Stress and Health 1 The Effects of Job Strain on Risk Factors for Cardiovascular Disease Arie Shirom and Galit Armon Faculty of Management, Tel Aviv University Shlomo Berliner and Itzhak Shapira Tel-Aviv Sourasky Medical Center, Tel-Aviv, and Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel Samuel Melamed Department of Epidemiology and Preventive Medicine, Sackler Faculty of Medicine, Tel Aviv University, Israel Please address future correspondence regarding this study to Arie Shirom, Faculty of Management, Tel Aviv University, Ramat Aviv, PO Box 39010, Tel Aviv 69978, ISRAEL, or email to: [email protected] Stress and Health 2 The Effects of Job Strain on Risk Factors for Cardiovascular Disease The Objectives and Scope of the Review In our review, we focus on work-related psychological stress. Within this stress research domain, we exclude from our review event-based types of stress – including acute and critical job events such as being demoted or going on involuntary vacation (Eden, 1982, 1990). We also exclude work-related hassles, such as being caught in a traffic jam while commuting to work (cf. Gajendran & Harrison, 2007). It follows that we focus on chronic stress at work. Researchers may disagree on the conceptual definition of work-related chronic stress (Cooper, 1998; Monroe, 2008). There is basic agreement, however, about the notion that work-related chronic stress, hereafter referred to simply as stress, may be implicated in cardiovascular disease risk factors, specifically physiological ones, such as elevated cholesterol and blood pressure levels, and in certain maladaptive behavioral responses (Aboa-Eboule et al., 2007; Chandola et al., 2008; Chandola, Brunner, & Marmot, 2006; Williams, 2008). We decided to focus on the effects of work-related stress on risk factors for cardiovascular disease (CVD). We define cardiovascular disease (CVD) as a composite of coronary heart disease, stroke, and cardiovascular mortality. This focus is due to the fact that CVD, including myocardial infarction (MI) and stroke, is a principal cause of death in most economically advanced countries; it is associated with multiple physiological, psychological, and socio-demographic risk factors that often interact in complex causal paths (Brotman, Golden, & Wittstein, 2007; Kiecolt-Glaser, McGuire, Robles, & Glaser, 2002; Williams, 2008). Chronic stress could be directly implicated in CVD by causing spasms of coronary blood vessels, Stress and Health 3 electrical instability in the heart conduction system, and abnormal heart rhythms (cf. Williams, 2008). Chronic stresses are thought to influence the pathogenesis of CVD by causing negative affective states such as burnout (Melamed, Shirom, Toker, Berliner, & Shapira, 2006), and anxiety and depression (Suls & Bunde, 2005), which in turn exert direct effects on maladaptive behavioral and physiological responses. There are two major physiological mechanisms that are considered as the most likely mediators of the effects of chronic stress on CVD: the hypothalamic-pituitary-adrenocortical axis (HPA) and the sympathetic-adrenal-medullary (SAM) system (Miller, Chen, & Zhou, 2007). Prolonged or repeated activation of the HPA and SAM axes can interfere with their control of other physiological systems, and could result in increased risk for a variety of physical and psychiatric disorders (McEwen, 2007). Cortisol, the primary endocrine response of the HPA axis, regulates a broad range of physiological processes, including the metabolism of fats and proteins represented in our review by blood lipids, and anti- inflammatory responses, represented in our review by biomarkers of micro-inflammation in the serum. SAM axis activation is associated with the secretion of catecholamines, which – interacting with the autonomic nervous system – exert regulatory effects on many organ systems in the body, including the cardiovascular system (Cohen, Janicki-Deverts, & Miller, 2007). SAM axis activation is represented in our review by both blood pressure and sleep disturbances. The chapter begins by describing a general theoretical perspective within which our review is embedded. The general theoretical perspective provides a comprehensive, system- based view of the antecedents of stress- CVD risk factors linkages. We then briefly review three leading theoretical models that have been used to investigate stress-CVD risk factors associations: the Person-Environment Fit model, the Effort-Reward model, and the Job Demand- Stress and Health 4 Control-Support model (JDC-S) model, also referred to as the Job Strain model (hence the term job strain in the title). We explain why we chose to focus, in the following sections of our review, on the JDC-S model. We then use the JDC-S model to present what is known about the effects of work-related stress on four major risk factors for CVD: elevated levels of blood lipids, blood pressure, and micro-inflammation biomarkers, and sleep disturbances. These risk factors represent only a subset of possible physiological and behavioral strains that may be impacted by work-related stress. For example, the effects of stress may include alterations in neuroendocrine factors, the autonomic nervous system, and immune functions. The concluding section discusses the limitations of this review and highlights promising avenues for future research in this field. Above we explained several physiological considerations that guided the choice of the CVD risk factors covered here. There were additional considerations, related to the availability of empirical studies and the contents of other chapters in this volume. This review focuses on empirical studies; therefore, a body of such studies should exist. Additionally, an explicit attempt was made to avoid duplication with other chapters of this volume, including chapters that specifically cover the maladaptive health responses of psychological distress, drug abuse, and alcoholism. The broad scope of this review necessarily limits the depth of the presentation. Readers should note that the range of the literature covered probably reflects the author’s personal viewpoints on several key issues. A General Framework for the Study of the Health Consequences of Stress at Work The theoretical model guiding this chapter is represented in Figure 1. Within the model, an individual’s state of health is viewed as being determined by multiple factors, including heredity, environment, early background, and socioeconomic influences. This theoretical model Stress and Health 5 draws on earlier conceptualizations (Macik-Frey, Quick, & Nelson, 2007; Quick, Quick, Nelson, & Hurrell, 1997, pp. 65-89). Among the multiple causal chains leading to maladaptive health responses is the effect of work-related stress. This effect is depicted as being moderated by individuals’ coping resources and personality factors. To simplify the presentation of the main effects, several arrows indicating moderating effects were omitted from Figure 1. We focus on the hypothesized arrow that leads from work-related chronic stress to maladaptive physiological and behavioral responses, primarily because of the considerations detailed above. Stress is posited in Figure 1 as precipitating the development of maladaptive health responses, like when it is implicated in raising a person’s blood pressure from normal to borderline. The following is a brief discussion of the theoretical model presented in Figure 1. It is introduced by a description of the conceptual approach followed by the definitions of stress and maladaptive health responses. Insert Fig.1 about here Early reviews of the vast area of work-related stress and physical health (Danna & Griffin, 1999; Ganster & Schaubroeck, 1991; Mackay & Cooper, 1987) mostly followed the theoretical framework depicted in Figure 1, defining health and well-being broadly to include psychological and physical health. By maladaptive health responses, we refer to a subclass of what has been labeled strain in the Michigan model, namely any deviation from the normal state of responses of the person (French, Caplan, & Harrison, 1982). This definition of strain included psychological strain, such as job dissatisfaction and anxiety, physiological strain like high blood pressure, and behavioral symptoms of strain such as sleep disturbances. Continuing high levels Stress and Health 6 of strain were postulated to affect morbidity and mortality levels (French et al., 1982). In this review, we refer only to the latter two types of strain. There are several reasons for introducing the model depicted in Figure 1 in this chapter. First, as argued by several researchers (Kasl, 1996; Marmot, Theorell, & Siegrist, 2002), studies of the relationships between stress and maladaptive health responses need to maintain a broad conceptual perspective of the etiology of these responses. Specific etiological factors leading from the work environment to health responses are embedded in a complex matrix of additional psychosocial influences. There are several classes of variables that were included in Figure 1, but were not discussed or reviewed here because of space limitations. The potential usefulness of each of those panels needs to be considered by future researchers. Salient examples are socioeconomic indicators (cf. Gallo & Matthews, 2003), stable individual differences (cf. Smith, Glazer, Ruiz, & Gallo, 2004; Smith & MacKenzie, 2006), and work role and work environment characteristics that represent individuals’ exposures to earlier work and job experiences (Theorell, 1998). Figure 1 depicts several bi-directional arrows. These double-headed arrows represent interactions or non-recursive processes between panels of variables. To illustrate, the bi- directional arrow between psychological and physiological maladaptive responses represents reciprocal feedback loops that can occur, as when distress - such as depression or burnout - affects the immune system (cf. Melamed, Shirom, Toker, Berliner, & Shapira, 2006). Again, given the confines of this review, it was not possible to discuss each double-headed arrow in detail. Stress and Health 7 The role of panels not discussed in our review could be illustrated by taking as an example the role of socioeconomic disadvantage, considered to have direct and indirect influences on maladaptive health responses. Decades of research have shown that socioeconomic status is a significant predictor of stress, strain, and state of health (Banks, Marmot, Oldfield, & Smith, 2006; Hemingway & Marmot, 1999; Marmot, 2006). Socioeconomic status differences are found for rates of morbidity and mortality for almost every disease and health condition (Adler et al., 1994). Components of socioeconomic status, income, education and occupation shape individuals’ early life experiences, including early-age health habits like diet and exercise, and significantly influence their work experiences, including access to coping resources such as social support at work (cf. Danna & Griffin, 1999). Researchers have often posited a strong relationship between perceived stress - an individual’s coping resources and coping mechanisms - and the etiology of stress-related maladaptive health responses (Lazarus, 1999; Taylor & Stanton, 2007). How an individual handles stress plays an important role in determining the health outcomes of the individual’s encounter with stress. Coping may be loosely defined as things we think and actions we take to ameliorate or remove the negative aspects of stressful situations, including indirect coping like avoidance (Taylor & Stanton, 2007). The ability to cope with stress is represented in Figure 1 by the panel of work-related coping resources. These resources interact with individuals’ subjective appraisal to determine their experienced stress. If a situation is not appraised as taxing or exceeding one’s coping resources, it is not likely to be experienced as stress (Lazarus, 1999). Personality factors like hardiness represent additional coping resources. Because of space limitations, this chapter does not cover the issue of effective coping mechanisms, which may Stress and Health 8 prevent psychosocial and physiological disequilibria that may in turn lead to stress-related illnesses. Adaptive and maladaptive responses to stress represent a complex set of an organism’s reactions intended to reestablish psychosocial and physiological equilibriums. As indicated, we focus only on a specific set of risk factors for CVD. The hypothesized effects of stress may appear in any combination of the physiological, behavioral, and psychological domains of strain. To illustrate, high blood pressure, sleep disturbances and high levels of "bad" serum cholesterol and obesity often co-occur. The synergic relationships among the panels of Figure 1 indicate that there is not any single consistent maladaptive health response applicable to most people in all work situations. This basic premise of inter-individual variability in stress response is related to the direct and indirect effects of coping resources and coping effectiveness considered above, and in addition to other individual difference variables depicted in other panels of Figure 1. Models Explaining the Effects of Work Stress on Physiological Risk Factors In recent years, occupational health researchers have devoted considerable attention to possible paths of influence linking work and job characteristics with employees’ physical health. They have been guided in their attempts to gain additional understanding of the pathways linking the world of work and employee physical health by several important and often used models. Each of these models represents a distinct way of reducing the complex reality into a comprehensive yet parsimonious model. Each of these models focuses on specific core elements in order to explain work-related physical health. One of the most important models is referred to as the Job Demands-Control-Support (JDC-S) model; for reasons explained below, we focus on this model in our review. However, we also briefly review in this section two additional models Stress and Health 9 that have been used to explain the effects of chronic work-related stress on the risk factors for CVD considered in our review. We describe the two other models largely because they are based on similar theoretical principles and therefore could be combined with the JDC-S model in future research explaining the pathways linking work-related stress with risk factors for CVD. The Person-Environment Fit Model One of the earliest (French & Caplan, 1973) models focusing on stress and health outcomes is the Person-Environment Fit model (P-E Fit). It has been widely applied to predict a variety of risk factors for disease (Edwards, Cable, Williamson, Lambert, & Shipp, 2006; Edwards, Caplan, & Van Harrison, 1998), including those that we focus on (cf. French, Caplan, & Harrison, 1982). Fit, in this model, includes the relationships between environmental supplies and individuals' values and needs, referred to as the S-V (supplies-values) Fit, and the relationships among environmental demands and individuals' abilities, skills and knowledge, referred to as the D-A (demands-abilities) Fit. The model postulates that the more pronounced a misfit, either S-V or D-A, the higher the level of the resulting strains will be. An additional postulate of the model is that the major components of S-V, and in turn also D-A, interact to influence one's level of strain (Edwards et al., 1998). Many additional types of Person- Environment Fit that could conceptually exist were described in a major conceptual review of the area (Kristof, 1996). Kristof (1996) pointed out that the type of fit mainly investigated up till now has been the D-A rather than the S-V. Several other approaches to the study of stress-strain relationships have also incorporated elements of this model, such as Cybernetic Stress theory (Edwards, 1998). A recent meta-analysis of the P-E Fit literature at large, including studies that followed Stress and Health 10 the specific P-E Fit model described above (Kristof-Brown, Zimmerman, & Johnson, 2005), concluded that fit was strongly associated with several attitudinal and behavioral outcomes. For example, Person-Job Fit was found to be strongly linked with job satisfaction, job performance, and turnover, while Person-Organization Fit was closely associated with organizational commitment. However, the number of studies linking the model with physical health related outcomes was found to be small (Schnall, Landsbergis, & Baker, 1994), and therefore this was not the model chosen to guide our review. Effort-Reward Imbalance at Work Model The Effort-Reward Imbalance (ERI) model (Siegrist, 1995) builds upon the notion of social reciprocity, a fundamental principle of interpersonal behavior which lies at the core of employment relationships. In the context of the ERI, social reciprocity is interpreted as representing the norm of return expectancy. Return expectancy refers to employees' expectations that the effort they invest at work would be equal to the rewards they receive. Efforts, in the context of ERI, represent job demands and requirements that are imposed on the employee. Rewards, in turn, refer to money, job security, self-esteem, and career opportunities, mostly distributed by the employer (but also by society at large). Reward in the ERI model is probably closely related to the notion of supplies in the P-E Fit model, while efforts resemble the notion of demands in the P-E Fit model. Therefore, it could be argued that the ERI model is embedded in the PE-Fit model. A job situation characterized by high efforts and low rewards represents a reciprocity deficit. Perceived lack of reciprocity is hypothesized to lead to strong negative emotions. These negative emotions, in turn, lead to sustained autonomic and endocrine activation and to negative health outcomes (Ursin & Eriksen, 2007). The wider the discrepancy between the