The chapters that follow in this part of the report (Chapters 4-8) present a systematic examination of potential explanations for the pervasive U.S. health disadvantages documented in Part I. A number of factors distinguish the United States from other countries, but their contribution to the nation’s health is unclear. For example, the United States has a very large economy and a large geographic footprint. At 3.7 million square miles, its land mass is much larger than any other high-income country but Canada (United Nations, 2012b) and it encompasses large rural expanses, although comparable cross-national data on the urban-rural mix are limited.1 The United States is also a much younger nation than most
1According to the OECD, fully 78 percent of the U.S. land mass is “predominately rural” (defined as more than 50 percent of the population living with fewer than 150 inhabitants per square kilometer). An even larger proportion of the land mass is predominately rural in some other peer countries, including Austria (79 percent), Norway (84 percent), Australia (85 percent), Sweden (90 percent), Finland (93 percent), and Canada (96 percent). Compared with Americans, larger proportions of the populations of Austria, Sweden, Denmark, Finland, and Norway live in predominately rural areas. Among 13 peer countries for which data were available, the United States had the third lowest percentage (3.8 percent) of people living in “remotely rural” areas, defined as having to drive more than 60 minutes to reach a locality with more than 50,000 inhabitants (OECD, 2011d). These definitions differ from those commonly used in the United States to define rural areas, such as those developed by the U.S. Office of Management and Budget and the U.S. Census Bureau, which themselves are not entirely consistent (Aron, 2006).
of its wealthy European counterparts. However, there is little empirical evidence to link these distinctive conditions to adverse health outcomes.2
The panel recognized the need to identify and organize the leading factors that could plausibly contribute to cross-national health differences, which naturally led to the question of what factors affect health in the first place. This chapter begins by examining the determinants of health and then turns to two issues that helped us frame our approach in looking for potential explanations for the U.S. health disadvantage: the need for a social-ecological perspective that reflects both upstream and downstream influences on health and the need for a life-course perspective that considers influences over time. These concepts were instrumental in persuading the panel to map out a systematic approach to examining the role of health systems (Chapter 4), individual behaviors (Chapter 5), social factors (Chapter 6), the environment (Chapter 7), and policies and social values (Chapter 8).
Attempting to explain the cause(s) of the U.S. health disadvantage leads one to the question of what causes health and disease, a topic that social epidemiologists and other social scientists have studied for decades. Some factors are innate biological characteristics, such as age, sex, and genes, that generally cannot be modified.3 Age and other standard sociodemographic factors are highly predictive of health outcomes but, as documented in Part I, the U.S. health disadvantage persists even after adjusting for these factors.
Both the general public and policy makers often assume that health is determined primarily by health care (Robert and Booske, 2011). Thus, it is reasonable to wonder if the U.S. health disadvantage reflects a deficiency in the U.S. health care system. For example, in contrast to many other countries, a large proportion of the U.S. population is uninsured (National Center for Health Statistics, 2012). But even if health care plays some role, decades of research have documented that health is determined by far more than health care. The seminal article by McGinnis and Foege (1993) highlighted the important role of health behaviors. By some estimates, approximately 40 percent of all deaths in the United States are associated
2Subsequent parts of this report do discuss the role of rural conditions in the United States in contributing to health disadvantages, such as access to medical care.
3Although it is possible that differences in population gene pools or other innate biological characteristics contribute to observed cross-national health differences, these and other nonmodifiable risk factors receive little emphasis in this report due to their unlikely contributory role. However, Chapter 6 does address the important topic of gene-environment interactions and the potential role of epigenetics as a causal pathway for social and environmental influences on health.
with four health behaviors: tobacco use, unhealthy diet, physical inactivity, and problem drinking (Mokdad et al., 2004, 2005).
Yet health care and health-related behaviors are still not the whole story and raise a bigger question: Why are adverse behaviors or deficiencies in health care more common in the United States than in peer countries? Analytic disciplines often draw from different theoretical traditions to answer this question. For example, the behavioral sciences have developed elegant theoretical frameworks for understanding the complex influences on human behavior (Glanz et al., 2008), and economists have proposed human capital models to explain the tradeoffs people make to optimize their “health capital” (Galama and Kapteyn, 2011).
The fact remains that the U.S. health disadvantage is pervasive, cutting across multiple population subgroups and diverse health outcomes (as different as chronic diseases and injuries). This pervasiveness highlights the need to look beyond individual behaviors and choices by examining systemic processes that may influence multiple health outcomes through various specific and often interrelated pathways. Two conceptual models that are useful in framing these more distal, or upstream, health influences are the social-ecological framework and a life-course framework.
Extensive research documents the behavioral and biological consequences of income, occupation, education, and social and physical environments (see, e.g., Adler and Stewart, 2010; Braveman et al., 2011b; Commission on the Social Determinants of Health, 2008; Kawachi et al., 2010). In a knowledge economy, education is the main pathway for economic security, and socioeconomic conditions influence the ability of individuals and families to make healthier choices and to live in healthier neighborhoods. The environment influences one’s ability to engage in healthy behaviors, receive health care, and protect oneself from direct environmental threats. Enhanced recognition that “place matters” has permeated both research and public policy discourse (see California Newsreel, 2008). Television, advertising, and other media also influence health and health-related behaviors, negatively or positively, for example, by promoting (un)healthy foods and (un)safe sex (Center on the Developing Child at Harvard University, 2010; Grier and Kumanyika, 2008; Harris et al., 2009).
All the determinants of health noted above—health care, health behaviors, neighborhood conditions, education, and income—are shaped by both public- and private-sector policies. Decisions made by government officials, business leaders, voters, and other stakeholders affect access to health care, the location of supermarkets, school lunch menus, crime rates, public transportation,
toxic waste sites, employment opportunities, and the quality of schools (Institute of Medicine, 2011d). Environmental factors and policies can affect everyone in the population and can influence multiple disease processes, thereby having consequences for multiple outcomes and offering a potential explanation for the recurring patterns observed in Part I.
The multifactorial nature of social-ecological influences on health has long been recognized (see Evans and Stoddart, 1990; Dahlgren and Whitehead, 1991). Many diagrammatic representations of the social-ecological model have appeared in reports of the Institute of Medicine (e.g., 2003, 2011e), in the conceptual framework developed by the World Health Organization (WHO) Commission on Social Determinants of Health, and in a schematic for public outreach developed by the Robert Wood Johnson Foundation Commission to Build a Healthier America (Braveman and Egerter, 2008). Figure 3-1 shows one such model, which was adopted for Healthy People 2020, the federal government’s national health objectives (U.S. Department of Health and Human Services, 2012a). All such models show (with nesting concentric circles or boxes with multiple arrows) connections between individuals, families, and communities, and the biological, physical, social, and economic environments that surround them. They all emphasize the existence of proximate, or “downstream,” health
FIGURE 3-1 Model to achieve Healthy People 2020 overarching goals.
SOURCE: Secretary’s Advisory Committee on Health Promotion and Disease Prevention Objectives for 2020 (2008, p. 7).
influences (e.g., smoking) that are shaped by distal, or “upstream,” factors (e.g., social norms regarding smoking, tobacco regulations). The models all attempt to highlight societal policies, cultures, and values as the important upstream context for conditions that promote, or undermine, good health downstream.4
Adverse conditions may also affect some population groups more than others,5 thereby contributing to health disparities among vulnerable groups and, possibly, to the health disadvantage experienced in aggregate by the U.S. population (Bleich et al., 2012). The panel focused on its charge to examine and explain cross-national differences in health, not the causes of health differences (or disparities) within countries. Yet we recognize that the two might be interconnected in ways that the field is only beginning to understand. Within-country health disparities may contribute to aggregate cross-national health differences if the former are more pronounced in one country than in another; however, they cannot explain all of the U.S. health disadvantage. The evidence reviewed in Part I suggests that differences persist even among similarly advantaged groups. Understanding why advantaged populations in the United States appear to be less healthy than their counterparts in other countries led the panel to explore upstream processes that affect everyone, including national policies and other aspects of American life.
The complex relationships just described are in play at any given moment of time, but they also play out over time. For example, the absence of green space today may be the product of zoning decisions two decades ago. Such influences also extend over a person’s lifetime: that is, the upstream-downstream continuum can also be a temporal experience for an individual. An individual’s struggle through middle age with exertional angina from coronary artery disease may have originated in adolescence with the adoption of cigarette smoking, perhaps as a coping mechanism for a stressful childhood (Richter et al., 2009) or simply because the family lived in a poor neighborhood where smoking was the norm. In turn, the family’s move into that poor neighborhood may have resulted from financial setbacks that occurred before the child was born. Health trajectories unfold not only over a lifetime, but also across generations as people are
4Policy is typically portrayed as the outer ring of models with concentric circles or as the far left-hand starting point of left-to-right logic models of influences on health outcomes.
5Such effects would be heightened if adverse conditions are more prevalent, more adverse, or have stronger effects in these populations.
subject to changing health influences stemming from family, neighborhood, and public policies.
This continuum of influences offers a potentially important systemic explanation for the higher prevalence of disorders (e.g., obesity, cardiovascular disease), violence (e.g., homicide), and risk factors documented in Part I. Over the past several decades, a great deal of social and biomedical sciences research has been devoted to looking across individuals’ life spans in an effort to better understand developmental and health trajectories over time and, more specifically, how characteristics and experiences early in life may influence health and biobehavioral pathways much later in life (Billari, 2009; Braveman and Barclay, 2009; Braveman et al., 2011b; Gluckman and Hanson, 2006; Halfon and Hochstein, 2002; Keating and Hertzman, 1999; Kuh and Ben-Shlomo, 2004; Lynch and Davey Smith, 2005; Mayer, 2009; Palloni et al., 2009; Pollitt et al., 2005; Power and Hertzman, 1997). The life-course framework casts health as a developmental process influenced by multiple nested social, environmental, and biological spheres that continually interact over the course of one’s life and shape the quality and nature of each person’s growth, health, and development (Halfon and Hochstein, 2002).
Biological, physical, and social influences over the life course can exert both positive and negative effects on health trajectories. As one study explained, health early in life “lay(s) out the foundation for a life time of well-being” (Center on the Developing Child at Harvard University, 2010, p. 1). Conversely, harmful experiences early in life “are built into our bodies and significant adversity early in life can produce biological ‘memories’ that lead to lifelong impairments in both physical and mental health” (p. 3). The same DNA can be expressed differently during development as an outcome of adaptive responses and may be modified over time.
Ontogeny—the development or course of development of an individual organism—involves the development of both organ systems and personality. “Developmental plasticity” refers to how the fetus, child, and adolescent are sensitive to the physical environment and to interaction with other human beings. As noted in Chapter 2, the acquisition of certain cognitive traits and abilities, such as social skills, impulse control, self-image, motivation, and self-discipline, may occur during critical “transition periods” (Viner, 2012). Although protective factors during these sensitive periods can foster resilience and healthy development, harmful exposures may cause an individual to adapt abnormally or in unhealthy ways (Gluckman and Hanson, 2006). These patterns may be relevant to understanding some of the health problems documented in Part I that are more common among American youth than among their peers in other countries. Although research in genetics, molecular biology, developmental neurobiology, psychology, economics, and sociology is only beginning to elucidate the causal pathways that link
early experiences with later disease, at least five general causal pathways have emerged as potentially important.
First, much of the association between early and adult health conditions results from the persistence of material, social, and cultural conditions over time. Children who grow up in poverty not only endure immediate health risks such as poor nutrition and exposure to lead, allergens, and other pollutants, but they are also more likely to face poverty and its consequences as adults (Case and Paxson, 2010, 2011; Case et al., 2005; Currie and Widom, 2010; Delaney and Smith, 2012; Isaacs et al., 2008). To the extent that such exposures remain stable across the life course, disadvantages during childhood and adolescence will persist later in life. Accordingly, poor health status early in life is associated with poor health status during adulthood (Kuh and Ben-Shlomo, 2004; Power et al., 1991; Wadsworth, 1991).
Second, biological exposures during early life, notably to certain infectious diseases—from human papilloma virus to HIV—can produce clinical complications later in life (Elo and Preston, 1996; Finch, 2010; Fong, 2000).
Third, adverse experiences and stimuli can result in biological embedding6 (Hertzman, 2000). For example, Barker and colleagues were among the first to hypothesize that nutritional status in utero and during infancy and early childhood can cause organ damage that is responsible for hypertension, heart disease, metabolic syndrome, insulin resistance, obesity, and diabetes much later in life (Barker, 1998, 1999; Barker et al., 1993; Godfrey, 2006; Hales et al., 1991; Ravelli et al., 1998). Similarly, other investigators have documented associations between early life conditions (e.g., social class, income, nutrition, infections) and subsequent mortality and morbidity from diabetes, cancer, disabilities, and drug use (Freedman et al., 2008; Galobardes et al., 2008; Kuh and Ben-Shlomo, 2004; Kuh et al., 1997; Maty et al., 2008; Mayer, 2009; Palloni, 2006; Palloni et al., 2009; Turrell et al., 2007; Wadhwa et al., 2009; Warner and Hayward, 2006). Research on developmental plasticity is examining the mechanisms through which social and environmental exposures become biologically embedded or reversed during critical periods (Gluckman and Hanson, 2006; Gluckman et al., 2008, 2010). For example, sustained stress is thought to affect the brain and the neuroendocrine and immune systems (see Chapter 6).
Fourth, the relatively new field of epigenetics (see Chapter 6) has documented gene-environment interactions, in which permanent changes in
6Embedding is a process that leaves signatures in the brain and other organ systems that may remain latent for many years and then become manifest and influence adult physical and mental health.
gene expression, triggered by environmental cues, are thought to precipitate chronic conditions (Sandoval and Esteller, 2012) and may be inherited by offspring (Gluckman and Hanson, 2006; Institute of Medicine, 2006b).
Fifth, adversities during childhood and adolescence may cultivate dysfunctional traits (such as poor self-control, limited social skills, lack of perseverance and resilience, and shortsightedness), maladjusted personalities, and susceptibility to antisocial behavior (Caspi, 2000; Caspi and Moffit, 1995; Center on the Developing Child at Harvard University, 2010; Dozier and Peloso, 2006; Elder and Shanahan, 2007; Felitti et al., 1998; Knudsen et al., 2006; Trzesniewski et al., 2006). Stress may alter gene expression in ways that predispose children to specific behavioral phenotypes (Caspi et al., 2002, 2003; Taylor et al., 2006). These traits may in turn influence physical activity, diet, and risky behaviors (e.g., smoking, substance abuse, unprotected sex) (Richter et al., 2009) and they may exacerbate propensities for self-abuse, mental illness,7 suicide, violence, unintended pregnancy, obesity, and diabetes (Dong et al., 2004; Edwards et al., 2003; Felitti et al., 1998; Guyer et al., 2009; Hillis et al., 2004; Horwitz et al., 2001; National Research Council and Institute of Medicine, 2000; Schilling et al., 2007; Shonkoff et al., 2009).8 As noted in Chapter 2, adolescence is especially critical in shaping cognitive and social skills that fuel success in education, careers, and parenting9 (Bowles et al., 2005; Heckman, 2007).
Life circumstances provide an essential context to properly interpret bivariate correlations between health and observed predictors, such as education or income. For example, educated adults may not necessarily experience better health as a direct result of what they have learned in school: the benefits may come from early experiences that foster self-efficacy, a desire to succeed in various endeavors—education among them—and an interest in good health (Pampel et al., 2010; Ross and Wu, 1995). On similar grounds, these factors may be relevant in understanding the evidence in Part I regarding the inferior health status and higher risk exposure of young people in the United States compared with those in other high-income countries. Higher rates of obesity, adolescent pregnancy, motor vehicle deaths, and
7Mental illness has other innate causes and can only partially be attributed to childhood experiences.
8Adults with cardiovascular disease, lung disease, depression, alcoholism, and mental disorders are more likely to report a history of child abuse (Anda et al., 2006; Caspi and Moffitt, 1995; Danese et al., 2007; Moffitt et al., 1992). Longitudinal studies of humans in particular situations (orphaned children, children living in refugee camps) and animals (McEwen, 2000; Zhang et al., 2005) suggest that inadequate maternal attachment and disruptions in protective adult behavior can trigger responses like those observed with child abuse, albeit less extreme (Thompson et al., 2002).
9Thus, adverse conditions during early childhood can breed adverse conditions in subsequent generations.
homicide may have a common explanation in the life experiences of the victims.
In short, existing research points to a variety of pathways that could link early childhood exposures and late adult conditions (e.g., diabetes, cardiovascular disease). The same pathways might also explain correlations between health status across contiguous stages of the life cycle, spanning shorter time intervals, such as between infancy and early childhood or between adolescence and young adulthood.
However, early life experiences are hardly the only influences on health and mortality later in life. It is important to differentiate between the health trajectories experienced by individuals and the epidemiologic trends observed in populations over time. For the latter, “period effects” can play a major role in explaining cross-national differences. For example, increases in smoking during the first half of the 20th century and in obesity in the second half were largely period effects that touched adults of all age groups and eventually extended to adolescents and children as well. Similarly, U.S. tobacco control efforts designed to reduce smoking among adults, which achieved success from the 1960s onward (see Chapter 5), benefited people of all ages.
The panel was acutely aware of one of the most well-known and vexing challenges when studying nonmedical influences on health, that of describing and empirically demonstrating causal pathways between a given health factor and a biological health outcome. In other words, how do the conditions presented by family, community, and national environments get under a person’s skin to affect health? And how do these conditions affect people differently? Even twins do not experience the same health development trajectories over time (Madsen et al., 2010). Unlike the study of clinical interventions or biological effects, research on social, environmental, and policy factors involves more multidisciplinary and varied methodological approaches, including different notions in the epistemology of what constitutes “evidence” (Anderson and McQueen, 2009; Braveman et al., 2011c; McQueen, 2009; Rychetnik et al., 2002; Victora et al., 2004).10 As the Measurement and Evidence Knowledge Network (Kelly et al., 2006, p. 33) of the WHO Commission on the Social Determinants of Health noted:
The data and evidence which relate to social determinants of health come from a variety of disciplinary backgrounds and methodological traditions. The evidence about the social determinants comprises a range of ways of knowing about the biological, psychological, social, economic and material worlds. The disciplinary differences arise because social history, economics, social policy, anthropology, politics, development studies,
psychology, sociology, environmental science and epidemiology, as well as biology and medicine, may all make contributions. Each of these has its own disciplinary paradigms, arenas of debate, agreed canons and particular epistemological positions. Some of the contributions of these disciplines are highly political in tone and intent.
The panel used both a social-ecological framework and a life-course perspective in determining what factors to consider in search of an explanation for the U.S. health disadvantage. Many theories or constructs could be pursued, but the panel identified five domains of particular interest that set the agenda for the next five chapters (see Figure 3-2).
We begin with health systems and individual behaviors, both of which can have important proximate influences on health, and are addressed in Chapters 4 and 5, respectively. The social-ecological model emphasizes that interactions with health systems, individual behaviors, and disease processes themselves are shaped by social factors and the environment. Chapters 6 and 7, respectively, examine their potential contributory role in the U.S. health disadvantage. Finally, Chapter 8 addresses how all these factors might be influenced at the macro level by the societal context of life in America, ranging from life-style to policies, governance, and social values.
Once again, we emphasize that the division of topics across the following chapters is not meant to imply that these categories of health influences operate independently or contribute in a simple additive fashion to the U.S. health disadvantage. The key dynamic trajectories of health, risk factors, socioeconomic circumstances, and physical and institutional environments are all integrally linked and cannot be decomposed in a reductionist fashion. As the following chapters make clear, these synergistic interactions are central to understanding the U.S. health disadvantage.