The concept of ‘lifeworld’ and the mixed mechanisms of health and disease

The concept of lifeworld is central to our argument and for this reason we present it first. Our approach is derived, but distinct from, the work of Alfred Schutz (1972; 1975). We also draw on the social theory of Anthony Giddens (see Giddens, 1979; 1986; Giddens and Dallmayr, 1982), and Pierre Bourdieu (2000; 2008). The basic idea is that each and every one of us inhabits our own lifeworld. It is deeply subjective in the sense that it is constituted of the assumptions, understandings and taken-for-granted aspects of our everyday existence. It is the seat of our sense of self, and the ideas we have about who and what we are, and who and what others are. We anticipate the actions of others and we anticipate the effects that our actions will have on others in our lifeworld. But the lifeworld also consists of the things we do, the actions we take, the practices in which we engage on a day-to-day basis. It is what we do, and our bodies are like a book in which we inscribe all the many things we do and we experience. Any aspect of our life can be described in terms of the lifeworld experience, including health and disease. For health and disease, in particular, our point is that all this information is crucial in order to decide which factor(s) to intervene in, whether to target specific individuals, groups or the whole population, and to identify the most suitable media to promote an intervention.

It is important to note that our individual lifeworlds abut others’ lifeworlds, particularly others who share the same or similar life circumstances to us, and share our placement in the social world. Our practices, our actions, have consequences at a supra-individual level. While we engage in practices in our own apparently singular lifeworlds, simultaneously billions of other people are likewise engaged in similar practices. Social patterning at population and community level arises from the intersecting practices of different social groups producing the rich variegation in the social world. In the context of public health interventions, this explains why one size does not fit all, and why some interventions will be more effective if targeting specific groups, or the whole population. The aggregate consequence of billions of people engaged in trillions of practices is not random chaos; rather, it produces patterns at the social level. It is by detailing these patterns, at different levels of aggregation, that we can intercept the social practices that are likely to influence health and illness.

But there is more. Actions have health consequences. The patterns of people’s lives, their lifeworld across time – the life course – will inscribe on the body, or get under the skin not just metaphorically, but biologically through human ‘omics’ as a timeline of our life (Kelly and Kelly, 2018). To understand this, we need to supplement the concept of lifeworld with that of ‘mixed mechanism’, in which biological and social factors are on a par to explain health and disease, and as candidates for intervention. Biological factors may have an effect on social factors; social factors may have an effect on biological factors; and the two are de facto interacting all the time. Ongoing research in exposure science, epige.netics and metabolomics, allostatic load, and the life course approach to the ‘social-to-biological transition’ tries to detail aspects of these mixed mechanisms, including the pathways between the social and the biological (Kelly-Irving et al, 2015; Delpierre et al, 2016; Kelly and Kelly, 2018; Radford, 2018; Kelly, 2021).

It is precisely because the mechanisms of health and disease are mixed that some public health interventions should target social factors in early life, even if the health outcome is expected much later in the life course. Some people have greater ability and resources to control their lifeworlds across the life course, because of access to resources, power and so on, but all of us are the product of the recursive nature of the timelines of our lives. Therefore we suggest that the mechanisms that bridge the body and the social sphere are to be located in the recursive nature of the practices in which we engage, or in which we are enmeshed.

Once we agree that the concepts of lifeworld and of mixed mechanisms hold potential to explain the phenomena of health and disease in their many facets, namely as biological and social phenomena and as individual- and population-level phenomena, then the next challenge is how to make them actionable, rather than just explanatory. And to make them actionable, lifeworld and mixed mechanisms need detailing and quantification. Our constructive proposal is a synergistic use of qualitative and quantitative methods across the social and biomedical sciences of health. This combined methodology will, in turn, allow us to incorporate appropriately the social dimension in the understanding of the mixed mechanisms of health and disease and in the design of policy intervention (see Greenhalgh, 2020). Before we explain our proposed methodology, it is important to further qualify the concepts of lifeworld and mixed mechanisms with respect to some existing, similar approaches in the sociology of health.

Structures without mechanisms

Over the years, sociological theories and methods applied to medicine have developed approaches close to the idea of lifeworld we have now sketched out. These approaches have paid attention to the role of institutions, social support, power structures and communication. These are all important elements for a proper understanding of the social dimension of health. These sociological approaches have not generally been dominated by quantitative analyses. They have tended towards the role of these elements at a high level of abstraction – social structures or systems. These are all very important, but not very actionable for the purpose of the design of public health interventions.

Knudsen and Vogd (2015), for instance, revive and develop the system theory approach. At its basis we find the concept of ‘polycontexturality’; this means studying health/patients from different perspectives (of the patient, of the doctor, of the healthcare provider and so on). A system theory approach (especially that of Niklas Luhmann) is then applied to the sociology of health and illness. Here, various aspects of the many ‘contextures’ involved in health, how they are arranged, and how they interact are considered. The main focus is communication as a fundamental element of society (and therefore of medical social systems too). Communication is used to shed light on decision-making processes, which may involve science-based decisions about diagnosis or treatment as well as the economic and financial aspects of healthcare management. In this way one can, for instance, account for the tensions between the (scientific) principles of evidence-based medicine and their implementation in healthcare management.

The main difference with the lifeworld and mixed mechanism approach, already discussed, is the following. The tendency in the structure or systems approach is to obscure or deliberately eschew the specific explanatory power of the biosocial mechanisms affecting health, and their usefulness for policy design or intervention effectiveness. Individual and group behaviour (including preferences, choices and values) become somehow secondary to higher-level sociological structures and while we grasp – at this higher level – the extent to which medicine and medical practice are fundamental social activities, we might miss more concrete, down-to-earth explanations of how and why certain socio-economic factors impact on health and disease.

In sum, while much potentially important policy and practice data and evidence has emerged, what is missing in this literature is that the mixed mechanisms underlying the phenomena of health and disease remain largely undefined. This is still more surprising because since the 1980s there has been a vibrant subfield, which emerged in medical sociology, about precisely the sociology of the body (Turner, 1992; Shilling, [1993] 2012). This is an important and dynamic literature, but one that, in our view, paradoxically has not established a bridge with the fields studying the biological sphere. Our view is that a system approach is useful to the extent that the role of high-level social structures is also explained in terms of the mixed mechanisms in which health and disease happen, at the individual and group level. In the next section, we attempt to explain how this can be done.

What to measure and how to measure it

The lifeworld of health and disease is very much about social factors and their role in health and disease of individuals and groups. But the problem is that the level of description of a lifeworld does not lend itself easily to the quantification of these factors. The temptation is therefore to default to statistical approaches which take a social factor (education, socio-economic status, access to health infrastructure), and measure it as precisely as possible. We call this ‘the problem of more measurement’: increasing the granularity of measurement of social factors does not carry explanatory power on its own. Going for more granularity can be taken as the core of methodological individualism, which, to put it simply, says that any phenomenon at the macro-social level is ultimately to be explained by the actions of the individuals (Zahle and Collin, 2014). Our reservation is, to be sure, not with the measurement of individual characteristics per se, but with confining the explanation (and the discourse on actionability) at the level of the individual. It is for this reason that we think that ‘more measurement’ does not solve the problem. The problem of ‘more measurement’, moreover, is exacerbated (rather than addressed) by the adoption of digital health records and of digital electronic devices for gathering data. Having more data helps very little if data are not accompanied by appropriate reflection on what these data mean and represent.

Our solution does not rely on quantity but on quality. In a nutshell, we need to consider more closely which (social) variables to measure and to specify what aspects of health and disease these variables are in fact supposed to measure or be proxies for. Providing theoretical underpinning to measurement of the social, hitherto, provides in turn (partial) descriptions of the lifeworlds in which mixed mechanisms operate. Thus, the suggestion is not to provide more measurement in qualitative approaches, but to take advantage of the (thick) description that qualitative studies can provide, also in the context of quantitative modelling. Let us explain further.

It is not a new idea that measurement in social science is complex and challenging. There is a well-established tradition in social science that reflects on how to measure social characteristics (Zeller and Carmines, 1980; Blalock, 1982; Pawson, 1989). This section builds on this tradition, drawing specific attention to the following. There are variables/factors that are easy to measure but problematic to interpret (such as age), and there are variables/factors that are difficult to define but then apparently easier to measure (socio-economic status, for example). In trying to establish a connection between ‘the social’ and ‘the biological’ dimension of health and disease, within a complex mixed-mechanisms approach, difficulties and challenges in measurement are exacerbated. There are also some aspects of social life where measurement is extremely problematic, and the question arises whether dynamic social processes are amenable to conceptualisation as variables at all (Blumer, 1962). There is also the risk that conceptualising things mechanistically resurrects naive positivistic attempts to apply some of the methods of the natural sciences to the social world – this is not the line of reasoning we advocate.

Take the example of age. Nowadays, chronological age is very easy to measure, especially in high-income countries, where demographic registries are long established and properly working. Age may provide very useful information to classify/stratify the population according to given socio-economic-demographic characteristics to be mapped onto health conditions (and vice versa). But what kind of information is age really giving? This is where a quantitative-oriented approach to social measurement needs to be complemented with more qualitative-oriented approaches that go into the details of habits, culture and behaviour – basically all key aspects of a lifeworld. Is it the same to be 15 years old in industrialised Western countries and in rural developing countries? It certainly isn’t. Thus, an appropriate description of the lifeworld should be able to make this distinction visible and meaningful, exploiting not only the ‘quantitative’ information about age (easy to retrieve), but also the ‘qualitative’ information about what it means to have a given age in a given culture, context and environment. This qualitative information would constitute the kernel of the lifeworld, providing the details of the mechanisms through which the social operates to affect health, and potentially providing variables that are finer grained than ‘age’ or ‘level of education’.

The quantitative measurement of chronological age carries another problem. The rate at which we age biologically is determined not by our chronological age but by the interaction of our biological inheritance with the experiences of exposures to microbiological, toxicological and traumatic insults across the life course as well as social exposures including diet and nutrition, and broader environmental and socio-economic phenomena. The impact of these factors is strongly patterned by social class, ethnicity, education, income and gender. Thus, although chronological age has a deceptively easy and quantifiable character, its meaningfulness, as a way of understanding the social dimension of health and disease, is highly contingent. This becomes clear when considering current research done in molecular epidemiology and epigenetics, where the concept of ‘allostatic load’ is meant to capture a ‘biological’ state of the body that is caused by a number of stressors, from the environment to important events in life or continuous stressful situations (see, for example, Delpierre et al, 2016; Vineis and Russo 2018).

Conversely, numerous socio-economic variables and factors are difficult to measure and yet their effect on health is known to be pervasive. The challenge is, then, to explain the mechanisms through which given socio-economic conditions get under the skin. Numerous studies exemplify this point. For instance, Case and Deaton (2015) ran a thorough analysis of midlife mortality and morbidity in non-Hispanic Americans in the 21st century. In their work, they clearly relate the rise of mortality in the said group to a rise in (self-reported) morbidity. Their analysis includes aspects of the lifeworld of the said group that, while amenable to a quantitative assessment (for example, how much respondents were able to walk, whether they could socialise with friends), require deep explanation of the social dynamics and how these directly affect health.

In none of these cases are quantitative approaches alone going to address appropriately the challenge of understanding the biosocial nature of health and disease, because having ‘more data’ does not automatically mean more understanding. Likewise, qualitative approaches cannot, on their own, provide us with an explanation for how and why the social turns into something biologically visible and quantifiable and real. In both quantitative and qualitative approaches theory, prior studies and theoretical assumptions play a major role, and these should be an explicit part of an integrated operationalisation of the lifeworld. Measuring the social is an important exercise in quantification that has delivered many useful results and that needs to be further pursued. Ghiara and Russo (2019) argue that to understand how the social gets under the skin, we need to attain a finer level of granularity and measure at the individual level socio-markers, in a way analogous to biomarkers. They suggest that this can be a step in understanding how mixed mechanisms work, how the social and the biological should not be considered dichotomously but as part of a same whole. Ghiara and Russo’s approach, while pushing for the need of identifying markers of social factors, also makes the point that their identification rests on the possibility and ability to understand people’s lifeworlds.

What we advocate is exemplified in ongoing work on stress and inflammation, allostatic load, and the social-to-biological transition, which is theorised by scholars working on ‘exposome’ initiatives (see, for example, Vineis and Kelly-Irving, 2019; Vineis et al, 2020; Vineis and Barouki, 2022), and found in numerous empirical studies (Albus, 2010; Fagundes et al, 2011; Eisenberger et al, 2017; Bilal et al, 2018; Diamond et al, 2021).

In the remaining part of the section, we explain how, in our view, qualitative and quantitative research can, together, contribute to specifying and measuring salient aspects of the lifeworld, thus ‘measuring the social’ in a useful way.

Beyond the qualitative and quantitative divide

In order to elucidate mixed mechanisms, we must develop a detailed and precise understanding of the characteristics of lifeworlds and the life course that goes beyond broad statistical generality and that provides enough details to enable action. One of the most significant gaps in public health understanding, we contend, is an inability to describe fully social variation in populations – in other words the lifeworld in which mixed mechanisms operate – in sufficient detail to act effectively. Measures such as class, education, income, gender and ethnicity are of course to be found in the literature, but the way these social differences intersect with each other in the lifeworld to produce patterning at local level and differences at individual level is hard to find in the public health canon. Yet medicine is very familiar with, and works with, the grain of individual biological variation – at its simplest individual people respond differently to biological interventions, like the ingestion of a pharmacological agent. Social variation manifested by the variation in lifeworlds has the same variable effect on interventions and outcomes, and yet it is neither properly theorised nor properly empirically investigated. Not surprisingly, then, the facility to elaborate the mixed mechanisms both of aetiology and prevention are hidden way – a problem that often just seems to be too complex to even try to unravel.

Nevertheless, methodologically, unravelling mixed mechanisms may actually be within our grasp, and policy makers need not fall back on simple heuristics to deliver effective interventions. There are a number of existing approaches that try to move beyond the quantitative vs qualitative dichotomy. The most paradigmatic is perhaps mixed methods research (MMR). Applied researchers in MMR combine both quantitative and qualitative approaches to address the same research question. Methodologists in MMR, at the same time, also reflect on how this combination of methods ought to happen (Guba and Lincoln, 1994; Johnson et al, 2007). Some, for instance, prefer calling it ‘multi method’ (rather than mixed method) to emphasise plurality and to keep distinct aims and goals when applying one approach or another in a study. There is an ongoing debate in the field as to whether MMR constitutes a new paradigm in social science research or whether MMR blends different existing paradigms, in a Kuhnian or post-Kuhnian sense. We build here on the work of Ghiara (2020), according to which MMR combines different, existing paradigms in social science. This has a rather specific meaning, for instance combining epistemologies as different as those underpinning qualitative and quantitative studies, or different ontologies, such as single-case or generic causal relations. We also build on a recent contribution on the foundations principles of MMR by Johnson et al (2019), who point to the importance of adopting a pluralist stance, notably concerning the notions of causation, evidence and mechanism. In the remainder of this section, we further explain how both quantitative and qualitative approaches can contribute to elucidating the mechanisms through which the social affects health and disease.

The general idea is that quantitative and qualitative approaches contribute to understanding mechanisms, but at different levels. On the one hand, with qualitative, small-scale studies we can typically grasp the details of the lifeworld: we can describe very precisely what happens between actors, and why, and under what conditions and so on. We can also compare different qualitative, small-size studies and enhance our understanding of cultural differences, or similarities. On the other hand, with quantitative large studies we can generalise, or test how stable certain correlations are across cultures, geographical differences and even times. MMR is a flourishing area in which very specific modelling approaches have been developed; our point, to be sure, is to highlight the potential of mixed methodologies, and of pluralistic methodological approach more generally, to address the limitations of either side of the qualitative and the quantitative divide. The approach is not to emphasise the incommensurability of methods and epistemologies, but to acknowledge that human life, as an object of study, cannot be captured by singular theoretical categories or methods, requiring ways of grasping the physical, biological, social and cognitive totality of the subject matter. MMR offers a very promising avenue in this regard.

The value of keeping both quantitative and qualitative studies, and of fostering a mutual support in these research traditions is twofold. First, methodologically, quantitative and qualitative approaches really are complementary: detailed qualitative studies may give hints about what to test at large scale, population level; conversely, quantitative studies may allow unexpected correlations (or absence thereof) to emerge deserving an in-depth qualitative study. Second, at the level of policy making, we need to know and understand which mechanisms are really culture-specific and which mechanisms are instead more general. This should help us design public health policies that are very specific and tailored to a population, or group, and decide when to export successful intervention to different contexts.

The biological and aetiological approach

The dominant paradigm in public health is based on a biological and aetiological understanding of health and disease, focusing on the biological causes of disease. For communicable diseases, this is thought to be the most appropriate approach, using ‘the social’ as classificatory, but not an explanatory or aetiological factor. In the case of non-communicable diseases, the approach takes ‘the social’ to be factors associated with risk, again largely in a classificatory rather than explanatory or mechanistic way.

The scientific literature is enormous, ranging over heart disease, cancer, obesity, alcohol problems and sexual health, for example. Consequently, we know a great deal about the origins of these diseases – the biological causes – and the things that put people at risk of them. These risks are typically located in behaviour and lifestyle. However, the mechanisms linking the risk factors to disease are much better defined biologically than the mechanisms linking the social to pathology. Consequently, the way to explain which aspects of the social are relevant in which specific context, and how to act on the social for public health purposes, is not clear. Most policy therefore falls back on high-level generalisations about, for example, the relationship between social disadvantage or ethnicity and poor health, or on simple recipes for behaviour change. In practical terms, this does not take us very far, because it does not specify what precisely needs to be done to change or improve things. These models also assume that if aetiology is sufficiently well understood, then effective preventive action can follow (Kelly and Russo, 2017).

The efforts at prevention have been geared, across the world, either to trying to change the behaviours that lead to exposure to risk, or less often at the so-called wider or social determinants – education, social class, income, poverty, disadvantage and social exclusion (Rutter et al, 2017). Neither approach has been conspicuously successful. This, we suggest, is because the linkages between the social and the biological remain ill-defined and under-described in mechanistic, and particularly mixed mechanistic ways. The logically important fact that the mechanisms involved in aetiology and in prevention are different is sidestepped. This is also the case for communicable disease. The aetiology from microorganism to infection is well understood in principle, and usually in its specifics, for most communicable diseases. However, the importance of human social affairs in the spread of infection, and the acceptance or rejection of vaccines, while known about, are not subject to anything like the detailed mechanistic understanding of the virology or the bacteriology. Although the emphasis in what has been called ‘population health’ is leading towards redefining epidemiology as population health science, this in itself does not address the specific point with which we are concerned with (see, for example, Keyes and Galea, 2016; Valles, 2019). In part, this is due to the fact that this approach, while explicitly focusing on social factors, is not explicitly explanatory or aetiological, as we explain next.

The social determinants approach

The biological, behavioural and social factors are conventionally correlated and associated with disease outcomes. Part of this literature shows that socio-economic inequalities correlate with poor disease outcomes (see, for example, House, 2002; Marmot, 2005; Wilkinson and Pickett, 2010; Bambra, 2016; Bartley, 2017).

This approach addresses the fact that socio-economic factors, and especially inequalities, map onto health patterning at the societal level; it is closely linked to social epidemiology. The literature is extensive and has been around a long time (Gairdner, 1862; Kadushin, 1964; Antonovsky, 1965; Marmot et al, 1978; Erikson and Goldthorpe, 1993; Mackenbach, 2006). Since the 1950s, this literature has been enlarged to include the idea of risks linked to social factors, especially risks associated with practices like smoking, diet, alcohol, consumption, sex, drug use and physical inactivity (Doll and Hill, 1964; Sytkowski et al, 1996). Globally, all forms of disease are patterned socio-economically and some infectious diseases are unequivocally diseases of poverty (Kelly and Doohan, 2012). The original impetus for this type of work was in the 19th century, when infectious disease was the largest cause of mortality, particularly associated with deprivation. The links between infection and disadvantage have been most recently underlined in data from the US and UK on COVID-19 deaths.

The results of this line of work have been relevant and groundbreaking, and have shed much light on the social dimension of health and disease and in the case of smoking led to declining rates of smoking and corresponding disease prevalence. The results of this research (and especially that of Marmot) are routinely endorsed and used by the WHO, which recognises the tight relation between social factors and health and urges the implementation of public health interventions that specifically tackle social factors related to health (WHO, 2008).

There are several important points to note about this approach (see, for example, Kriznik et al, 2018). Relevant to our argument, in particular, is that this approach establishes that socio-economic factors are crucial to health. However, it does not establish why/how it is so, nor how to go about changing things. In other words: while the social determinants approach does establish meaningful and robust correlations between social and behavioural factors and health, it does not elucidate the mechanisms or the causal pathways through which ‘the social’ affects health. In this ‘social determinants’ approach, the tendency, especially in recent years, has been to go as granular as possible in the measurement of the social. Thanks to progress in sampling, statistical techniques, availability of medical records, geographical information systems, data linkage and so on, we can establish such correlations at increasing levels of granularity. This means, for instance, being more precise about the social and geographical groups involved. Alternatively, it means refining socio-economic characteristics with more precise definitions and measurement tools, or relating not just ‘classes’ of pathologies but finer-grained types of disease. Think, for instance, of how many types of breast cancer we can now differentiate. Our main concern with this approach is that, after all, correlations remain non-actionable or difficult to action. Instead, for the purpose of public health interventions, we need to elucidate the mechanisms, or rather the mixed mechanisms at work in the lifeworld of individuals.

Why evidential pluralism can bridge the biology and the social dimensions of health and disease

The two approaches just presented have been successful but unfortunately remain largely distinct, and therefore unable to bridge the world of ‘the biological’ and the world of ‘the social’. Our own approach is in line with evidential pluralism, a line of research in the philosophy of causality and of medicine, in which both elements (correlations and mechanisms) are important (Clarke et al, 2014; Parkkinen et al, 2018), and that can help bridge the biological and the statistics-based approaches.

Evidential pluralism is a position developed in philosophy of science and in philosophy of biomedicine, according to which, in order to establish a causal claim, we typically need evidence of correlation and evidence of mechanisms (Russo and Williamson, 2007; Illari, 2011; Parkkinen et al, 2018). This is an epistemological and methodological thesis about disease causation, rather than an ontological account. In the field of (philosophy of) biomedicine, evidential pluralism has been used to argue, against evidence-based medicine (Sackett et al, 1996), that evidence of mechanisms is also important, and in a way that is not reflected in evidence hierarchies (Clarke et al, 2014) or even in the GRADE system (Parkkinen et al, 2018). Interestingly for our argument, evidential pluralism is not only about exploiting mechanisms and correlations synergistically for a more solid evidence base. It is also about enlarging the scope of what we take evidence of: both mechanisms and correlations can be about the biological and about the social dimensions. In particular, evidential pluralism has, until now, largely focused on how biochemical mechanisms should be part of the evidence base to establish causal claims, besides correlations. But our argument is that mechanisms of health and disease are not just biochemical but also inherently social or, better said, biosocial (see also: Kelly et al, 2014; Kelly and Russo, 2017; Kelly and Kelly, 2018; Ghiara and Russo, 2019).

Let us return to COVID-19 to exemplify the usefulness of evidential pluralism and of biosocial mechanisms. On the aetiological side of the equation, the vector of transmission involves a large number of social factors. These include socialising, touching, sneezing and coughing, working environments and work practices, domestic circumstances, numbers of people and generations in shared domestic households, and locality, age, ethnicity, sex, occupational, educational and income composition of communities. On the preventive side of the equation, these same factors were equally important, although the mechanisms involved are not the same as the aetiological ones. The problem is that the knowledge base about the preventive and the aetiological social mechanisms are not known about in sufficient detail to enable jurisdictions to act forensically. Thus, blanket, rather than targeted, interventions were the best that most authorities could do – and then hope for the best. Most governments favoured non-targeted and unspecific action based on the known social coordinates, and opted for heuristics – short-cut thinking to provide answers to complex problems – without knowledge of unknown or intended and unintended consequences. Heuristics allow fast thinking, but they are usually biased and often wrong (Kahneman, 2013). The fact that high-risk groups, such as the BAME community and the elderly, should have been the target for protection was ignored. And, of course, the epidemiological models used to justify the decisions taken by governments were invariably about aetiology, not prevention, and did not deal in mechanistic social evidence about either aetiology or prevention (Aronson et al, 2020).

More generally, many preventive efforts relating to non-communicable disease, in particular, have not produced results that have made much impression on the mortality and morbidity associated with the patterning of inequalities, or the prevalence and incidence of disease (excepting smoking related disease). This had led to commentators raising two issues. First, given what is known about aetiology and risks, many diseases ought to be by now preventable, and yet some epidemics are on the rise – obesity and some cancers for example (Cavalin and Lescoat, 2017; Horton, 2017). And, second, despite the deep understanding of the biology of non-communicable disease, public health interventions mostly target either whole populations or behaviour in ways that have been less than optimal (Marteau et al, 2015; Rutter et al, 2017). It is not clear why. The interactions between different social dimensions are complex, self-evidently, but they are not unknowable. Very clear accounts exist in the sociological literature such as social practice and structuration theory (Giddens, 1979; Giddens and Dallmayr, 1982; 1986; Bourdieu, 2000; 2008), which provide high-level, but very informative frameworks to understand the interactions between people’s actions and the social structures that they inhabit, and the ways that these interactions ingrain themselves in human biology. At the very least, they provide a route map with the key coordinates for forensic action to prevent transmission and facilitate preventive action.