are taken to be the site of social order and change. Interest and analysis consequently centres on the history and contemporary characteristics of everyday practicesordinary examples might include eating dinner, commuting, watching TV or smoking. Those who seek to analyse and understand the lives of practices emphasise a series of related features. Crucially, practices are, by definition, social: they are always shared. Whilst practices persist across space and time (Giddens, 1984, p. 2), they are never entirely static: they emerge, endure, change and disappear. They also interact, combining to form more extensive complexes and bundles that condition future possibilities (Schatzki, 1996). These core ideas inform a number of more specific propositions that are useful in thinking about how practices are constituted and how they develop.
Building on Reckwitz (2002), Shove et al. (2012) suggest that enacting social practices, doing things like cycling to work or eating dinner as a family, involves the active integration of generic 'elements', including materials/tools/infrastructures, symbolic meanings and forms of competence and practical know-how. Understanding how specific practices come and go is, in part, a matter of understanding the circulation and availability of requisite elements.
Social practices also interact. In some situations, they compete for resources, including those of people's time and energy. They can also support and sustain each other (e.g. eating often depends on cooking). Either way practices connect and in so doing form complex systems or bundles that have something of a life of their own. The trajectory of any one practice is likely to affect the trajectories of others and to be of consequence for different aspects of daily life.
The survival and persistence of a practice depends on its ability to recruit and retain cohorts of 'practitioners' (people) through whose performances/enactments the practice is reproduced and transformed. Whilst some practices gain more recruits, others lose out. It is therefore important to think about how people become the carriers of a practice like smoking, how do their careers (as the carriers of that practice) evolve and how do some defect? People are crucial both to the survival of practices and to their ongoing transformation. After all, if practices were not more and less faithfully (re)enacted, they would not persist.
We apply these ideas, in turn, to the case of tobacco smoking as a means of articulating a distinctive method of conceptualising and addressing both the diffusion and the persistence of NCDs. We conclude by reflecting on the potential and the pitfalls of developing the field of 'practice oriented' public health policy.
The elements of social practice Reckwitz says that a practice 'consists of several elements, interconnected to one other … "things" and their use, a background knowledge in the form of understanding, know-how, states of emotion and motivational knowledge ' (2002, p. 249). In other words, practices depend on the ongoing integration of at least three key elements: materials (objects, consumer goods and infrastructures); competence (including understandings of the situation; practical know-how); meanings (including embodied understandings of the social significance of the practice and past experiences of participation) (Shove et al., 2012).
To give a concrete example, smoking on a regular basis depends on an integration of materials: not only, cigarettes, matches and lighters; but also tobacco crops, factories, transport systems, retail infrastructures, an economy and so on; competence: to know where, when and how to smoke, for example, not only how to light a cigarette and inhale, but how to smoke in the 'correct' fashion for a given social situation (e.g. smoking in a beer garden is clearly different to smoking during a break at work) and meaning: understanding smoking as a normal and socially acceptable thing to do, variously associated with relaxation, sociability, masculinity, glamour and toughness.
It is difficult to smoke if one or more of these elements are missing: if there is no lighter, no notion of smoking as a normal thing to do, or no embodied knowledge of how to smoke, for example. In so far as practices are constituted by their elements, social practices change as and when these elements are reconfigured. Indeed, the material and symbolic elements of smoking have undergone a series of transformations over the past century. In 1914, in The social history of smoking, Apperson celebrated the fact that: The introduction of the cigarette completed what the cigar had begun; barriers and prejudices crumbled and disappeared with increasing rapidity; until at the present day tobaccosmoking in Englandby pipe or cigar or cigaretteis more general, more continuous, and more free from conventional restrictions than at any period since the early days of its triumph in the first decades of the seventeenth century. (1914/2006, preface) At this time, smoking was considered a social good: social reformers were keen that 'lower classes' should also have access to this wonderful panacea rather than it being restricted only to the affluent. Governments took to providing tobacco rations for troops and sailors. Meanwhile, a key material innovationthe development and introduction of mass-produced cigaretteswas significant for other elements of the practice, shifting meanings and loosening previously strong associations with social class.
At the beginning of the 1950s, Hill (1950, 1952) published the first evidence linking smoking to lung cancer, challenging understandings of smoking as a healthy thing to do. This has led to greater emphasis on quitting smoking and higher levels of state intervention, including banning advertising and smoking in public, raising excise taxes on cigarettes and introducing tools to support quitting, including nicotine patches, gum, electronic cigarettes and the provision of smoking cessation services by the NHS.
These developments have impacted on how smoking has evolved. Smoking, and the elements of which it is composed, is inherently unstableas with other practices, they change all the timebut in one form or another smoking remains resilient, with the prevalence of smoking in the UK persisting at around 20%, compared to 82% in 1948 (Action on Smoking and Health, 2013). In this section, we have underlined the point that practices like smoking evolve as new and different elements are integrated.
A second feature of practices is that they do not exist in isolation. Rather they are linked together to form bundles of practices that organise the time-space of social life (Schatzki, 2002). Schatzki suggests that the connections between practices can be characterised as either in harmony or in conflict. So while some practices might be incompatible, others become so closely coupled that they depend on each other, meaning that entire bundles are themselves routinely reproduced. A further method of intervening then is to pay attention to the ways in which bundles of practices co-evolve with a view to strengthening or weakening connections between them.
At different moments in its history and in specific socio-economic and cultural contexts, smoking has been variously connected to, and variously dependent on other practices and bundles of practices. We can consider the following examples: tobacco smoking has been a central part of religious rituals and shamanistic practices in many civilisations. In these situations, the characteristics of smoking depended on an array of related practices including those of prayer, meditation, divination and healing. In sixteenth-century England, smoking was initially viewed as one amongst other forms of medicinal practice, defined by its status within and as part of related bundles of practices including healing, diagnosing, prescribing, resting and recovery. However, Pollard (2004) explains that subsequently smoking lost these associations and formed new linkages with concepts and experiences of pleasure as it developed alongside Britain's colonial tobacco trade. At around the same time in Japan, smoking was connected to a very different arrangement of practices. Gately (2001) describes how smoking was taken up by Samurai knights who created ornate silver pipes and other instruments for smoking, bringing them to meetings of smoking clubs that were held on special and ceremonial occasions. In this instance, smoking's reproduction was dependent on the various practices that make up the life of a Samurai, including craftsmanship, combat training and ceremony. More recently, practices of smoking and drinking in pubs and clubs or smoking and going out for a meal in a restaurant were strongly interconnected in the UK, until the ban on smoking in public places (Bauld, 2011;Sims, Maxwell, Bauld, & Gilmore, 2010). In various other parts of the world, these close connections persist; but in the UK at least, eating at a restaurant and smoking, and drinking in a bar and smoking, are currently incompatible.
These examples demonstrate that practices can hold each other together in bundles (in various configurations in various socio-historical situations), they can become mutually dependent on each other, and they can break apart in ways that are important for their routine reproduction. At specific moments in time and space, such bundling depends on the exclusion of other practices, or of other configurations. This suggests that those seeking to intervene in the evolution of smoking as a practice should turn their attention not only to the integration of particular elements of smoking, but also to strengthening and weakening relationships between related practices at specific sites.

Processes of recruitment and defection
If practices are to persist they need people who are willing and able to enact them, and to keep them alive. Those who are interested in promoting specific practices, or stemming others, need to think about how practices capture and recruit their 'carriers' and about how carriers defect from a practice. Rather than treating individual motivations and desires as explanations of what people do, the more relevant questionfrom a practice-theoretical point of viewis to ask how it is that some people are recruited to specific practices and not to others, and how participation is sustained. In this analysis, individual commitment, motivation and desire figure as outcomes of engagement in the practice rather than as preconditions for it.
So how is it that certain individuals become the carriers of smoking whilst others do not? Practices depend on the coexistence and availability of requisite elementscompetences, materials and meaningsbut these are not evenly distributed across society. This is important for patterns of actual and potential recruitment. For example, practices that require extremely fit and flexible bodies are less likely to recruit from amongst the frail or the very elderly. The chances of becoming a practitioner consequently depend on what the practice itself demands and on previous life histories and resources (in terms of know-how, material elements, etc.) accumulated along the way. The structuring of opportunities and access to requisite elements is not random but is instead closely linked to what were earlier referred to as the 'wider determinants' of health (contextual or structural conditions). Not surprisingly, social inequalities play out in 'practice': that is in the range of practices which different social groups encounter and of which they do or do not become 'carriers'. This is a critical insight and one that helps bridge the gap between generalised accounts of 'structural conditions' and contexts, and a more detailed specification of the unequal social distribution of more and less healthy social practices.
Practices clearly differ in the demands they make of those who do them, and in the types of resources and commitments they command. This is especially obvious if we consider habits, many of which are important for public health. Rather than seeing these as individual traits, the challenge is to establish how certain practices manage to secure the resources, including time, money, etc., required for frequent, recurrent and habitual reproduction (Reckwitz, 2002). Understanding how habits take hold is a matter of understanding how the many practices that are reproduced in the course of daily life are synchronised and coordinated, and how some become more deeply embedded than others.
In this brief discussion, we have focused on a small number of concepts extracted from the literature on practice theory (Reckwitz, 2002;Schatzki, 2010;Shove et al., 2012), and applied them to the case of smoking. This is an illustrative, not an exhaustive exercisesocial theories of practice have much more to offerbut it is enough to give a sense of the potential that such a paradigm affords. At a minimum it is now obvious that the strategy of taking smoking, not the smoker, as the focus of analysis and intervention stands in stark contrast to the notion that smoking is either a single behaviour that is chosen by an individual or something simply determined by broader forces over which there is no control. It is also clear that focusing on social practices, on the 'elements' of which they are composed and on where, how and by whom they are enacted provides a means of showing how 'wider determinants' and structural conditionsthemselves an outcome of past and present practiceshave effect. They do so not as the context of individual behaviour but as conditions and outcomes of the diffusion and distribution of specific practices, and of bundles of practice. In short, patterns of health inequality are closely related to the patterning of social practice: that is, to how specific practices do and do not 'capture' recruits and to the unequal distribution of competences, materials and meanings on which participation depends. In other words, access to the various social meanings (understanding smoking as cool/sophisticated/unhealthy), materials (having or refusing access to different kinds of smoking paraphernalia) and skills (knowing different methods, rhythms and situations in which to smoke and not smoke) is not evenly distributed across society. Patterns of socio-economic inequality are, in effect, outcomes of past practices that are, in turn, relevant for the circulation and accumulation of the requisite elements of which contemporary practices are formed. In this respect, social practices constitute the 'missing' link: social inequalities are mediated and maintained by bundles of social practices that are, in turn, of great consequence for health outcomes.
Eating, physical activity and alcohol consumption can all be analysed in the same way, as can forms of personal hygiene, occupational practices, and the myriad of things that create lifeworlds and patterns of inequalities in health (Kelly & Doohan, 2012;Kelly et al., 2009). From this it follows that health inequalities and epidemics of NCDs are the product of the lived experiences of social disadvantage. The lived experience of social disadvantage itself consists of bundles of social practices with their own lifecourse. In showing how these connections are made, social theories of practice provide more compelling insight into the dynamic reproduction of health inequalities than do either the social determinants approach or individually oriented lifestyle explanations of behaviour and behaviour change.
We now comment on what a practice-theoretical orientation means for the design and implementation of strategies and policies that seek to improve public health.

Implications for public health policy
Those who take social practice as the unit of analysis and intervention are unlikely to proceed in the same manner as those who attribute health outcomes to individual behaviour or to 'wider determinants' and structural/contextual conditions. This is because theoretical commitments have methodological and practical implications. If we want to know how social practices develop over time, or what can be done to change them, it makes little sense to ask what motivates or constrains individuals to adopt more or less active lifestyles. To frame the question this way is to set it in terms of a behavioural model of choice and change and to assume that this model is useful for fostering change, guiding interventions and predicting outcomes. One problem is that the strategies which followsuch as providing individuals with more or different informationare unlikely to be especially effective as means of enhancing the development (or demise) of specific practices. Instead of addressing individuals, one at a time, practiceoriented public health would seek to understand and influence the emergence, persistence or disappearance of shared social practices like cycling to work, walking for fun or taking a smoking break. Since these practices are evidently critical for public health, the question to which we now turn is whether the lives of individual practices and bundles of practice can be governed and steered, and if so, what this might entail.
Our discussion of smoking suggests that practices can be treated as sites of intervention. One option is to focus on the configuration of elements that establish smoking as a normal or necessary thing to do. Another is to consider ways of influencing relationships between the various practices with which smoking is associated. Such techniques call for engagement not just with smoking but with the dynamics of different hybrid combinations, some of which might be developing, extending and attracting new recruits while others might be in decline (e.g. smoking on a night out).
It is important to notice that there is nothing especially new about the methods of intervention that might follow: for example, promoting new meanings, providing relevant infrastructure and assisting or preventing the development and diffusion of specific competences and skills, etc. Programmes and policies of this kind have been carried out to great effect, one of the best examples being that of banning smoking in enclosed public places. Since the elements of a practice also interconnect, it is possible and indeed likely that interventions focusing on the materials and infrastructures that smoking requires (e.g. by making it more difficult to find a place to smoke), affect competences (knowing where and when not to smoke) and even meanings (by reducing the sociability of smoking for example). Banning smoking in public places has consequently made a real difference (Bauld, 2011;Sims et al., 2010) not only to levels of smoking, but also to understandings of where smoking might go on.
However, smoking is demonstrably resilient and is therefore capable of adapting to changing conditions. For instance, new meanings of smoking are formed when people have to go outside to do it, and as these meanings take hold, new variants of the practice emerge. Further interventions will be required to root smoking out of the spaces and places to which it has retreated and in which it is now reproduced. Strategies like banning smoking in public places therefore contribute to and are consequences of the changing arrangements and practices both of smoking and of health policy-making. This suggests that a self-conscious practice-oriented public health policy would recognise its dual position, actively contributing to the constitution and reproduction of specific practices, and at the same time acknowledging that goals and priorities in public health are themselves outcomes of social practices/bundles of practice.
As with any theoretical approach, there are strengths and limitations. Whilst a turn to practice makes great strides in overcoming critical, problematic dichotomies between individuals and social structures, rational actions and habits, it does not generate simple guidelines for intervention. The question of how to deliberately steer practices and bundles of practices requires further thought. As Cohn (2014) comments, defining exactly where one practice begins and another ends is a matter of analytic judgement, meaning that explanations of social action depend on correspondingly debatable attributions of causality. This is always so, but recognising this to be the case is theoretically and practically challenging.
More pragmatically, whilst some established forms of intervention are compatible with a practice orientation, others are not. By implication, some currently popular techniques, such as exercise on prescription, would be abandoned or radically redefined. Similarly, providing generic advice for individual smokers is very unlikely to modify the coordinated and synchronised sets of practices (e.g. working and going on a night out) into which smoking fits.
Instead, practice-oriented forms of public health policy would be actively involved in continuously monitoring and adapting to changes in the arrangements of social practices that make up everyday life. The purpose of such monitoring would be to detect the cross-cutting impact of changes across a range of practices, (for example, in working, eating or travelling practices), to respond to these movements and changes and to intervene in and promote certain kinds of healthy practices. This might be achieved through targeted interventions aimed at making and breaking links between practices (the smoking ban effectively broke the link between smoking and eating at a restaurant), or, for example, promoting competitive practices around physical activity that exclude or are less compatible with smoking.
Taking these ideas seriously and taking them forward is undoubtedly challenging, not because the concepts are difficult to understand or mobilise, and not because the dynamics of practice are so chaotic that it is impossible to intervene. The problem has to do with the dominance and the power of, on the one hand, the individualistic behavioural paradigm, and, on the other, the 'wider determinants' approach, and with that follow from both.

Challenging dominant paradigms
In public health, as in many other areas of public policy, change is routinely understood as an outcome of individual choice, even when that is constrained by 'wider determinants' and structural conditions. The primacy of the individual as the focus of behaviour change interventions has been a key driver (and outcome) of economic theory and of psychological research into the mechanisms and means of behaviour establishment, behaviour change and the maintenance of change. These theoretical paradigms inform priorities and programmes of action: they matter for how resources are used, and for how policy impacts are discerned and evaluated.
Some of the most recent approaches within public health and psychology tend to be even more reductionist, with efforts being made to identify specific components of behaviour change interventions that may be effective (that is, effective in generating some detectable change in the individuals concerned). Such approaches are entirely consistent with a post-Thatcher society (in the UK) focused on the rights and responsibilities associated with individualism and the sociopolitical imperative of self-management. They are also consistent with the present UK government's aim to 'encourage, support and enable people to make better choices for themselves' (H.M. Government, 2010, p. 8).
In short, the individual 'decision-maker' and sometimes the context in which he or she makes decisions constitute the primary targets for intervention, and the primary point of reference when evaluating impact. The idea that behaviour is, at heart, a matter of individual choice, and the unspoken assumption that what people do is somehow separate from wider society, underpins the design and provision of guidance on food and alcohol consumption, along with efforts to persuade people to exercise on a regular basis. This is not an especially conducive context in which to advocate an approach that is paradigmatically opposed to the forms of individualism on which so much contemporary research and policy depends.
It may be an uphill struggle, but we argue that one way, and perhaps the only way, out of the difficulties now facing public health policy in relation to NCDs is to refresh the terms in which problems of 'behaviour change' are defined, framed and evaluated. In this context, the idea that practices (like smoking, eating, drinking or taking exercise) are the proper target for intervention makes a lot of sense. If taken to heart, this calls for a major change in the theoretical foundation of public health policy and for corresponding forms of methodological inventiveness and ingenuity.
We have sought to lay the foundations for practice-oriented public health policy and make the case for an approach which takes the practice, and related webs of practice as the primary focus of attention and intervention. Identifying possible and viable means of modifying the ways in which practices develop (who they 'recruit', how they extend and shrink and how they change) is a matter of identifying the preceding conditions and elements of the practice in question. It involves thinking about recruitment to and migration from the practice, and about how such processes relate, for instance, to the life course, or to issues of access and equity. It also depends on understanding the types of skills, practical know-how and interpersonal relationships involved, and the social meanings and emotional responses that practices engender and reproduce. Above all, it bids us to stop thinking about risks to health as if they were some latter day equivalent of a bacteria or a germ working their pathogenesis on the individual human body and to see not individuated selves who can be prompted to change, but people enmeshed in social arrangements that are defined and constituted through the practices they enact, whether for good or for ill.