Framing disease: An underappreciated mechanism for the social patterning of health

https://doi.org/10.1016/j.socscimed.2008.02.017Get rights and content

Abstract

The emerging fields of social epidemiology and population health seek to understand the social determinants of health. Especially, with regards to how income inequality causes health disparities, attention has been focused on material and psychosocial mechanisms. I use examples from the epidemiological and social science literature to argue for a third broad etiological framework: the role played by the ways we generally recognize, define, name, and categorize disease states and attribute them to a cause or set of causes. These framing effects shape population health by influencing: health and illness beliefs; patterns of consumption and other behaviors; perceptions of what interventions and policies work; class, ethnic, and other social dynamics; and clinical and public health practices. Important characteristics of many framing phenomena are their capacity to be self-perpetuating and their performative power. A better understanding of framing effects can lead to deploying them more deliberatively and flexibly to improve individual and population health.

Introduction

Clinicians, researchers, and policy makers have rediscovered the social determinants of health and health inequalities in the United States and other modern, industrialized countries. In parallel, new disciplines have emerged – most prominently, social epidemiology and population health – whose core definition is not being risk factor epidemiology and rejecting the emphasis on individual-centered public health practices. These reconfigured fields aim to find new means of improving societal health outside the provision of health care services, for example by influencing social and economic policies, mass communication, and the built environment.

One contentious issue in establishing this emerging paradigm is the mechanism by which social factors lead to health and disease in the bodies of individuals. Frequently, the shorthand for this issue is “how does culture get under the skin?” Most everyone recognizes that individual behavior, environmental exposures, and differential access to health care make a difference, but what these new fields want to understand and influence are contextual factors “above” the level of the individual. Two broad causal conceptions have emerged, especially salient in debates over the health consequences of income inequality. One emphasizes material mechanisms over the lifecourse – such as the availability and quality of schools, housing, and health care (Lynch, Davey Smith, Kaplan, & House, 2000). The competing conception emphasizes the important role of psychosocial mechanisms, such as perception of one's position on the socio-economic ladder, in shaping societal health (Wilkinson, 1996). Psychosocial proponents point to research on stress and heart disease and the role played by relative not simply absolute income on health outcomes.

I want to make the case for a third broad etiological framework for understanding the social patterning of health and disease. These are ways that the health of individuals and populations results from how societies generally recognize, define, name, and categorize disease states and attribute them to a cause or set of causes. These framing ideas and practices can have profound effects by influencing individual and group behavior, clinical and public health practices, and societal responses to health problems. Many of these causal pathways are frequently uncovered in standard epidemiological studies but are categorized as biases or confounders rather than as causes or mediating influences. This dismissal makes sense if we assume, as most health researchers do, that social and historical variation in health beliefs, diagnostic practices, and what problems are medicalized are mere epiphenomena in relation to “real” health status. But I will argue below that these framing phenomena are sometimes the very mechanisms by which the social patterning of health and illness emerges.

I use framing somewhat euphemistically for what is often referred to as the social construction of disease (Rosenberg, 1992). This usage may avoid a few unwanted connotations sometimes associated with constructionist arguments – a style of dated cultural relativism, a lack of common sense, and a reflexive opposition to biomedicine. In recent socio-historical scholarship, not only have there been terminological shifts but also a substantive one from case studies of obvious borderline diagnostic categories (e.g. homosexuality as a disease in psychiatric nosology) to more complex and systematic research into normative medical categories and classifications (cancer, heart disease, etc.). Although there is no inconsistency between attributing a causal role to framing phenomena and also to materialist and psychosocial ones, they are rarely combined and often sort by discipline (e.g. historians focus on framing phenomena while social epidemiologists focus on materialist and psychosocial ones). Yet, knowledge and insights stored in these different silos might profitably borrow from one another, resulting in more complete accounts of the social patterning of health.

Below I survey a range of framing mechanisms and consider the plausibility of understanding them as etiological in the social patterning of health. Some of the examples are drawn from the epidemiological literature. Investigators typically explored associations hitherto understood in material or psychosocial terms but upon further and more subtle analysis proved to better explained as reporting and selection biases, misclassifications, misapprehension of changed diagnostic criterion and patterns for true disease change, or misleading representations of efficacy.

Other examples come from the social sciences. Consideration of framing phenomena in disease has a long history within medical history, sociology, and anthropology, albeit under different labels, each with their own (often contested) scope and connotations – social construction, medicalization, labeling, and others. It is much less widely recognized, especially in the emerging field of population health and the policy attention given to health disparities, that framing phenomena (1) in aggregate represent a mechanism causing or mediating some of the social and temporal patterns of health and disease in the contemporary U.S. and (2) that much of the evidence for “framing as mechanism” is right before our eyes in the medical and epidemiological literature and yet is not recognized as such, in large measure because from the perspective of everyday medical and public health rationality, these phenomena are obstacles and diversions from understanding “true” causal mechanisms.

As a final introductory note, I want to point out that while I evoke “how culture gets under the skin” to situate my argument about framing as mechanism, I do so only because it is the most common shorthand that epidemiologists and clinicians use to map social conditions to biological phenomena. Epidemiologists and clinicians use this shorthand because they cannot precisely or even imprecisely define the interface they intuitively want to capture – what exactly is inside and outside the body or culture, or what culture is and is not. In other words, I am using an existing and problematic term of reference. Yet, by arguing for the inclusion of framing phenomena I want to challenge medicine's and epidemiology's everyday if poorly articulated assumptions about the location and meaning of this culture/body interface. Again, this is by no means a unique challenge but I hope to make a contribution to the burgeoning population health field by using findings from within the medical and epidemiological literature and the interpretive social sciences, which are non-trivial in their scope and implications, and that are not confined to borderland health conditions.

While I will highlight some beliefs, linguistic and classificatory norms, and social dynamics not usually understood in etiological terms, they are not down payments on a new conceptual model. The expanding literature on the social determinants of health does not need another “arrow salad” in which everything causes everything else.1 I want to focus on a few identifiable pathways because the very complexity of multi-causal and multi-level models can and has been used as a pretext to give up population-based health interventions as critics argue that it is too difficult to imagine sensible and effective points of leverage (Farmer, 1999, Tesh, 1988).

Section snippets

Social and structural framing of diagnoses

It has been widely observed in the United States that asthma's incidence, mortality, and hospitalization rates have been rising and that these trends are dramatically worse for the urban poor and ethnic minorities. Biomedical and social scientists have generally looked to micro-environmental (e.g. dust mites and cockroaches) and psycho-biological (stress formulations with immune mechanisms) explanations. Yet, there are clues in the existing clinical and epidemiological literature that the

Technological change affects frame

In a widely cited 1985 article, epidemiologist Alvan Feinstein et al. coined the term the Will Rogers phenomenon to explain an apparent paradox in cancer statistics (Feinstein, Sosin, & Wells, 1985). Feinstein recalled that humorist Will Rogers had once quipped that when the Okies migrated from Oklahoma to California during the Depression, the average IQs of both states went up. This could only happen if the migrating Okies were on average less intelligent than the average Oklahoman but more

Risk, fear, and demand for control

In some other interactions, a cycle of perceived risk, fear, and demand for control has created a self-sustaining process similar in dynamics to the Will Rogers phenomenon, but perhaps even more difficult to recognize, understand, and modify. The history of breast cancer risk (Aronowitz, 2001, Aronowitz, 2007) has shed some light on how this worked earlier in the century. The American Society for the Control of Cancer (later the American Cancer Society) began in the early 20th century a public

Economic and structural contributions to rising number of newly defined health risks and how efficacy is framed

In cancer screening and many other contemporary health risk interventions, we have created a momentum from which it has been difficult to pull back, change direction, or even question. Large scale public health or marketing campaigns have changed the routines of not only ordinary men and women but of general practitioners, radiologists, pathologists, and other important players in the modern “risk factor” rapid reaction force. These interactions among changing epidemiological perceptions,

Dynamics of consumption – feedback loops between consumers and producers

A related and relatively unstudied framing mechanism which has contributed to the emergence and social patterning of many conditions, is the interaction between patterns of consumption and disorders of consumption. Let me illustrate with an example from a recent case study (see also the historical production of the medical consumer by Nancy Tomes, 2001, the rise and fall of the cigarette (Brandt, 2007)).

Schull (2005) carried out a largely ethnographic study of relationship between the producers

Framing health disparities: consequences of or ways to maintain social inequalities?

Researchers and policy makers have not generally questioned why there has been a recent upsurge in interest in health disparities in the United States. Health disparities are not new phenomena and have been noted before. One reason for the prominence of some health disparities might also be a clue to why some patterns exist over and above their production by inequalities in the material conditions that produce health and disease and/or the resources available to prevent and treat disease. Could

Implications

The “framing” mechanisms outlined above may explain only a small fraction of the social patterning of health and disease. They also may act in concert with more direct, material mechanisms. For example, the incidence of breast cancer in the United States and other industrializing countries probably grew considerably in the 19th and early 20th century. The most likely explanations are socio-material ones, that is, a series of demographic and other social factors that together led to greater life

References (26)

  • R.A. Aronowitz

    Lyme disease: the emergence and social construction of a new disease

    The Milbank Quarterly

    (1991)
  • R.A. Aronowitz

    Do not delay: breast cancer and time, 1900–1970

    The Milbank Quarterly

    (2001)
  • R.A. Aronowitz

    Situating health risks

  • R.A. Aronowitz

    Unnatural history: Breast cancer and American society

    (2007)
  • H. Becker

    Outsiders

    (1963)
  • P. Bourdieu

    Distinction: A social critiques of the judgement of taste

    (1984)
  • A.M. Brandt

    The cigarette century: The rise and fall of the drug that defined America

    (2007)
  • J. Cunningham et al.

    Race, asthma, and persistent wheeze in Philadelphia schoolchildren

    American Journal of Public Health

    (1996)
  • P. Farmer

    Infections and inequalities: The modern plagues

    (1999)
  • A.R. Feinstein et al.

    The Will Rogers phenomenon. Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer

    New England Journal of Medicine

    (1985)
  • P. Gergen

    Social class and asthma – distinguishing between the disease and the diagnosis

    American Journal of Public Health

    (1996)
  • J. Greene

    Prescribing by numbers: Drugs and the definition of disease

    (2006)
  • R. Hertz et al.

    Racial disparities in hypertension prevalence, awareness, and management

    Archives of Internal Medicine

    (2005)
  • Cited by (86)

    View all citing articles on Scopus

    This paper benefited from the thoughtful comments of David Asch, Charles Bosk, Charles Rosenberg, Jason Schnittker, Steve Feierman, and the Robert Wood Johnson Foundation (RWJF) Health and Society Scholars at Penn. This work was supported in part by a RWJF Investigator Award in Health Policy.

    View full text