Introduction
Research over the past four decades has greatly increased our understanding of the proximal determinants of cardiovascular disease, beginning with the pioneering work of Ancell Keys and his colleagues in the Seven Countries Study, which identified the role of cholesterol, 1 and progressing by means of international collaborations such as the MONICA project2 to, most recently, the INTERHEART Study.3 The last of these studies, which included subjects from 52 countries in all regions of the world, showed how nine risk factors (abnormal lipids, smoking, hypertension, diabetes, abdominal obesity, psychosocial factors, consumption of fruits, vegetables, and alcohol, and regular physical activity) could account for almost all the variation observed in myocardial infarction. Studies such as these provide an answer to the question “Why are some people healthy and others not?”.4 However, they raise the further question, first posed by Geoffrey Rose, of why some populations are healthy and others are not.5 These risk factors are not distributed at random in the population but are geographically and socially patterned.
The consequence is that deaths from cardiovascular disease vary markedly among and within populations and their distribution is constantly changing. Beginning as a disease of the upper classes in industrialized countries, its subsequent increases among the lower classes would soon reverse the social gradient.6 More recently, marked declines among those better off have not been seen among the poor, leading to a widening social gap.7 The long-standing advantage enjoyed by countries in Southern Europe is now being eroded,8 at a time when once very high rates in Nordic countries are falling rapidly. In the former Soviet Union death rates have fluctuated dramatically over the course of the past two decades, while remaining consistently very much higher than in Western Europe.9
It is also important to note, in a situation where medicine is increasingly dominated by the search for genetic explanations of disease, that levels of risk change when individuals move from one setting to another, elegantly described in studies of Japanese migrating to Hawaii and California.10 The risk of cardiovascular disease is substantially influenced by the environment that the individual inhabits.
The reasons for some of the patterns that can be observed are either obvious or have been understood through research. Those living in the countries bordering the Mediterranean benefited from their geographical location and consequently their climate, which allowed them to produce the components now referred to as the Mediterranean diet, rich in fruit and vegetables and olive oil. Contemporary increases in mortality in some parts of Greece can plausibly be explained by a failure to reduce smoking rates. In Russia and its neighbors, the high cardiovascular mortality can be explained by, in addition to the traditional risk factors, the widespread practice of episodic heavy drinking, especially of surrogate products such as aftershaves.11 coupled with weaknesses in the healthcare system and, especially, the management of hypertension.12 Within societies, differences in smoking and diet, which are shaped by the environments that people inhabit, can also go some way to explaining social inequalities, and indeed some research has suggested that they may themselves be sufficient.13
However, there is an extensive body of research suggesting that psychosocial factors play a part. The largest of these studies, INTERHEART, found that those suffering a myocardial infarction were significantly more likely to have experienced stress at work or at home, financial stress or major life events in the past year. Put another way, those suffering several episodes of stress or permanent stress at work were more likely to suffer an infarction (respectively, odds ratios (OR) 1.38, 99% confidence interval (CI) 1.19–1.61, and OR 2.14, 99% CI 1.73–2.64), after adjustment for age, sex, geographic region, and smoking. The corresponding figures for several episodes or permanent stress at home were OR 1.52 (99% CI 1.34–1.72) and OR 2.12 (1.68–2.65). Severe financial stress and stressful life events in the past year showed similar associations (OR 1.33, 99% CI 1.19–1.48, and OR 1.48, 99% CI 1.33–1.64, respectively).14
This chapter begins from the intentionally simplified position that individuals inhabit three often overlapping communities, defined by their families, their work environment, and their social life. In each of these they have relations with other individuals. The roles that they fulfill and the relationships that they experience within each of these communities may impact on their risk of cardiovascular disease or their ability to survive it. Individuals also live within a larger community, where ties may be much looser but which surrounds them and whose environment is characterized by its physical, political, economic, social, and cultural attributes. This environment may impact on their risk of cardiovascular disease by virtue of the ease which it enables them to make healthy choices or constrains them from doing so. This framework is illustrated in Figure 17.1, with five individuals who are connected to each other through families, work, and social ties. Much of the rest of this chapter will focus on the evidence linking the nature and intensity of these connections, in these three domains, to cardiovascular disease, and on the way that the environment in which people live influences their degree of risk.
The most intense social relationships are within the immediate family. There is considerable evidence that marriage is protective against cardiovascular disease, at least for men, who have been most intensively studied. Some research has been able to explain this on the basis of differences in risk factors.15 Divorce is associated with taking up smoking or, among ex-smokers, relapsing, for both men and women.16,17 In the Health Survey for England, single men and those with low social support were less likely to have hypertension adequately treated.18 A small Israeli study found a more harmful pattern of lipids among divorced, separated or widowed women than in their married counterparts that was unexplained by age, smoking or physical activity.19
However, there is also considerable evidence that the excess risk associated with being unmarried persists after adjustment for conventional risk factors. In the British Regional Heart Study, single (never-married) men were at increased risk of dying from cardiovascular disease (relative risk (RR) 1.5, 95% CI 1.0–2.2) after adjustment for potentially confounding variables.20 Men who were divorced or widowed at recruitment did not exhibit an increased risk but risk was greater among those who divorced during follow-up (RR 1.9, 95% CI 0.9–3.9), again unexplained by conventional risk factors. An increased risk associated with divorce was also found in the Multiple Risk Factor Intervention Trial, which included men without evident cardiovascular disease at recruitment but who were at above-average risk on the basis of conventional risk factors.21 In contrast, an analysis of the Whitehall Study found an excess risk of ischemic heart disease mortality associated with widowhood (RR 1.46, 95% CI 1.08 –1.97). As it is possible that this could be associated with acute stress associated with bereavement, a subsequent analysis excluding deaths occurring in the first two years was undertaken, with the effect persisting.22 In a cohort of over 90 000 Japanese men and women followed over 10 years, never-married men, compared with married men, were at a greater risk of cardiovascular mortality (RR 3.05, 95% CI 2.03–4.60) after adjustment for potentially confounding variables. Never-married women were also at greater risk, although less so than for men (RR 1.46, 95% CI 1.15 –1.84). An increased risk was also seen among divorced and widowed men but not women.23
Some studies have also looked at the association between marital status and survival among those with established cardiovascular disease. Among men and women found to have significant disease at angiography, those who were unmarried and without a close confidant were more likely to die within three years than those who were married with a close confidant (OR 3.34, 95% CI 1.8–6.2).24 Among those in the placebo arm of the Multicenter Diltiazem Post-infarction Trial, living alone was associated with an increased risk of major cardiac events, which included deaths and non-fatal myocardial infarction.25 However, other research, in a German MONICA sample, found an increased risk of death among men living alone but not among women living alone, after adjustment for conventional risk factors,26 with another study showing that men but not women living alone are also disadvantaged in terms of survival following myocardial infarction.27
There has long been evidence that social networks are ben-eficial to health, with lower all-cause mortality observed in those engaging most frequently in activities such as cultural events.28
A few studies have looked at the effects of social networks on the incidence of myocardial infarction. A Swedish study following men over 15 years examined the association between incident coronary heart disease and two measures of social support: “emotional attachment” and “social integration”. The hazard ratio (HR) for those in the highest quartile for social integration compared with the lowest was 0.45 (95% CI 0.24–0.84). The corresponding figure for emotional attachment was 0.58 (95% CI 0.37–0.91).29 A cohort of over 13 000 Swedes, followed for three years, found that low social participation was significantly associated with first-time acute myocardial infarction, after adjustment for age, sex education, economic stress, daily smoking, physical activity and body mass index.30 However, some other studies have not found an increased risk of incident disease in those with poor social networks.31,32
Other research has used mortality as an endpoint. For example, a study of approximately 30 000 male health professionals found that those defined as socially isolated men (not married, fewer than six friends or relatives, no membership in church or community groups) had a greater risk of cardiovascular disease mortality than those with the highest level of social networks at both four32 and 10 years (relative risk. adjusted for a range of factors including traditional cardiovascular risk factors, 1.82, 95% CI 1.02–3.23).33 In the Japanese Komo-Ise Study over 11 000 subjects were followed over seven years. Men who did not participate in hobbies, club activities or community groups had significantly higher relative risk of circulatory system mortality (RR 1.6), after adjustment for established risk factors.34
This seeming discordance between some of the evidence on incidence and the much more consistent finding with mortality has focused attention on how social networks may be more important in reducing case fatality than in preventing incident disease.35 Participants in the Beta Blocker Heart Attack Trial who were socially isolated were twice as likely to die during three years of follow-up than those who were not after adjustment for conventional risk factors.36 Those who scored highly on social isolation and on a measure of life stress (which included divorce, exposure to traumatic events, and financial difficulty) had an over fourfold increase in risk.
Participants in a population-based study in New Haven who were hospitalized with myocardial infarction were more likely to die in hospital and at up to one year if they had no social support. This result was consistent for men and women, at all ages, and regardless of the severity of disease. In a multivariate model adjusting for sociodemo-graphic, psychosocial, and clinical factors, the excess risk (odds ratio) of mortality among those with no support compared to those with at least one source of support was 2.9 (95% CI 1.2–6.9).37
There is also some evidence that the nature of networks is important, with one study of patients undergoing coronary bypass grafting finding that those who were not members of voluntary organizations, including religious ones, or who do not draw support from faith, were seven times more likely to die.38
The preceding studies examined the networks that individuals inhabited. There is also a growing body of research looking at the characteristics of the communities in which individuals live. Some of this work employs the concept of social capital, a term that has been defined in a variety of ways.39–41 However, the definitions are all characterized by notions that it relates to a community and not an individual and that its existence brings benefits that go beyond those who generate it. It is often measured by characteristics such as trust in others, perceived reciprocity, and density of membership in civic organizations.
Social capital can be assessed in a variety of ways. In one ecologic study, social surveys undertaken in 39 American states were used to obtain information on levels of trust and of membership of voluntary organizations, which were then related to state-level mortality from cardiovascular disease. Lower levels of both were associated with increased cardiovascular mortality, after adjusting for poverty.42 Another study categorized Swedish neighborhoods in terms of violent crime and unemployment rates.43 Residents aged 35–64 were followed up for a year. Those living in neighborhoods with higher levels of violent crime or unemployment had a higher incidence of coronary heart disease, affecting both men and women. For violent crime, the odds ratios were 1.75 (CI 1.37–2.22) and 1.39 (CI 1.19–1.63) for women and men, respectively. The corresponding figures for neighborhood unemployment were 2.05 (CI 1.62–2.59) and 1.50 (CI 1.28–1.75). These figures changed very little when a range of individual variables were included.
Finally, one study has examined prognosis. This used the Petris Social Capital Index, which characterizes neighborhoods according to level of employment in religious and community organizations.44 Of almost 35 000 people in California previously admitted with acute coronary syndrome, the recurrence rate was lower in neighborhoods with higher scores on the index (after adjustment for medication, previous revascularization, and standard sociodemographic variables), but only among individuals living in poorer areas.
An individual will only spend a relatively small proportion of their entire lives at work, perhaps 40 hours a week between the ages of 20 and 65. Yet, their experiences at work have a disproportionate impact on their lives, influencing how they spend their leisure time, their well-being in retirement, and their psychological well-being.
Interest in the potential impact of the work environment arose from research suggesting that observed social inequalities in cardiovascular disease could only partially be explained by the risk factors then being studied45 (a more limited set of those included in more recent research such as the INTERHEART study).3 It has grown in the light of concerns about the changing nature of work and, in particular, the greater insecurity of employment in many countries.
The literature on the health effects of the work environment is dominated by two explanatory models: the demand–control model and the effort–reward imbalance model. Both begin from the position that a work environment that provides opportunities for contributing in a meaningful way to the task in hand and which provides appropriate feedback is likely to be health enhancing while an environment that fails to do so may be health damaging. Two elements in particular have been highlighted. The first is self-efficacy or the belief that one can do what is asked of one.46 An environment that encourages self-efficacy will allow those in it to make greatest use of their skills and to exercise control over their work. The second is self-esteem or feeling of one’ s own worth. An environment that encourages self-esteem will allow an individual to interact positively with others and obtain feedback following good performance. In contrast, repeated failure to recognize effort and success can be corrosive.
The demand–control model was developed in Sweden in the early 1980s.47 In a study that combined a cohort and case–control design, those in what was described as “a hectic and psychologically demanding job” had an increased risk of developing cardiovascular disease and of dying from it, as had those in jobs where they had little discretion about how they did their work or how they scheduled it. The associations were especially marked among those with the least education and persisted after controlling for age, education, smoking, and overweight.
The demand–control model has been developed to include the extent to which the work environment encourages active learning, by which the worker can develop their self-efficacy, and some researchers have suggested inclusion of a third dimension, social support at work.48 It has also been suggested that a high level of self-efficacy has effects that extend beyond the work environment into other aspects of the worker’ s life, including their involvement in social and cultural activities which, as the previous section showed, may itself be beneficial for health.49
The second, and complementary, model involves effort–reward imbalance.50 It arises from the concept of social reciprocity, whereby individuals expect that their efforts will be rewarded appropriately. This may break down in a number of situations, such as where there are few alternative sources of employment, but can also arise in individuals who underestimate the demands placed upon them while overestimating their ability to cope.
Interpretation of the empiric evidence is complicated by the diverse range of measures employed to capture the work environment, the use of both objective measures and self-report, the use of both cohort51 and case–control designs,52 differing outcomes, and differences in the extent to which other risk factors are adjusted for. Much of the existing literature has been summarized in some recent systematic reviews.53,54
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
