THE OCCUPATIONAL STRESS INDEX An Approach Derived from Cognitive Ergonomics and Brain Research for Clinical Practice
.D...
96 downloads
506 Views
996KB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
THE OCCUPATIONAL STRESS INDEX An Approach Derived from Cognitive Ergonomics and Brain Research for Clinical Practice
.DUHQ%HONLü
CAMBRIDGE INTERNATIONAL SCIENCE PUBLISHING
Contents
1
2
Acknowledgements
iv
Introduction
1
The Need for the Occupational Stress Index—An Approach Derived from Cognitive Ergonomics and Brain Research for Clinical Practice
1
The Impact of Stressful Work on Health
4
How research in this area has proceeded: The crucial role of models The Job Strain Model The Effort Reward Imbalance Model
4 4 5
Evidence concerning exposure to psychosocial work stressors and adverse health outcomes
5
Cardiovascular Disease Hypertension Musculo-Skeletal Disorders Mental Health Outcomes Occupations with Evidence of Risk for Adverse Stress-Related Health Outcomes Professional Drivers Health Care Professionals Teachers Air Transport Professionals Sea Pilots Other Occupational Groups
5 10 11 11
12 12 13 14 14 15 15
ii
3
4
5
How Insights from Cognitive Ergonomics and Brain Research Inform our Assessment of the Work Environment
16
How we handle Information: A Neurophysiologic View Levels of Information Transmission: High Demands versus Underload Knowledge-Based versus Skill-Based Information Processing Threat Avoidant Vigilance The Conflict Dimension Physically Aversive Exposures
16 22 27 30 34 36
Occupation-Specific versus Generic Self-Report Measures To Assess Workplace Exposures The Occupational Stress Index as an Additive Burden Model to Help Bridge The Gap
38 38
The Occupational Stress Index (OSI) Model – Revised Version
40
The Organization of the OSI Levels of Information Transmission Stressor Aspects The Two-Dimensional Matrix The OSI Summations
41 41 41 42 42
The Generic versus Specific OSI’s The Generic OSI Occupation Specific OSI’s OSI for Professional Drivers OSI for Physicians OSI for Clerical Workers OSI for Teachers Other Specific OSI’s in Development
43 43 44 44 49 50 50 51
The Occupational Stress Index In Clinical Practice Preparing a Narrative Occupational History, which includes Psychosocial Stressors as This Informs a Work-Related Diagnostic and Management Plan
53 53
The Clinician’s Challenge An Approach to taking a Work History which includes Psychosocial Stressors, Based upon the OSI
53 54
iii Part I—Pedagogical Occupational Histories based upon the OSI A. Physician Specialist in Neuropsychiatry--using the OSI for MD’s B. Long Route Truck Driver--using the OSI for professional drivers C. Administrative Assistant--using the OSI for those who work daily with computers D. Automobile Assembler—using the Generic OSI
58 58 79 92 120
Part II--Pedagogical Clinical Cases based upon the OSI-derived Occupational Histories 139 A. Physician with Paroxysmal Supraventricular Tachycardia 139 B. Truck Driver with silent myocardial ischemia & complex arrhythmias 142 C. Administrative Assistant with Angina Pectoris 146 D. Automobile Assembler – Status Post Acute Myocardial Infarction 150
6
7
Assessment And Approach To Management: Answers To Questions From The Cases
154
The Revised OSI Questionnaires and Score Sheets
163
The Revised Version of the Generic OSI Questionnaire Score Sheet
165 180
The Revised Version of the OSI for Professional Drivers Questionnaire Score Sheet
197 206
The Revised Version of the OSI for Physicians Questionnaire Score Sheet
219 233
The OSI for Teachers Questionnaire Score Sheet
248 264
Conclusions and Future Perspectives
280
Bibliographic References
282
iv
Acknowledgements I would like to express my appreciation to family, teachers and colleagues for support, inspiration and dialogue. I also wish to thank my patients and participants in research projects, as well as many other working people who have shared their concerns and experiences. For their cooperation and help throughout the years, I thank many librarians. Thanks are also due to my students who have participated in these research and pedagogical endeavors. In the most recent period, I have had a chance to communicate electronically with numerous clinicians and researchers throughout the world about the Occupational Stress Index. I am grateful for their queries and interest – these have been a prime motivation for this work. The pedagogical cases and parts of Chapter 3 were prepared for Courses in Occupational Health Psychology and Occupational Cardiology at the University of California, Los Angeles and Irvine, respectively, and this was supported, in part, by a teaching grant from the U.S. National Institutes for Occupational Health and Safety (NIOSH). Permission to reproduce figures was granted by Elsevier Science Publishers for Fig 3.1 and 3.17; and by Taylor and Francis publishers (http://www.taylorandfrancis.com) for Fig. 3.12 and 3.13. Those who have participated in the development of the specific OSI’s are acknowledged in the corresponding sections of Chapter 4. This work is dedicated to the memory of Professor Ludwig Edinger, my great granduncle, whose neuroanatomic discoveries have been an impetus for inquiry.
Chapter 1
Introduction The Need for the Occupational Stress Index—An Approach Derived from Cognitive Ergonomics and Brain Research for Clinical Practice Paraphrasing the late Professor Bertil Gardell, a pioneer in the effort to develop multi-facetted strategies to humanize the work environment: work is one of the most important potential sources of social and psychological well-being, which can provide much of the meaning and structure in adult life (Gardell 1987). Unfortunately, however, for many working people, this potential is far from reality. Instead, the contemporary work environment has all too frequently become the locus in which employed adults spend the majority of their waking hours performing activities that are characterized as demanding, constraining, and otherwise stressful. Reflecting pressures of global competition, trends in working life are towards increasing job demands, working hours and job instability. Growing dependence on computer technology, which could improve working life, has de facto lead to greater workload and pressure. The toll taken by unhealthy work is enormous. Mental health problems and other stress-related disorders are recognized to be the largest overall cause of premature death in Europe (Levi 2002). Cardiovascular disease, the major cause of morbidity and mortality in the industrialized world, and for which the stressful work environment is increasingly recognized as an important risk factor (Belkic 2000a, Karasek 1990, Kristensen 1998, Schnall 2000), is projected within the next twenty years, to become the leading cause of death worldwide (Braunwald 1997). In purely economic terms, a recent estimate is that job-related cardiovascular disease costs for the U.S. alone amount to $22.5 billion annually (Leigh 2000). At the same time, there have been tremendous strides made in our understanding of the human nervous system. This knowledge could potentially be used to help create working conditions that are in harmony with human needs. As stated in the Tokyo Declaration on Work-Related Stress and Health (1998): “The growth of neuroscience and stress science has allowed elucidation of the links between social structures and processes (at work and outside it), the way in which these are perceived and appraised and the resulting interaction between the central nervous system and other organ systems to promote or counteract workers' health, based on a bio-psycho-social approach to all relevant aspects of the [human] – environment ecosystem and its dynamics” (p. 2). One of the critical stumbling blocks in this process has been confusion between the objective characteristics of a given set of working conditions, and how the individual perceives and responds to these. Rohmert’s formulation (1971) is a helpful starting point for clarification: “It seems to be an advantage to distinguish between the evaluation of work and the assessment of the human operator. The independent factors of work (or control tasks) affect as stressors, stress. The dependent physiological, psychological and social
2 reactions…[of a given individual] due to these stressors affect strain. The amount of strain always depends upon the given components of stressors and the individual capacities of [a given person] at work” (p. 546). This distinction becomes particularly important for occupations in which many of the most taxing stressors are not readily apparent, and it is here, as we will outline, that insights from brain research and cognitive ergonomics prove to be invaluable. Our basic motivation in this work is practical—to offer the reader a way to apply a methodology derived from cognitive ergonomics and brain research for assessing work stressors: the Occupational Stress Index (OSI) (Belkic 1989). In order to do so, we needed to present some basic information about how the brain receives and handles information: the aversions and affinities of the human nervous system in relation to the environment. This is done very briefly in Chapter 3, with illustrations to facilitate this aim, and, hopefully, to spark interest and curiosity to delve further. This also represents a unified vantage point from which various types of stressors: “mental”, physical, ergonomic and organizational can be considered in concert, with respect to their effect upon the central nervous system. Chapter 3 provides an introduction to the OSI framework, which is presented formally and in detail in Chapter 4, entitled, “Occupation-Specific Versus Generic Self-Report Measures To Assess Workplace Exposures: The Occupational Stress Index as an Additive Burden Model to Help Bridge the Gap.” Here, we discuss two divergent trends in occupational psychosocial research. One is represented by theory-based, generic approaches, which tend to be remote from actual work experiences, and therefore are often not helpful for assessing within-occupation variance, the very level at which intervention strategies are developed, in practice. The other trend has been that of occupation-specific inquiries, which provide rich, detailed information often useful for identifying key areas for intervention. These have usually been so focused upon a given occupation, that more generalizable conclusions based upon between-group analyses are often missed. This is precisely where the OSI can offer a potential solution, namely by providing a series of occupation-specific instruments that are all mutually compatible within the OSI framework: allowing betweenoccupation comparisons, but at the same time far more operationalized and streamlined than a single generic instrument. The utility of this approach is illustrated with respect to the OSI for professional drivers, which has been applied most widely, and for which within-group and between-group criterion validity have been demonstrated. We also present initial results from the OSI for Physicians, and describe the progress made in developing OSI’s for teachers, factory workers, those who work daily with computers, as well as for other occupational groups. We especially focus on the clinical arena, which has heretofore remained relatively isolated from the rapidly growing field of occupational psychosocial health. In Chapter 2, for that audience in particular, we describe the Job Strain (Karasek 1979) and Effort Reward Imbalance Models (Siegrist 1991, 1996), and how their evolution and wide application have represented a key turning point for this field of research. We then briefly summarize the large body of empirical evidence concerning exposure to Job Strain and to Effort Reward Imbalance and health outcomes including cardiovascular and cerebro-vascular disease, hypertension, musculo-skeletal problems (repetitive-motion disorders) and adverse mental health outcomes, especially burnout and depression. Data are also reviewed regarding specific occupations at
3 increased risk for deleterious stress-related disorders. Of note here are the limitations of the Job Strain Model as currently operationalized, e.g. for single occupation studies of ambulatory blood pressure responses during work, as well as for detecting the reasons why city mass transit drivers and other professional drivers are at such high risk for cardiovascular disease. Chapter 5 is devoted to the application of the OSI in clinical practice. The OSI provides the necessary information for obtaining a comprehensive occupational history, including not only psychosocial exposures, but also those of a physical and chemical nature, as well as considering other key stressors such as long work hours and shift work, inter alia. From each of four completed OSI’s: for a physician, a truck driver, an administrative clerical worker who works daily with computers, and an automobile assembler, we provide step-bystep instructions on how to prepare a narrative history. These can be viewed as examples of occupational histories that could be incorporated into the general medical history. In Part II of this Chapter, four complete clinical cases are presented, with diagnostic and management issues informed by each of the OSI’s and their corresponding narrative histories from Part I. This chapter was designed for pedagogical purposes1. We therefore pose a set of questions for each case, and provide some possible answers at the end of the Chapter. Perhaps the most important questions to ask oneself in reading Chapter 5 are: (1) what would have been missed if the occupational history were not included? (As is now very often the standard of care, particularly in clinical cardiology, which is the major domain of these clinical cases), and (2) how can the insights gained from the work history inform various possible diagnostic and management scenarios? The development of the OSI framework and the specific questionnaires has been a long-term process. In Chapter 6 we present for the first time, the revised versions of the generic OSI and of the specific OSI’s for professional drivers and physicians, as well as the newly developed OSI for teachers. Together with the questionnaires are explicit instructions for coding and data analysis. Permission to use any of the OSI instruments should be obtained from this author. Our policy is to provide permission free-of-charge for all research endeavors aimed at improving the job conditions and health of working people. We will be happy to answer questions concerning its application, and to discuss how the OSI might be best implemented in a given setting, and look forward to dialogue and feedback.
1
The cases are not based on any individual patient.
4
Chapter 2
The Impact Of Stressful Work On Health How research in this area has proceeded: The crucial role of models It has long been suspected that exposure to stressors of the modern work environment may be related to adverse health outcomes. Karasek and Theorell (1990), in discussing how this area originally developed, note that the question of whether the social organization of work caused serious physical illness would require scientific evidence of such associations, that evidence of subjective perceptions, such as job dissatisfaction, would not be sufficient to generate the political will to redress worker hazards related to psychosocial exposures. They pointed out (ibid) that this evidence would be far more difficult to accumulate, compared to that for physical or chemical work exposures, where the cause of injury was often obviously jobrelated. The critical obstacle was, in fact, the theoretical conceptualization, modeling and measurement of workplace stressors. The Job Strain Model A pioneering breakthrough came in 1979 with the introduction of the Job Strain (DemandControl) Model (Karasek 1979). The model was developed for work environments in which stressors are “chronic, not initially life-threatening and the product of sophisticated human organizational decision making. In decision making the controllability of the stressor is critical, and it becomes more important as increasingly complex and integrated social organizations develop, with ever more complex limitations on individual behavior”. The model has two components: “psychological demands, and a combined measure of task control and skill use, or decision latitude”. Job strain occurs when the human organism is overloaded psychologically and at the same time deprived of control over his or her work environment, a combination which is predicted to give rise to increased risk of stress-related illness (Karasek 2000, p.78). A second hypothesis of the model is that high demands together with high levels of decision-making latitude lead to the “active learning” of new, salutogenic behaviors, e.g. improved coping, and may thereby lead to improved health (Ibid). Later, Johnson and Hall (1988) added social support, as a third dimension of the Job Strain Model. When workers are faced with adverse working conditions, a dynamic process of improving these conditions via "collective control" emerges from joint, supportive efforts. This, and other forms of social support at work, can serve as an important buffer against ill health. On the other hand, workers who are faced with high psychological demands and low decision-latitude and who are also socially isolated at work are in the worst situation. This Demand-Control-Support model has high face validity. The model is readily embraced by working people, who tell us that these general concepts coincide very well with
5 their real life experience. As stated recently by Karasek and Theorell (2000): "The ... model is useful educationally when a worksite is being explored. The model has great face value, and the employees immediately grasp the importance of it in the practical exploration of the psychosocial work environment." (p. 78) Exposure to job strain can be assessed from self-report via questionnaire, with the dimensions operationalized in the form of short, general instruments, most frequently the Job Content Questionnaire (JCQ) or the Psychosocial Job Strain Questionnaire (PSJSQ) (Johnson 1988, Karasek 1998, Theorell 1988, Landsbergis 2000). These are feasibly administered in field and epidemiological studies. Data linkage methods have been developed in the U.S. and in Sweden, so that exposure to Job Strain (as well as “iso-strain” in Sweden) can also be inferred from occupational title alone, i.e. the imputation method (Johnson 1993, Schwartz 2000a). External assessment of job characteristics (e.g. by an expert observer) is yet another method for obtaining exposure data. (For an in-depth discussion of methodological issues, see Landsbergis, Theorell et al. (2000)). The Effort Reward Imbalance Model An alternative, yet complementary way of looking at psychosocial work stressors is embodied in the Effort-Reward Imbalance (ERI) Model (Siegrist 1991, 1996). This model emphasizes lack of reciprocity between efforts spent and rewards received. The latter include monetary rewards, as well as esteem, career opportunities and job security. Efforts can be both extrinsic (job demands and obligations) and intrinsic (over-commitment by the individual to work). Compared with the Job Strain Model, with its emphasis on moment-to-moment control over the work process (namely, decision-making latitude), the ERI Model focuses upon macrolevel, longer-term control as reflected in rewards such as income, recognition and chances for job advancement. As pointed out by Belkic, Schnall, Landsbergis and Baker (2000(a)): “Key dimensions are shared by the Job Strain and ERI Models: both control as well as challenge (demands) are an integral part of each. However, control varies—from micro (task) level in the former, to macro level in the latter. The nature of the challenge varies from model to model, but there is a challenge of some kind in each.” (p. 310) The components of each of these two models are shown in Table 2.1a and b, respectively. A burgeoning body of epidemiological studies has emerged examining these exposures in relation to a number of health outcomes. Evidence concerning exposure to Psychosocial Work Stressors and Health Outcomes Cardiovascular Disease The intimate connections between the social environment and the central nervous system (CNS) and the CNS and the cardiovascular system via the autonomic and neuroendocrine
6 systems, together with clinical observations, have long suggested that work stressors may impact upon cardiovascular morbidity and mortality.
TABLE 2.1A Components of the Job Strain Model ___________________________________________________________________________ Psychological job demands: Working very hard Working very fast Excessive Work Conflicting Demands Not having enough time to get the job done
Decision latitude: Skill Discretion Job requires learning new things Job provides opportunities to develop one’s skills Job requires a high level of skill Job requires creativity Job entails a variety of things to do Job does not involve a lot of repetitive work
Decision Authority Job allows making one’s own decision Job provides a lot of freedom as to how the work gets done Job provides a lot of say on the job Job allows taking part in decisions affecting oneself ____________________________________________________________________________________________ Derived from: Karasek RA, Russell RS, Theorell T. Physiology of stress and regeneration in job-related cardiovascular illness. J Hum Stress 1982; 8: 29-42.
7
TABLE 2.1B Extrinsic Components of the Effort Reward Imbalance Model _______________________________________________________________ Extrinsic Effort Constant time pressure due to heavy workload Many interruptions and disturbances on the job Pressured to work overtime Physically demanding work Job has become increasingly demanding
Reward Esteem Respect from superiors Respect from colleagues Respect and prestige based on efforts and achievements Adequate support in difficult situations Not treated unfairly Monetary Gratification/Security and Career Opportunities Adequate salary/income given efforts and achievements No undesirable change at work Promotion prospects Job security Job reflects education and training Adequate work prospects given efforts and achievements
___________________________________________________________________ Derived from: Siegrist J, Peter J. Measuring effort-reward imbalance: Guidelines. University of Duesseldorf,1999.
8
Job Strain and Cardiovascular Disease The Job Strain Model has been the most widely used means of evaluating the psychosocial work environment as it may impact upon cardiovascular disease (CVD), with some studies incorporating the third dimension of social isolation, as well. Since the introduction of the Model, a large number of empirical investigations have been published concerning the relation between job strain and CVD outcomes, including acute myocardial infarction (MI), other manifestations of ischemic heart disease and CVD-related mortality. Many of these studies report significant positive findings, and job strain is increasingly receiving attention as a potential contributor to CVD risk (Barsky 2001, Hemingway 1999, Kristensen 1998, Schnall 1994). On the other hand, however, there have been several non-confirmatory findings concerning job strain and CVD outcomes published from large-scale studies. These results spurred some questions concerning the strength and consistency of the evidence. Recently, using a pre-defined set of criteria, we systematically examined the empirical studies on job strain and CVD. The criteria2 were developed to assess the methodological issues affecting internal validity of studies on this topic, and, whenever possible, to identify the direction in which the results would most likely be affected. Other major elements of causal inference besides strength and consistency of the association were also reviewed. We thereby sought to provide a more definitive answer to the question: Is job strain a major CVD risk factor (Belkic 2002, 2003)? We found that notwithstanding their high methodological quality, in all but one of the fourteen examined longitudinal studies, biases towards the null were predominant, due most often to use of the imputation method and long follow-up times during which there was no assessment of exposure or even employment status. Viewed in this light, we considered that six investigations, including several of the largest, showing significant positive results, plus another three studies with positive, though not statistically significant findings, provide strong and consistent evidence, particularly among men, that exposure to job strain is associated with an increased risk of future cardiac events and death from cardiovascular disease. The magnitude of this association appears to have been substantially underestimated, since bias towards the null was present in nearly all of these longitudinal studies. We also concluded that the six of nine case-control studies with significant positive results, provide consistent evidence supporting an association between job strain and cardiovascular disease among men, 2
The empirical studies examining the relation between exposure to job strain and CVD outcomes were assessed using 15 criteria that could affect internal validity. Issues most likely creating bias to the null were: use of the imputation method and of a dichotomous variable to define job strain, both leading to non-differential misclassification; assessment of exposure to job strain temporally distant from outcome, e.g. from longitudinal studies with protracted follow-up periods without repeated assessment of exposure status; selection bias in assembly of the sample, if participants exposed to job strain but without CVD preferentially enter the study; selective attrition, if those exposed to job strain or related work stressors selectively stop working during the follow-up period; likely confounding by other factors, if the relationships were in the opposite direction of the tested association; and lack of gender-stratified analysis. Overestimation of association could occur with: information bias, if outcome and exposure were both self-reported; selection bias in assembly of the sample, if participants exposed to job strain and with CVD preferentially enter the study; selective attrition if those not exposed to job strain or related work stressors selectively stopped working during the follow-up period; and likely confounding by other factors, if the relationships were in the direction of association.
9 and some, though not as consistent, support for this association among women. The crosssectional studies provide further evidence of an association between job strain and CVD among men, although biases leading both to over-estimation, as well as to the null, may have been present in some of the studies. Based upon these analyses, the conclusion of the previous focused review on this topic (Schnall 1994), has been corroborated, namely that that there is strong and consistent evidence of an association between exposure to job strain and CVD, across study designs and across a somewhat limited number of examined populations. The data among women are much more sparse, and not quite as consistent, though, as is the case among men, the majority of the studies are likely to have underestimated existing effects. Several other elements of causal inference were also supportive of this hypothesis, particularly the biological plausibility of the association between job strain and risk of CVD (Belkic 2003). Effort Reward Imbalance and Combined Effects upon Cardiovascular Disease A substantial body of cross-sectional and longitudinal investigation, primarily among men, has also shown a significant positive association between Effort-Reward Imbalance and acute MI, as well as CVD-related mortality. The magnitudes of the effect have been found to be similar or even higher than for job strain studies. (For overviews, see Belkic, Landsbergis et al. (2000(b)) and Brisson (2000)). Peter and colleagues (2002) recently examined the combined effects of exposure to Job Strain and Effort Reward Imbalance upon risk of acute MI in the Stockholm Heart Epidemiology case-control study. Among men, exposure to job strain together with high extrinsic effort and low rewards, yielded a considerably higher adjusted effect estimate, Odds Ratio (OR) =2.02 (1.34 – 3.07), compared to being exposed only to job strain or only to ERI (1.42 and 1.30, respectively). This was a gender-specific finding: among women; it was only intrinsic effort (over-commitment) plus job strain, which yielded a combined effect. These authors point out that assessing the joint effects of the two models is much more informative than handling the alternative model as a confounder. Controlling one model for the other, in order to test independent effects did not result in systematically increased effect estimates in their study. Bosma and colleagues (1998a) found that although job control remained a significant independent predictor of self-reported CHD after adjusting for Effort-Reward Imbalance, the effect estimate diminished. They also reported a significant association between job control and effort-reward imbalance: those with low job control reported ERI more often than those with high job control. This association is not surprising, since the control dimension is integral to both models, though, as mentioned, for job strain this is mainly control over task performance, whereas ERI views control at the “macro-level”, over larger issues such as salary, career advancement, etc. The extrinsic effort and psychological demand dimensions have substantial similarity, and show moderate statistical correlation (Peter 2002). Thus, as we have noted above, while the two models have clear conceptual and operational differences, they also overlap. Most importantly, the “combination of information derived from the two models [captures] a broader range of stressful experience at work, and thus, result[s] in an improved risk estimate” (Ibid, p.294).
10
Hypertension There is also strong empirical evidence linking job strain to hypertension. A sizable amount of data has accumulated, indicating that exposure to job strain is cross-sectionally and longitudinally associated with significant elevations in ambulatory blood pressure of clinically important magnitude, greatest at work, but also at home and during sleep among heterogeneous working populations (LaFlamme 1998, Melamed 1998, Schnall 1992, Schnall 1998). (For detailed numerical analyses and discussion of methodological issues regarding job strain in relation to ambulatory blood pressure, see Landsbergis et al. (1994), and for reviews of the empirical data, see Belkic et al. (2000b) and Brisson (2000)). One of the largest investigations, the Work Site Blood Pressure Study, which addresses this issue, includes longitudinal follow-up. Three-year follow-up of 195 men showed that those chronically to job strain had a +11.1/+9.1 mmHg adjusted difference in work systolic/diastolic ambulatory blood pressure, compared to the men unexposed at baseline and at follow-up (Schnall 1998). Increased ambulatory blood pressure, particularly during work, is closely linked to left ventricular hypertrophy (Devereux 1983, Liu 1999, Verdecchia 1994). Furthermore, exposure to job strain has been directly associated with increased left ventricular mass (Schnall 1990). It is therefore plausible that sustained exposure to job strain leads to sustained elevation in blood pressure, which in turn causes structural changes in the left ventricle. Considering the strong, independent relation between increased left ventricular mass and cardiac events, this pathway may account for a substantial part of the reported association between job strain and CVD-related morbidity and mortality (Schwartz 2000b).
Studies assessing the relationship between job strain and ambulatory blood pressure among single occupations have often yielded null results. These include studies among nurses (Goldstein 1999), firefighters (Steptoe 1995), and teachers3 (Steptoe 1999). An investigation by Theorell, Ahlberg-Hulten, Jodko and colleagues (1993) which included female registered and licensed vocational nurses as well as hospital aides, revealed a significant association between systolic and diastolic blood pressure at work and exposure to job strain, based upon selfreport. Two cross sectional studies (Peter 1997, 1998) show a significant positive association between exposure to Effort-Reward Imbalance and elevated blood pressure. At 6.5 year follow-up, an increased risk for the co-manifestation of elevated blood pressure and high LDL-cholesterol was found among blue-collar workers exposed to ERI, after adjusting for age, body mass index, smoking and exercise (Siegrist 1991, 1996). Repetitive Motion Injury/Musculo-Skeletal Disorders
3
Work ambulatory blood pressure did not differ significantly between teachers in the high strain versus low strain groups. However, the difference between day and evening blood pressure was significantly greater among those in the low strain group. The authors state that “failure of subjects with high job strain to show reduced blood pressure in the evening may be a manifestation of chronic allostatic load” (p. 193):
11 There has been substantial investigation of the relationship between adverse psychosocial work conditions and musculo-skeletal disorders. Toomingas and colleagues (1997) examined men and women in various occupations, and found that low social support at work, high psychological demands and high job strain were associated with soft tissue tenderness on physical examination in the central body regions. Analyses were stratified by age, gender and physical load at work. The authors conclude that their results are confirmatory of findings from earlier studies regarding associations with symptoms from the neck and back regions. Moreover, they note that studies not separating clinical signs and body regions may have attenuated risk estimates. Leslie and coworkers (1998) showed that lean production and changes in the layout of the shop floor lead to increased risk of injury, in particular repetitive strain injury. Exposure to job strain or its major dimensions has been associated with risk for musculoskeletal symptoms among nurses and nurses’ aides (Ahlberg-Hulten 1995, Josephson 1997), sales persons (Skov 1996), municipal workers (Myers 1999) and among public transit operators (Krause 1997a, 1998). In most of these studies, this association remained significant after adjusting for physical factors. In her review of workers who use video-display units (VDU), Punnett (1997) concluded that based both on self-reported symptoms, as well as objective findings, “For disorders of the hand and wrist, we found evidence that the use of the VDU or the keyboard was a direct causative agent, mediated primarily through repetitive finger motion and sustained muscle loading across the forearm and wrist” (p.1). Working four or more hours per day at a VDU carries an OR of about 2. High work demand and postural stress from poor workstation design were associated with upper extremity disorders (Ibid).
Adverse Mental Health Outcomes Exposure to job strain has been found to significantly predict depressive symptoms in a longitudinal study of over 10,000 electrical company employees (Niedhammer 1998). Crosssectional relationships between job strain and negative emotions have also been reported (Bourbonnais 1996, Williams 1997), though null findings have been seen, as well (Landsbergis 1992). Recent preliminary results from the Czech Republic, Russia and Poland reveal that exposure to Effort-Reward Imbalance, as well as to Job Strain, is associated with depressive symptoms in cross-sectional analyses (Pikhart 2002). There are also some data indicating a relationship between the main components of job strain and various psychological markers. In a 1.5 year follow up of 11,121 working men, psychological work load was associated with 1.4 times higher risk of a new visit for psychiatric treatment (Uehata 1993). In an Israeli study among female blue-collar workers, shortcycle repetitive work was significantly related to psychological distress (anxiety-irritability, depression and somatic complaints) (Melamed 1995). Low job control was significantly associated with negative affectivity among male civil servants studied by Bosma et al. (1998b). However, risk estimates of self-reported heart diseases due to low job control were not substantially changed in models with and without adjustment for negative affectivity in the paper of Bosma et al (1997), suggesting that this was not a mediator of the job control-CHD association.
12 Job strain has also been associated with burnout or vital exhaustion in cross-sectional studies among nurses (Amick 1998, Bourbonnais 1998) and teachers (Cropley 1999). Nurses in Germany exposed to Effort-Reward Imbalance were found to have high levels on two of the three core dimensions of burnout (Bakker 2000). Occupations With Evidence of Risk for Adverse Stress-Related Health Outcomes The burden of disease is unequally distributed across various occupations. Identification of occupational groups at increased risk for adverse, stress-related health outcomes can be helpful in generating etiological hypotheses (Tuchsen 2000). Here, we briefly summarize salient findings concerning occupational groups for whom there is some evidence of increased risk for one or more of these health outcomes. A number of methodological issues arise with this type of analysis. Of particular importance for cardiovascular disease is the very strong selection effect in hiring and periodic examination for many of these high stress occupations, leading to a bias towards the null. These issues are discussed by Tuchsen (Ibid). Professional Drivers Professional drivers, especially urban transit operators, are at exceptionally high risk for hypertension and ischemic heart disease (IHD). The data accumulated over three decades in various countries, despite “super-healthy worker” selection against these diseases at screening and follow-up. Compiling the focused reviews of Winkleby, Ragland, & Fisher et al. (1988), Belkic, Savic & Theorell et al. (1994), and van Amelsvoort (1995), Belkic, Emdad and Theorell (1998) reported that thirty-four of forty empirical studies on this topic showed a significant positive association. Such a consistent and large body of evidence concerning hypertension and ischemic heart disease cannot be found for any other occupational group. Of particular note is that the acute cardiac events often occur prematurely, such that professional drivers are over-represented among series of young myocardial infarction patients (Riecansky 1988, Villarem 1982). In the paper of Villarem, Thieleux, & LaBlanche et al (1982) of thirty-eight consecutive patients who had a first myocardial infarction before the age of thirty, 20% were long-route truck drivers. Riecansky, Milichercik & Kasper et al (1988) reported that 40% of their series of patients younger than forty with acute myocardial infarction were professional drivers. City mass transit drivers are especially prone to develop ischemic heart disease, as well as hypertension (Alfredsson 1993, Backman 1983, Gustavsson 1996, Michaels 1991, Morris 1966, Netterstrom 1988, Ragland 1987, Ragland 1997, Rosengren 1991). Standard cardiac risk factor prevalence has been found to be high among professional drivers, however, risk factor status does not consistently distinguish professional drivers from other groups at lower risk (Belkic 1994). Rosengren, Anderson & Wilhelmsen (1991) found that the increased risk of coronary heart disease among middle-aged bus and tram drivers compared to referents in Gothenburg was independent of standard risk factor status. After a mean of 11.8 years of prospective study, these authors reported an odds-ratio of 3.3 (95% Confidence Interval = 2.0 - 5.5) for coronary heart disease in 103 middle-aged male mass transit drivers in Gothenburg compared to 6596 men from other occupational groups. With accounting for the major standard cardiac risk factors (age, serum cholesterol, blood pressure, smoking, body mass index, diabetes, positive parental
13 history of CHD and physical activity) as well as socio-demographic factors, the risk decreased only slightly (OR=3.0, 95% CI=1.8-5.2). More recently, Bigert and colleagues (2002) showed that that increased risk for myocardial infarction among bus and taxi drivers, was only slightly diminished by adjusting for selfreported job strain. The authors conclude, “bus and taxi driving in urban areas is a highly stressful occupation and all aspects of the psychosocial stressors in this environment may not be reflected by the demand/control model”. This point is discussed in more detail in Chapters 3 and 4. Empirical studies have also revealed an increased risk among professional drivers for stroke musculoskeletal disorders (Backman 1983, Hedberg 1988, Krause 1997b), peptic ulcer disease (Netterstrom 1990), and psychological distress (Orris 1997). (Tuchsen 1997),
Health Care Professionals Physicians and nurses, particularly early in their careers, have been reported to have a high prevalence of adverse mental health findings, including depression and burnout (Baldwin 1995, Hisashige 1991, Schweitzer 1994). Olkinuora and colleagues (1992) in their study of a representative, random sample of 2671 Finnish physicians, found the highest burnout scores among nonspecialists, and also among those often dealing with chronic, incurable or dying patients. A total of 22% and 26% of male and female physicians, respectively, in their sample had either contemplated or attempted suicide. Gunnarsdottir and colleagues (1995) compared 2159 female Icelandic nurses who had worked more than or less than 20 years in their profession, with the general population, and found an excess risk of suicide in those with the shorter employment time. High rates of suicide among nurses have also been reported in the U.S. (Katz 1983) and among female physician consultants in the U.K. (Carpenter 1997). Tan (1991) states, “among all professional groups, nursing has one of the highest rates of suicide” (p. 227). According to Heim (1992) suicide rates are high among physicians, especially females, although he states that these findings are “surprisingly little observed and reflected by the medical community when compared with increasing preventive activities in other job situations” (p. 207). Low standardized mortality rates due to cardiovascular disease have been reported among physicians (Carpenter 1997). However, Nedic and colleagues (2001) followed two groups with hypertension: 160 physicians and nurses, and referents: 122 hospital employees of other profiles without clinical duties. The doctors and nurses were found to have a RR=3.7 (95% CI=1.6 - 8.6) for developing cardiovascular or cerebrovascular complications (MI, angina pectoris, stroke) at 7 year follow-up. Smoking, obesity and alcoholism were similar in the two groups, while lack of physical activity, positive family history of CVD, and hyperlipidemia were significantly higher among the referent group. High rates of back pain have been found among nurses in several countries (Harber 1985, Stubbs 1983, Videman 1984).
14 Teachers In a study from the U.S., Eaton and colleagues (1990) report that teachers and counselors at the non-college levels4, were one of four occupations with a significant adjusted odds ratio (2.85, 95%CI = 1.2 – 6.8) for major depression using DSM III criteria measured by the National Institute of Mental Health Diagnostic Interview Schedule. This was part of a case-control study among twenty-eight selected occupations examined in a five-site Epidemiologic Catchment Area Program of 11,789 persons who had been full-time employed. A longitudinal investigation of Schonfeld (1992) of 255 first-year female teachers in the New York metropolitan area revealed that CES-D-assessed depression was higher among those working in the most adverse conditions (episodic and ongoing stressors such as threat of personal injury, confrontation, vandalism, overcrowding, unmotivated pupils, lack of disciplinary enforcement against unruly pupils), after controlling for other risk factors. Among 352 nursery teachers and guidance workers in homes for mentally retarded children in Japan, Takeda (1994) found that labor-related problems: dissatisfaction and sense of being overburdened were significant multivariate predictors of depression, using the Zung scale. However, their sample was not found to have a higher prevalence of depression compared to the general population. Burnout among teachers has received increasing attention, as reviewed by van der Berghe (1999). Air Transport Professionals Air Traffic Controllers Cobb and Rose (1973) performed a detailed study in the U.S. of 4,325 air traffic controllers (ATC) and 8,435 second class airmen, with respect to hypertension, peptic ulcer disease and diabetes. Since the licensing regulations are far more stringent for the latter group with respect to hypertension, their findings of significantly greater prevalence and incidence of diagnosed hypertension among the controllers is somewhat difficult to interpret. More informative is the significant association between average traffic density and both hypertension and peptic ulcer disease among the ATC. Comparison of 80 male ATC working in Milan with an age-matched group of 240 male workers in a variety of occupations, revealed no significant differences in ambulatory blood pressure and heart rate on 24-hour periods that included working shifts, while casual clinic systolic but not diastolic blood pressures were significantly higher among the ATC (Sega 1998). The authors (Ibid) note the rigorous selection procedure, which applies to this occupational group. Potter (1987) cites air traffic controllers as one of the high-risk groups for burnout. Feelings of burnout among ATC have been associated with a number of work-related factors: having experienced a “near miss” in the past 3 years, number of years of work as an ATC, working at the highest level of air traffic (“Level 5”), poor work organization, and poor social support 4
Classified by most-recent full-time job.
15 from supervisors and co-workers (Landsbergis 1998). Air traffic controllers were among the occupational groups found to be at increased risk for acute myocardial infarction in the population-based study of Hammar, Alfredsson and colleagues (1992). Airline Pilots Airline piloting is clearly a highly stressful occupation. However, as indicated above, there are extremely rigorous selection and licensing standards, particularly with respect to cardiovascular disease, hypertension, psychiatric disorders and substance abuse (Dark 1987, Joy 1992). These limit the value of comparisons with other occupational groups. Little data is available on the relationship between stressful working conditions and health outcomes. A field study of pilots by Jorna (1993), however, indicates that a total loss of heart rate variability (HRV) occurs during the time of landing, and that when pilots learn to handle a new type of aircraft, there was a prolonged duration of attenuated HRV during the approach period, prior to touch down. Sea Pilots Harrington (1972) reported that over a 12-year period, in comparison with age-matched employed referents, 393 English sea pilots age 35 to 49 have 3.5 times more cardiac deaths (p1.
16
Chapter 3
How Can Insights From Cognitive Ergonomics And Brain Research Inform Our Assessment Of The Work Environment? Complementary to constructs such as the Job Strain Model and Effort-Reward Imbalance, that are based heavily upon sociological theory, are approaches derived from cognitive ergonomics and brain research. The domain of cognitive ergonomics encompasses questions about how the human being processes information, makes decisions and, on that basis, carries out actions (Singleton 1997). Spectacular advances have been made in our understanding of how the human being-via the Central Nervous System (CNS)--handles information, transforming it into productive output of various kinds. The question which we are now addressing is: How can that knowledge be harnessed to inform us in our quest to organize work so that it becomes in better harmony with human needs and capacities? This approach helps us understand what are the affinities and the aversions of the human being at work, and then to describe, in more objective and quantitative terms, the burden of work processes upon the central nervous system (CNS). Thus, e.g., when speaking of psychologically demanding work, we can go far beyond queries about “working hard” and “working fast”, to analyze tasks in terms of allocation of mental resources. In other words, cognitive ergonomics and brain research help us address in a concrete way, the very question of “how fast is too fast” or “how hard is too hard”. Approaches to quantifying the mental burden of occupational endeavor using objective means can help circumvent some of the difficulties inherent in self-report methods (Cacioppo 1990, Greiner 1998, Kristensen 1996, Sackett 1979, Schnall 1994). Ideally, this information would complement the worker's own perceptions of his or her occupational tasks and in that way help guide participatory intervention strategies. Unfortunately, however, the tremendous knowledge gained from this research has generally not been harnessed to inform and enrich Psychosocial Work Assessment. These methods, when applied in the context of the real-life work environment, have mainly been used to improve performance, although there are exciting developments for humanization of work, some of which we will discuss here. How we handle Information: A Neurophysiologic View Figure 3.1 is a schematic representation of how the organism handles information--mapped along the time axis, as developed by Ivanitsky (1980). We see that the incoming signal is evaluated first in terms of its physical attributes, and then with respect to its meaning for the individual. On that basis, a decision is made, which may result in some action.
17
Perception
Psychophysics of perception
Sensitivity Index
Criterion Index
Evoked Potential
Brain Informational Processing
Msec.
FIGURE 3.1
Sensory Stage
Stage of Synthesis
Decision Making
Analysis of the physical characteristics of the signal
Synthesis of the physical and biological characteristics of the stimulus Sensation
Decision-Making
0--------50--------100--------150--------200--------250--------300
Schematic Diagram of the Stages of Information Processing
From: Ivanitsky AM. Evoked potentials and mental processes. In Lechner H, Aranibar A. Electroencephaologr Clin Neurophysiol Amsterdam: Excerpta Medica, 1980, pp. 727-732. With permission.
18 As we view this process along the time axis, it can be seen that at least 300 milliseconds (0.3 seconds) are needed to arrive at this stage of decision-making, at which time the P300 wave of the evoked potential may be seen. The P300 is a positively oriented averaged electrocortical wave appearing 300 - 500 milliseconds after stimulus presentation, and is most commonly elicited when the subject's attention is focused upon an infrequently-occurring signal, especially if this signal has some motivational or emotional significance. The P300 will be produced by task-relevant stimuli that occur relatively unexpectedly, and require either some kind of motor response and/or cognitive decision (Ritter 1968). For more information about Event-Related Potentials, such as the P300, see (Chiappa 1989, Coles 1990, Cooper 1980). The study of “Event-Related Potentials” (ERP) provides us with insight into the higher nervous processing resources required by a given task. These can be a gauge of "mental chronometry" (McCarthy 1981) by assessing brain activity over time as it processes information, makes decisions and lays the basis for task execution. Using the concepts from cognitive ergonomics, Event-Related Potentials can help quantify mental burden using the time dimension.
Time FIGURE 3.2 Schematic illustration that the P300 wave has a longer latency (occurs later), as shown by the curve with dotted line, when more time is needed to evaluate a given signal.
The latency of the P300 ERP component is related to the time required to evaluate and correctly categorize a signal. For example, in the visual modality, when detection is made more complex or when contrast is diminished, the latency of the P300 becomes prolonged (Walton 1987). This is shown above in Figure 3.2.
19
Reaction Time (msec)
600
400
200
0
1
3
5
7
9
Number of Alternatives
FIGURE 3.3
The relation between number of alternatives and reaction time: The “Merkel” curve
Derived from: Merkel J. Die zeitlichen Verhaltrisse der Willenstatigkeit. Philosophie Studion 1885; 2: 73-127.
As the incoming information exacts a greater demand upon the brain’s processing resources, not only the electrocortical activity (e.g. P300) but also the reaction time (RT) takes longer. Here, as shown above in Figure 3.3, from the classical study by Merkel (1885), as the number of alternatives increases, there is a marked rise in the mean RT: from 200 milliseconds (msec) with 1 signal versus 600 msec. when the choice is among 9 alternatives.
20
↑ cognitive or emotional significance of signal
Time
↑ complexity of primary task
Time FIGURE 3.4 resources
Increased P300 amplitude (dotted line) with increased demands upon mental
The amplitude of ERP components, notably the P300 wave, also reflects allocation of mental resources to a given task. As the complexity of a task increases, not only is the latency prolonged, but also the amplitude of the P300 wave heightens. The amplitude of the P300 also increases as the cognitive or emotional significance of the signal increases, as shown above in Figure 3.4. However, when a person performs two tasks, the P300 amplitude to a subsidiary task diminishes as the primary task becomes more difficult, indicating withdrawal of processing
21 resources from a lower priority task as the primary one consumes progressively more of the subject's mental energy (Sirevaag 1984). On the other hand, as the load upon memory increases, P300 amplitude falls. This may similarly reflect competing demands upon mental resources, so that less are available for the specific task at hand; there may also be more uncertainty or equivocation with increasing memory load (Kok 1997). The effect of exacerbating stressors can also be seen with respect to the amplitude of the P300. Laboratory exposure to noise and to sleep deprivation elicited an attenuation of P300 amplitude (Gunter 1987, Polich 1995). Among professional drivers, an inverse relation was found between number of work hours behind the wheel and the P300 amplitude to a visual oddball reaction time task. This finding was considered to be related to fatigue (Belkic 1996), which is known to attenuate P300 amplitude, as well as prolonging its latency (Polich 1995).
TIME PRESSURE
↓ P300 Latency (CR sub-component)
No ∆ P300 Latency (Stimulus Assessment Sub-component)
+ Shortened Choice Reaction Time
PERFORMANCE ERRORS
FIGURE 3.5 A Neurophysiologic View of Time Pressure Derived from data of: Hohnsbein J, Falkenstein M, Hoormann J. Effects of attention and time-pressure on P300 subcomponents and implications for mental workload research Biol Psychol 1995; 40: 73-81.
22
Subcomponents of the ERP waves can be even more illustrative of how these exacerbating stressors deleteriously affect mental processes. Hohnsbein and colleagues (1995) found that when a person was placed under time pressure to perform two-choice reactions, the choice reaction subcomponent of the P300 shortened, even though the stimulus assessment time remained unchanged. The consequence was a greater number of performance errors. These findings are graphically represented in Figure 3.5. In other words, the subject was making a forced decision to react to stimuli that had not been properly evaluated, because the brain did not have sufficient time to do so! These neurophysiologic findings provide insight into the brain mechanisms that may mediate the compromise of safe performance, including the observed increase in accident rate, associated with high levels of time pressure (Gardell 1983, Green 1991, Greiner 1998).
Levels of Information Transmission: High Demands and Underload In line with the neurophysiologic scheme and data that we have just shown, the main phases involved in any kind of labor can be described in terms of Levels of Information Transmission, as formulated by Welford (1960): Sensory input, central decision-making and effecter output or task performance. TABLE 3.6
Levels of Information Transmission
___________________________________________________________________________
Sensory Input
Central decision-making: Information Processing
Effecter output: Task Performance
___________________________________________________________________________ Derived from: Welford AT. The measurement of sensory-motor performance: Survey and reappraisal of twelve years’ progress. Ergonomics 1960; 3: 189-230.
23 In the Occupational Stress Index (to be presented in full subsequently) we have attempted to quantify some of the elements of high demand, broken down by level of information transmission. On the input level besides the number of signals to be processed, we need to consider their modality, complexity, dynamics, sources, inter alia.
TABLE 3.7
High Demand on the Input Level
_________________________________________________________________________
Several sources of information Heterogeneous signals Heavy burden on the visual system High frequency of incoming signals Three sensory modalities Communication essential ___________________________________________________________________________ From the OSI
We know e.g. that the visual modality places the greatest demands upon attentional resources, as do heterogeneous signals, especially if from various sources. Then, we go to the level of central decision-making and consider how many elements are involved (complicated) and their interrelation (complexity), as well as decision-making involving the work of others, and how quickly the decision needs to be made. Next, we consider the nature, heterogeneity and time exigencies involved in actual task performance. We point out here that when people themselves report working hard and working fast, they mainly look at the task performance level, rather than the first two levels (sensory input and decision-making), that are often “invisible”. Finally, we have added a fourth level which goes beyond the levels of information transmission, to considering working conditions on a more general level, that contribute to demand: long hours, lack of rest breaks, irregular schedule or night work, working at a piece rate, lack of vacations and holding down more than one job.
24
TABLE 3.8
High demand among Professional Drivers vs. Subway Attendants according to the Job Strain Model versus the OSI Professional Drivers (Mean +/- sd)
Level of Significance Subway Attendants (Mean +/- sd)
Job Strain
(N=34)
(N=23)
High Demand
11.9 +/- 3.8
Non-significant
12.3 +/- 3.4
Total High Demand
16.1 +/- 2.1
p < 0.001
6.4 +/- 3.8
Input High Demand
9.1 +/- 1.2
p < 0.001
3.4 +/- 1.2
p < 0.01
1.3 +/- 1.4
p < 0.001
1.0 +/- 1.1
OSI
Central High Demand 2.0 +/- 0.2 Output High Demand
3.9 +/- 0.5
From: Belkic K, Emdad R, Theorell T, Cizinsky S, Wennberg A, Hagman M, Johansson L, Savic C, Olsson K. Neurocardiac mechanisms of heart disease risk among professional drivers. Stockholm: Swedish Fund for Working Life, 1996.
Table 3.8 provides an example of how a more detailed and operationalized approach to the demand dimension distinguishes professional drivers, whom we consider to be among the most highly strained of occupational groups, but for whom the standard 5 questions for the demand dimension from the Swedish Psychosocial Job Strain Questionnaire (Theorell 1988) actually scored lower than the subway attendant referents. Note especially, that the demands upon professional drivers are nearly three times greater than referents at the level of incoming signals. Notably, there is actually a reciprocal relation between what would be described as working fast (i.e. driving fast), and the actual demands on the input and central level. In other words, when traffic conditions are the most demanding, as in rush hour, the possibilities to proceed quickly with task performance are, in fact, the lowest.
25
TABLE 3.9
High Demand Among City Bus Drivers Versus Truck Drivers-Using the OSI City Bus Drivers (N=130)
Level of Significance
Truck Drivers (N=69)
Input High Demand
10.0 ± 0.9
Central High Demands
2.0 ± 0.0
Output High Demands
4.0 ± 0.1
P