To better assess the biological mechanisms by which job strain may impact on blood pressure, several new tests are being added to the Cornell worksite study. One is for urinary catecholamines and cortisol, and another is an ultrasound procedure to measure the extent of atherosclerosis in the carotid artery. In addition to associations between job strain and standard coronary risk factors (e.g., blood pressure), other hypothesized pathways by which job strain might affect CVD are an increase in the mass of the heart's left ventricle (LVMI), an increase in the speed of coronary atherosclerosis through mechanisms such as coagulation, and the precipitation of myocardial infarction or arrhythmias among vulnerable persons with underlying heart disease.
The Cornell worksite study only assessed current exposure to job strain, which cannot adequately reflect interpersonal differences in cumulative exposure. For example, 22% of the participants changed job strain status over the first three years of the study. Some older participants with high blood pressure may have worked for years in high strain jobs, but were eventually promoted (to a job with more authority and/or skills), and were therefore classified as "not exposed" in this analysis. In theory, a measure of cumulative exposure should provide a more rigorous test of the job strain hypothesis. While the cross-sectional age x strain interaction (the apparent lack of effect among the 30-40 year olds in the study) suggests an induction period of at least a decade, cohort findings, stratified by age group, suggest that blood pressure increases occurred over three years even in the youngest (30-40 year old) group (Schnall et al., 1992b). Therefore, complete work histories are now being collected from study participants. From these, we will be able to estimate whether cumulative exposure (e.g., number of years in high strain work) is a better predictor of CVD risk than only current exposure. In addition, we will be able to estimate the induction period for hypertension due to job strain, whether exposures early in life or later in life are more consequential, and whether the negative effect of job strain on blood pressure can be reversed by a certain number of years of working in a non-strain job.
More objective measures of job strain
Self-report bias is a potential problem in many job strain studies since exposure has often been assessed through questionnaires completed by study participants. Self-reports may be inaccurate descriptions of job characteristics, or may be biased by personality characteristics. However, while self-reports of work characteristics may be affected by perceptions, this does not make them primarily subjective. Nor are they measures of "perceived" stress, since someone can work in a job with high demands and low decision latitude, and not report feelings of stress.
Concerns about the subjectivity of perceptions in self-report data are substantially addressed by 11 job strain studies in which national averages of job characteristics for a particular job title are assigned to individuals having that job title, thereby effectively excluding the subjective component of reported job characteristics data (Schwartz, Pieper & Karasek, 1988). This procedure, designed to establish a "consensus" of workplace observers, results in more objective measures of stressors since "they are independent of a person's cognitive and emotional processing" (Frese & Zapf, 1988, p. 378). The imputation technique may also be effectively applied within a workplace, by averaging people's ratings of job demands, job decision latitude, and other job characteristics within a jobtitle. A person's self-reported job characteristic score can be removed from the average score for his/her jobtitle, and the resulting adjusted jobtitle average can then be assigned to the person. This would ensure the goal of "independence" when predicting outcomes for this individual.
However, this strategy can result in substantial misclassification due to within-occupation heterogeneity of job characteristics, and a loss of statistical power for detecting job strain effects. Despite these limitations, seven of 11 CVD and CVD risk factor studies using the imputation method have yielded significant positive findings (Schnall, Landsbergis & Baker, 1994) thus providing strong support for the Karasek model. However, use of this method is becoming increasingly problematic in the U.S., since U.S. national averages are based on the 1969, 1972 and 1977 Quality of Employment Surveys conducted by the U.S. Department of Labor. National occupational averages may have changed substantially over the intervening period. While we believe that the imputation methodology has wide applicability in this and other areas of research, it requires that large nationally representative surveys like the Quality of Employment Surveys be conducted periodically.
In order to assess more objective cumulative exposure, Johnson & Stewart (1993) constructed a job exposure matrix in Sweden (with jobtitle-specific as well as age-group and gender-specific scores) in a study population which includes information about previous jobtitles. Job characteristics are then imputed to study participants based on length of employment in a particular jobtitle. A similar matrix could be created in the U.S., but, again, it should be based on updated national job characteristics surveys. Ideally, a research study would utilize the imputed data as well as self-reported data.
Other strategies need to be developed to obtain more objective measures of job strain and of job stress in general. One strategy is the use of outside expert observers. In one of the few studies where expert ratings were available (of education "expected of a particular occupation"), those ratings were highly correlated (r=0.64-0.69) with self-report of "intellectual discretion" (Karasek et al., 1981). (Lower correlations would be expected between expert observers and self-report for "psychological workload demands", which are more subjective, and have less between-occupation variance than measures of decision latitude, Schwartz, Karasek and Pieper, 1988.) Another strategy, being utilized in the Cornell worksite study, is to examine other features of the work environment that may be correlated with job strain, e.g., involuntary overtime, low education required by the job, assembly-line work, electronic monitoring, short cycle time work, etc.
While these techniques for obtaining more objective measures of job strain are valuable for internal validity, they are of limited value for our goal of understanding the social context of stress in a particular workplace, since they are inherently reductionist and exclude the meaning of events to participants from analysis. To accomplish the goals of understanding context, or conducting and evaluating interventions, standardized scales (such as those from Karasek's Job Content Questionnaire) need to be utilized in combination with qualitative data collection techniques and questions specific to a particular job and workplace. Qualitative data collection and analysis methods, e.g., focus groups, interviews, ethnographic observation (Patton, 1987; Neale, Singer & Schwartz, 1987), are among the techniques which overcome the limitations of quantitative instruments and analysis.
Formulations of the job strain model
Job strain has been defined as an interaction between job demands and job decision latitude (Karasek, 1979). However, a variety of mathematical forms of job strain have been utilized in CVD research studies. Only two of four studies found a significant interaction effect using the traditional partialled product interaction term for latitude x demands, after controlling for the main effects of latitude and demands. More frequently, a dichotomous, trichotomous, or four-level exposure variable was constructed (high demands plus low latitude) -- with typically significant positive effects. In addition, in all six studies that used a quotient term (demands divided by latitude) to operationalize job strain, significant associations were found (Schnall, Landsbergis & Baker, 1994). Main effects were often not reported in the studies utilizing dichotomous or quotient terms. Therefore, it remains to be seen whether a strict interaction formulation (with no main effects), or a model including some main effects of job demands and job decision latitude, as well as an interaction term, best describes the effect of job strain on cardiovascular health. Karasek (1989, p. 143) suggests that true interaction effects are frequently difficult to detect due to a lack of statistical power. With respect to the job strain model in particular, he argues that the exact form of an interaction term is not the main issue, since "the primary `interaction' claimed for the model is that two separate sets of outcomes (e.g., risk of illness and activity level) are jointly predicted by two different combinations of psychological demands and decision latitude -- an interaction of significant practical importance."
The possible existence of and location of a threshold of effect for job strain remains to be determined. Data from the Cornell worksite study suggests a significant effect of job strain (using the quotient term) on work and home ambulatory blood pressure at cutpoints beginning at about the upper tertile of the distribution of job strain and increasing in magnitude at narrower cutpoints. The group defined by the lowest tertile of decision latitude and the highest tertile of demands (6.5% of the sample) had work SBP about 9-12 mm Hg higher than those in low demand or high latitude groups. (Patterns for DBP were less clear.) Using national means for decision latitude and demands to define the "high strain" group (8% of the sample) increased effect sizes to 11.5 mm Hg SBP and 4.1 mm Hg DBP. Therefore, in this sample, effects were observed for a range of formulations of job strain. Job strain appeared to have a threshold of effect, and increasing effects at higher levels (or "doses") of job strain (Landsbergis et al., 1994).
Components of the job strain model
Associations between job strain, hypertension and CVD have been reported utilizing varied (and often rather narrow) operationalizations of demands and control. One of the next stages in research will be to assess whether broader operationalizations of these concepts will hinder or improve our ability to predict ill health, as well as whether certain dimensions of demands and control may be more strongly associated with hypertension and CVD. For example, levels of workers' influence, not just on their task, but higher level influence at the group, department or company level, either individually by workers or collectively, also need to be examined (Israel, Schurman & House, 1989; Johnson, 1989; DiMartino, 1992). In the Cornell worksite study, adding a measure of organizational influence to the task-level decision latitude variable produced a greater risk of hypertension due to job strain (OR=3.7, 95% CI 1.6-8.5) (Landsbergis et al., 1994).
The job demands scale in the Job Content Questionnaire (JCQ) (Karasek et al, 1985) primarily contains measures of workload demands, along with one item on role conflict. However, it has been suggested that a broader formulation of job demands may be appropriate. For example, the JCQ and the NIOSH generic job stress instrument (Hurrell & McLaney, 1988), derived from the Michigan job stress model (Caplan et al., 1975), also contain items on cognitive demands. In addition, the Michigan model and the NIOSH instrument contain items on job demands such as role ambiguity, responsibility for people, and threat of violence or injury.
In addition, workplace social support has been added to the job strain model as a third major psychosocial job characteristic in four CVD studies as well as a number of studies of psychological strain outcomes. The combined hazard has been referred to as "iso-strain", or socially isolated high strain work. Three of the four CVD studies provide evidence of interaction -- a "buffering" of the effects of job strain by social support (Johnson & Hall, 1988; Johnson, Hall & Theorell, 1989; Astrand, Hanson & Isacson, 1989). Therefore, it is recommended that future job strain studies examine the potential main and moderating effects of social support.
Potential confounding variables
Related job characteristics. Physical activity is protective of CVD, and overall physical activity at work (self-reported) was controlled for in many of the job strain-CVD studies (Schnall, Landsbergis & Baker, 1994). However, variations in physical position and activity during a work day will affect ambulatory blood pressure readings (Pickering, 1991). Therefore, participants in the Cornell worksite study record their position for each blood pressure reading and are now wearing a small device, known as an "actigraph", along with their ambulatory blood pressure monitor, to measure changes in physical activity throughout the day. Future analyses will include these measures as independent and control variables.
Other workplace characteristics (e.g., job insecurity, chemical and physical work hazards) also need to be assessed and controlled for. Modest elevations of blood pressure have been associated with noise (Talbott et al., 1985), lead (Pirkle et al., 1985) and job insecurity (Kasl & Cobb, 1983). CVD has been associated with exposure to carbon monoxide, solvents, other chemicals, and shiftwork (Kristensen, 1989). Blood lead levels of some participants in the Cornell worksite study will be measured to examine whether lead is possibly a confounder or effect modifier of the job strain-blood pressure relationship.
Non-work sources of stress
The research on job strain and both CVD and psychological distress has made an important contribution to our understanding of a major environmental source of stress. Of course, important sources of stress exist outside the work environment, and these have been carefully examined in major bodies of research, including those on stressful life events (SLEs), and on work/family role conflict. The concept of control as a critical moderator of demands has influenced recent work in those research traditions. For example, controllability of SLEs is considered an important variable in modifying the effects of events (Kasl, 1983, p. 95). Measures of control, demands, support and conflict (and their interaction) in home and family settings have also provided valuable insights (Hall, 1992). Person-environment fit researchers as well have now emphasized control as a significant variable in their model (Caplan, 1983). Such non-work sources need to be incorporated into any complete model of stress and health.
Most job strain researchers recognize that characteristics of the individual play a role in the development of stress symptoms and illness. We believe it is unlikely, however, that such characteristics explain the reported associations between job strain and hypertension or CVD. Type A behavior, for example, is primarily associated with higher status "success-oriented" jobs, making it an unlikely confounder of job strain and CVD. In the Cornell worksite study, Type A behavior (assessed by the Jenkins Activity Survey) was most prevalent in "active" jobs (Landsbergis et al., 1992), and was not associated with hypertension. (In addition, in this study, the Type A scale was controlled for.) In fact, hypertensives did not differ significantly from normotensives on anger, anxiety, hostility, Type A behavior, or other psychological measures in the Cornell worksite study (Schnall et al., 1990). While some research has suggested associations between some personality characteristics and hypertension and CVD (Friedman et al., 1994), in most positive studies, hypertensives have been aware of their blood pressure and therefore these associations could result from diagnosis and labeling rather than representing a cause of the illness (Rostrup & Ekeberg, 1992).
Research on the effects of personality measures also needs to consider the potential influence of job characteristics in shaping personality. For example, Kohn & Schooler (1982) demonstrated that the substantive complexity of work (closely related to decision latitude) predicted increased intellectual flexibility, non-authoritarianism, and intellectually demanding leisure time activities 10 years later. In Sweden, workers whose jobs became more "passive" (low demand-low latitude) over six years reported less participation in political and leisure activities. In contrast, workers in jobs which became more "active" participated more in these activities (Karasek & Theorell, 1990, p. 53). Intervention studies
In the U.S., efforts to reduce occupational stress continue to focus primarily on changing the individual behavior of employees, e.g., relaxation techniques, exercise, diet, cognitive/behavioral skills (Ivancevich et al., 1990). However, the accumulation of evidence linking job stress (including job strain) to CVD risk (and other negative health outcomes) has resulted in a variety of job redesign and public education programs (Karasek & Theorell, 1990; DiMartino, 1993), this joint APA/NIOSH conference, as well as several intervention studies designed to alter the work environment to reduce job stressors (e.g., Israel, Schurman & House, 1989; Cahill & Feldman, 1993). It is important that such programs and studies be promoted and flourish. It is also critical that work environment reform efforts, whether or not part of a formal intervention study, be carefully evaluated.
Interventions to reduce job strain will differ in character from traditional heart disease risk factor interventions (changes in diet, smoking cessation, exercise, weight loss) which are based on individual lifestyle changes. They will be guided by methods developed in social psychology and sociology, primarily participatory action research. This process involves researchers and organization members (usually management and employee representatives) jointly involved in all aspects of the intervention -- including needs assessment, developing targets for change, feedback, planning, intervention and evaluation (Israel, Schurman & House, 1989). The goals of such a process are both evaluation and workplace change. This will necessitate the use of both quantitative and qualitative research methods, multiple levels of intervention (e.g., individual, work group, organization), longitudinal designs, and the measurement of process, impact and outcome (thus requiring multidisciplinary teams skilled in these techniques) (Israel et al., 1992). While changes in "risk factors" and resulting health impacts will be more difficult to isolate than in a clinical intervention trial, such methods provide a key to both changing structural aspects of the workplace that cause stress, and understanding the change process.
The evidence supporting the existence of a relationship between
job strain and CVD is quite strong1. Prior to the Cornell
worksite study and several other recent studies, however, the
evidence for a relationship between job strain and CVD risk factors
was mixed. Results from the Cornell worksite study indicate a
relationship of job strain to:
1) hypertension (defined from casual blood pressure readings taken on two separate occasions)
2) ambulatory blood pressure at work, home, and sleep
3) enlargement of the heart
4) three-year changes in ambulatory blood pressure.
Several important directions for future research in this area were suggested. First, we would like to learn whether job strain is related to any of a variety of structural characteristics of the cardiovascular system, e.g., left ventricular mass, arterial wall thickness, or the presence of atherosclerotic plaques. Second, there is a need to investigate alternative biophysiological mechanisms by which job strain may elevate blood pressure, both acutely and chronically, e.g., activation of the sympathetic nervous system.
Much work also needs to be done on refining the job strain concept and its measurement. At the conceptual level, it remains to be seen whether job strain would predict CVD outcomes better if the two underlying dimensions, job demands and decision latitude, were broadened to include such factors as role conflict and responsibility for people, or job insecurity, supervisory authority, or influence on organizational decision-making. Most existing operationalizations of job strain have used narrower definitions. A major measurement issue in research on psychosocial risk factors for CVD is the potentially subjective quality of the predictors. Most of the job strain research is based on self-report data. Therefore, it is important to determine whether it is mostly the objective characteristics of jobs or an individual's subjective perception and evaluation of them that is most closely associated with changes in blood pressure or the development of CVD. This chapter has suggested several possible approaches that might be used for assessing job characteristics more objectively.
In addition, people change jobs and jobs change their characteristics. Most of the empirical analyses of job strain, blood pressure and CVD have relied on cross-sectional assessments of job strain. Yet clearly most of the hypotheses implicitly assume a putative effect of long-term (chronic) job strain, analogous to that of smoking. This chapter describes attempts by Johnson & Stewart (1993) and ourselves to retrospectively estimate cumulative exposure to job strain.
While refinements in the measurement of job strain and the assessment of additional intermediate outcomes are likely to improve our understanding of job strain's relationship to CVD, the most stringent tests of the job strain hypothesis are likely to come from intervention studies that have not yet been conducted. Successful intervention studies should be able to demonstrate that manipulation of job demands and/or decision latitude can result in clinically meaningful changes in blood pressure, other structural characteristics of the cardiovascular system, or CVD risk. The considerable literature on organizational interventions and change strategies provides a useful guide to efforts to reduce the CVD risk that results from job strain.
In summary, while existing research leads us to be fairly confident that a relationship exists between CVD, hypertension and job strain, much work needs to be done before we fully understand this relationship, and test and refine methods to modify job design and work organization in order to reduce the risk of these diseases.
1) The possible association between job strain and health outcomes other than CVD and hypertension also needs further investigation. Some evidence exists linking job strain with psychological distress outcomes (Sauter, Murphy & Hurrell, 1990) such as exhaustion or depression (Karasek, 1979) or job dissatisfaction (e.g., Karasek, Triantis and Chaudhary, 1982; Landsbergis et al., 1992). In addition, several recent studies suggest that job strain may play a role in the development of adverse pregnancy outcomes such as low birth weight (Brandt & Nielson, 1992; Homer, James & Siegel, 1990) or pregnancy-induced hypertension (Landsbergis, Hatch & Zhang, 1993). A recent review suggests that musculoskeletal disorders may be associated with psychosocial job stressors similar in concept to job strain such as monotonous work, high perceived workload, time pressure, low job control, and lack of social support (Bongers et al., 1993). Another job stressor, electronic performance monitoring may also be associated with the development of cumulative trauma disorders (Smith et al., 1992). Finally, several studies suggest that psychosocial job factors (although not job strain, specifically) may be associated with "sick building syndrome" (National Institute, 1991), and alterations of the immune system (Henningsen et al., 1992).
American Heart Association. (1992). 1993 Heart and Stroke Facts.
Astrand, N. E., Hanson, B. S., & Isacson, S. O. (1989). Job demands, job decision latitude, job support, and social network factors as predictors of mortality in a Swedish pulp and paper company. British Journal of Industrial Medicine, 46, 334-40.
Bongers, P. M., deWinter, R. R., Kompier, M. A. J., & Hildebrandt, V. H. (1993). Psychosocial factors at work and musculoskeletal disease: A review of the literature. Scandinavian Journal of Work Environment and Health, 19, 297-312.
Brandt, L. P. A., & Nielson, C. V. (1992). Job stress and adverse outcome of pregnancy: A causal link or recall bias? American Journal of Epidemiology, 135, 302-311. Caplan, R. D. (1983). Person-environment fit: Past, present and future. In C. L. Cooper (Ed.), Stress Research (pp. 35-78). London: Wiley.
Caplan, R. D., Cobb, S., French, J. R. P., Jr., Van Harrison, R., & Pinneau, S. R., Jr. (1975). Job demands and worker health (Publication No. 75-168). Cincinnati, OH: National Institute for Occupational Safety and Health.
Cahill, J., & Feldman, L. H. (1993). Computers in child welfare: Planning for a more serviceable work environment. Child Welfare, 72, 3-12.
DiMartino, V. (1992). Conditions of Work Digest: Anti-Stress Programs. Geneva: International Labor Office.
Frese, M., & Zapf, D. (1988). Methodological issues in the study ofwork stress: Objective vs. subjective measurement of work stress and the question of longitudinal studies. In C. L. Cooper & R. Payne (Eds.) Causes, coping and consequences of stress at work (pp. 375-411). New York, NY: Wiley.
Friedman, R., Schnall, P. L., Pieper, C. F., Gerin, W., Landsbergis, P. A., & Pickering, T. G. (1994). Psychological variables in hypertension: The "hypertensive personality" revisited (under review).
Hall, E. M. (1992). Double exposure: The combined impact of the home and work environments on psychosomatic strain in Swedish men and women. International Journal of Health Services, 22, 239-260.
Henningsen, G. M., Hurrell, J. J., Baker, F., Douglas, C., MacKenzie, B.A., Robertson, S. K., & Phipps, F. C. (1992). Measurement of salivary immunoglobulin A as an immunologic biomarker of job stress. Scandinavian Journal of Work Environment and Health, 18(suppl 2), 133-136.
Homer, C. J., James, S. A., & Siegel, E. (1990). Work-related psychosocial stress and risk of preterm, low birthweight delivery. American Journal of Public Health, 80, 173-177.
Hurrell, J. J., & McLaney, M. A. (1988). Exposure to job stress -- A new psychometric instrument. Scandinavian Journal of Work Environment and Health, 14(suppl 1), 27-28.
Israel, B. A., Schurman, S. J., & House, J. S. (1989). Action research on occupational stress: Involving workers as researchers. International Journal of Health Services, 19, 135-155.
Israel, B. A., Schurman, S. J., Hugentobler, M. K., & House, J. (1992). A participatory action research approach to reducing occupational stress: Phases of implementation and evaluation. In V. DiMartino (Ed.) Conditions of Work Digest: Anti-Stress Programs (pp. 152-163). Geneva: International Labor Office.
Ivancevich, J. M., Matteson, M. T., Freedman, S. M., & Phillips, J. S. (1990). Worksite stress management interventions. American Psychologist, 45, 252-261.
Johnson, J. V. (1989). Collective control: Strategies for survival in the workplace. International Journal of Health Services, 19(3), 469-480. Johnson, J. V., & Hall, E. M. (1988). Job strain, work place social support, and cardiovascular disease: A cross-sectional study of a random sample of the Swedish working population. American Journal of Public Health, 78, 1336-1342.
Johnson, J. V., Hall, E. M., & Theorell, T. (1989). Combined effects of job strain and social isolation on cardiovascular disease morbidity and mortality in a random sample of the Swedish male working population. Scandinavian Journal of Work, Environment and Health, 15, 271-279.
Johnson, J. V., & Stewart, W. (1993). Measuring work organization exposure over the life course with a job-exposure matrix. Scandinavian Journal of Work, Environment and Health, 19, 21-28.
Karasek, R. A. (1979). Job demands, job decision latitude, and mental strain: Implications for job redesign. Administrative Science Quarterly, 24, 285-308.
Karasek, R. A. (1989). Control in the workplace and its health-related aspects. In S. L. Sauter, J. J. Hurrell, C. L. Cooper (Eds.) Job Control and Work Health (pp. 129-159). New York: Wiley.
Karasek, R. A., Baker, D., Marxer, F., Ahlbom, A., & Theorell, T. (1981). Job decision latitude, job demands, and cardiovascular disease: A prospective study of Swedish men. American Journal of Public Health, 71, 694-705.
Karasek, R. A., Theorell, T. (1990). Healthy work. New York, NY: Basic Books.
Karasek, R. A., Triantis, K. P., & Chaudry, S. S. (1982). Coworker and supervisor support as moderators of associations between task characteristics and mental strain. Journal of Occupational Behavior, 3, 181-200.
Kasl, S. (1983). Pursuing the link between stressful life experiences and disease: A time for reappraisal. In C. L. Cooper (Ed.) Stress Research (pp. 79-102). London: Wiley.
Kasl, S. V., & Cobb, S. (1983). Variability of stress effects among men experiencing job loss. In L. Goldberger & S. Breznitz (Eds.) Handbook of stress (pp. 45-65). New York, NY: The Free Press.
Kristensen, T. S. (1989). Cardiovascular diseases and the work environment. A critical review of the epidemiologic literature on chemical factors. Scandinavian Journal of Work Environment and Health, 15, 245-264.
Landsbergis, P. A., Schnall, P. L., Dietz, D., Friedman, R., & Pickering, T.G. (1992). The patterning of psychological attributes and distress by job strain and social support in a sample of working men. Journal of Behavioral Medicine, 15, 379-405.
Landsbergis, P. A., Hatch, M., & Zhang, H. (1993). Psychosocial workplace risk factors for pregnancy-induced hypertension (abstract). American Journal of Epidemiology, 138, 606.
Landsbergis, P. A., Schnall, P. L., Schwartz, J. E., Warren, K., & Pickering, T. G. (1994). Association between ambulatory blood pressure and alternative formulations of job strain. Scandinavian Journal of Work, Environment and Health (in press).
Light, K. C., Turner, J. R., & Hinderliter, A. L. (1992). Job strain and ambulatory work blood pressure in healthy young men and women. Hypertension, 20, 214-218.
National Institute for Occupational Safety and Health. (1991). Indoor Air Quality and Work Environment Study (HETA 88-364-2102). Cincinnati, OH: Author.
Neale, M. S., Singer, J. A., & Schwartz, G. E. (1987). A systems assessment of occupational stress: Evaluating a hotel during contract negotiations. In: A. W. Riley & S. J. Zaccaro (Eds.) Occupational stress and organizational effectiveness (167-203). New York, NY: Praeger.
Patton, M. Q. (1987). How to use qualitative methods in evaluation.Newbury Park, CA: Sage. Pickering, T. G. (1991). Ambulatory monitoring and blood presure variability. London: Science Press.
Pirkle, J. L., Schwartz, J., Landis, J. R., & Harlan, W. R. (1985). The relationship between blood lead levels and blood pressure and its cardiovascular risk implications. American Journal of Epidemiology, 121, 246-258.
Rostrup, M., & Ekeberg, O. (1992). Awareness of high blood pressure influences on psychological and sympathetic responses. Journal of Psychosomatic Research, 36, 117-123.
Sauter, S. L., Murphy, L. R., & Hurrell, J. J. (1990). Prevention of work-related psychological disorders. American Psychologist, 45, 1146-1158.
Schnall, P. L., Pieper, C., Schwartz, J. E., Karasek, R. A., Schlussel, Y., Devereux, R. B., Ganau, A., Alderman, M., Warren, K., & Pickering, T. G. (1990). The relationship between 'job strain', workplace diastolic blood pressure, and left ventricular mass index. Journal of the American Medical Association, 263(14), 1929-1935. (With letter to the editor: Schnall, P. L., Devereux, R. B., Pickering, T. G., & Schwartz, J. E. (1992). Journal of the American Medical Association, 267(9), 1209.)
Schnall, P. L., Schwartz, J. E., Landsbergis, P. A., Warren, K., & Pickering, T. (1992a). The relationship between job strain, alcohol and ambulatory blood pressure. Hypertension, 19, 488-494.
Schnall, P. L., Landsbergis, P. A., Schwartz, J. E., Warren, K., & Pickering, T. G. (1992b, September). The relationship between job strain, ambulatory blood pressure and hypertension. Paper presented at the Ninth International Symposium on Epidemiology in Occupational Health, Cincinnati, OH.
Schnall, P. L., Landsbergis, P. A., Pieper, C. F., Schwartz, J., Dietz, D., Gerin, W., Schlussel, Y., Warren, K., & Pickering, T. G. (1992c). The impact of anticipation of job loss on psychological distress and worksite blood pressure. American Journal of Industrial Medicine, 21, 417-432.
Schnall, P. L., Landsbergis, P. A., & Baker, D. (1994). Job strain and cardiovascular disease. Annual Review of Public Health, 15, 381-411.
Schwartz, J. E, Pieper, C. F., & Karasek, R. A. (1988). A procedure for linking psychosocial job characteristics data to health surveys. American Journal of Public Health, 78, 904-909.
Schwartz, J. E., Schnall, P. L., Warren, K., Pickering, T. G. (1992, November). The role of demographic factors in the job strain model: Independent or non-independent risk factors. Paper presented at the APA/NIOSH Conference on Occupational Stress, Washington, DC.
Smith, M. J., Carayon, P., Sanders, K. J., Lim, S-Y., & LeGrande, D. (1992). Employee stress and health complaints in jobs with and without electronic performance monitoring. Applied Ergonomics, 23, 17-27.
Talbott, E., Helkamp, J., Matthews, K., Kuller, L., Cottington, E., & Redmond, G. (1985). Occupational noise exposure, noise-induced hearing loss, and the epidemiology of high blood pressure. American Journal of Epidemiology, 121, 501-514.
Theorell, T. G., Ahlberg-Hulten, G., Jodko, M., Sigala, F., Soderholm, M., & de la Torre, B. (1993). Influence of job strain and emotion on blood pressure in female hospital personnel during work hours. Scandinavian Journal of Work Environment and Health, 19, 313-318.
Van Egeren, L. F. (1992). The relationship between job strain and blood pressure at work, at home, and during sleep. Psychosomatic Medicine, 54, 337-343.
|Risk Factor||Odds Ratio (95% C.I.)||p|
|Age 51-60 (vs. 30-40)||11.4 (4.3-30.4)||<.0001|
|Age 41-50 (vs.30-40)||6.4 (2.6-15.7)||<.0001|
|Job Strain||2.9 (1.3-6.6)||.011|
|Body Mass Index||1.3 (1.1-1.5)||.0009|
|Race (other than Caucasian)||1.7 (0.6-5.3)||NS|
|Physical Exertion||1.1 (0.7-1.8)||NS|
|Type A Behavior||1.1 (0.5-2.2)||NS|
|Urine Sodium||1.0 (1.0-1.0)||NS|
Risk estimates reflect the independent effect of each variable, controlling for all other variables in the model. The full model was computed by unconditional logistic regression. Worksite was also controlled for, coded as seven dummy variables.