The Whitehall Study

Contact: Professor Michael Marmot
Department of Epidemiology and Public Health
University College London
1-19 Torrington Place
London WC1E 6BT
England
Tel: 44 171 391 1717
Fax: 44 171 813 0280
Email: Michael@public-health.ucl.ac.uk

The Whitehall study examined mortality rates over 10 years among male British Civil Servants aged 20-64. The study was an attempt to avoid some of the problems created by the use of general social class groupings, e.g., the heterogeneity of occupations within a single class leaves room for differing interpretations. The Whitehall study concentrates on one “industry” in which there is little heterogeneity within occupational grades and clear social divisions between grades (Marmot, Kogevinas and Elston, 1987).

An inverse association between grade (level) of employment and mortality from CHD and a range of other causes was observed (78). Men in the lowest grade (others = messengers, doorkeepers, etc.) had a three-fold higher mortality rate than men in the highest grade (administrators) (Marmot, Shipley and Rose, 1984).

Grade is also associated with other specific causes of death, whether or not the causes were related to smoking (Marmot, Kogevinas and Elston, 1987). While low status was associated with obesity, smoking, less leisure time physical activity, more baseline illness, higher blood pressure, and shorter height (78), controlling for all of these risk factors accounted for no more than 40% of the grade difference in CHD mortality (Marmot, Shipley and Rose, 1984; Marmot, Kogevinas and Elston, 1987). After controlling for standard risk factors, the lowest grade still had a relative risk of 2.1 for CHD mortality compared to the highest grade (Marmot, 1994).

One possible explanation of the remaining grade differences in CHD mortality is grade differences in job control and job support (Marmot, Kogevinas and Elston, 1987). In addition, blood pressure at work was associated with “job stress”, including “lack of skill utilization”, “tension”, and “lack of clarity” in tasks. The rise in blood pressure from the lowest to the highest job stress score was much larger among low grade men than among upper grade men. Blood pressure at home, on the other hand, was not related to job stress level (78).

Thus, a second longitudinal study of British Civil Servants (Whitehall II) was initiated to investigate occupational and other social influences on health and disease (Marmot, 1994). The final sample was 6900 men and 3414 women aged 35-55 in the London offices of 20 civil service departments (Marmot et al., 1991). Employment grade was strongly associated with work control and varied work (measures of decision latitude) as well as fast pace (a measure of job demands) (Marmot et al., 1991; Marmot, 1994). Lack of control on the job is related to long spells of absence (> 6 days) (Marmot, 1994).

In addition, there was no decrease in the difference in prevalence of ischemia depending upon employment category over the 20 years separating Whitehall I and Whitehall II (Marmot et al., 1991). Plasma cholesterol concentrations did not differ by job category, and the small inverse association between job status and blood pressure in men was reduced from that seen in the Whitehall I study. There was a significant inverse relation between BMI and job status, but, especially in men, the differences were small. The risk factor that differed most between employment categories was smoking. Moderate or vigorous exercise was less common among subjects in lower status jobs (Marmot et al., 1991).

The Wolf Project

Selected Abstracts

Risk indicators for myocardial infarction and psycho-social working conditions. Peter Westerholm , Lars Alfredson, Evy Fellenius, Monica Soderholm, Tores Theorell, Ulf de Faire, et al.

This project aims at examining the association between workplace conditions – primarily psychological factors and social conditions – and risk indicators for cardiac infarction, i.e. biochemical variables (blood lipids, fibrinogen etc.) and increased blood pressure (Acronym: WOLF: Work, lipids, fibrinogen). One of the aims of the study is also to examine whether already known risk factors for cardiac infarction such as smoking, food habits, physical exercise etc. have any impact which can be modified by the psychological and social workplace factors. The study is carried out with self administered questionnaires and measurement of blood pressure and analyses of biochemical variables. The target population: 6000 employees in ages 30 to 60 recruited from the Stockholm labour market. The project is carried out in collaboration between the National Institute for Working Life, the Karolinska Institute (Institute of Environmental Medicine), Institute of Psychosocial Factors and Health, Department of Social Medicine Kronan, Sundbyberg and the Medical and Occupational Health Departments of Karolinska Hospital, Stockholm. The field organisation of the project consists of more than 20 occupational health service units.

Financial support: Swedish Council for Work Life Research.

Thanks to Yrkesmedicin for this information:
Department of Occupational Health, Karolinska Hospital
SE-171 76 Stockholm, Sweden
phone +46 8 517 730 56

Job strain and major risk factors for coronary heart disease. Baseline results from the WOLF study. Lars Alfredson, Niklas Hammar, Ulf de Faire, Johan Hallqvist, Tores Theorell, Peter Westerholm

In this study we analyzed the relationship of job strain (high pshychological job demands and low decision latitude) to hypertension, serum lipids and plasma figrinogen. The study population consisted of employed persons 15-64 years of age in Stockholm, Sweden, and the data collection was carried out during 1992-1995. A total of 5696 subjects participated in a medical examinatrion and completed a questionnaire (76% of all invited). No strong associations were seen between job strain and hypertension or plasma fibrinogen, but men reporting high psychological demands had higher fibrinogen levels than other men. Somilar tendencies were seen in women. Men reporting job strain also had more often an adverse LDL/HDL balance than other men, and this association was considerably stronger in subjects aged 15-44 than in older subjects. Our results do not support the hypothesis that low job strain has an adverse impact on hypertension, serum total cholesterol and plasma fibrinogen levels. However, an atherogenic cholesterol profile with low HDL levels and high LDL/HDL ratios for subjects reporting job strain was suggestede, in particular among men.

The SHEEP (Stockholm HEart EPidemiology) Project

SHEEP is a major epidemiological study on cardiovascular disease and it’s biomedical risk factors. It is a case-referent study of the first myocardial infarction in men and women (1500 men and 700 women in the ages 45-70) and the same number of referents. It has been established, along with WOLF (WOrk, Lipids and Fibrinogen) as a collaborative effort with several various occupational health care teams, and regional research departments and institutes, hospital departments of medicine. Both projects are part of an EU collaboration including Tores Theorell, Michael Marmot, Johannes Siegrist, and others.

Anders Ahlbom, Johan Hallqvist, Christina Reuterwall, Ulf de Faire, Finn Diderichsen, Christer Hogstedt, Tores Theorell, Lars Alfredson, Per Gustavsson, Nils Plato, Eva Vingard, Piroska Ostlin et al.

The incidence of myocardial infarction (MI) has been shown to vary across different occupational, social, and ethnical groups, between geographic regions, and over time; those variations, however, can only partly be explained by between-group differences in the risk factors hypertension, hyperlipidimi, and smoking. A large number of other risk factors – some of them work related – have also been suggested. The present case-referent study will test several hypotheses concerning risk factors for MI, such as occupational exposure to chemical and/or psychosocial factors, social and environmental factors, dietary and smoking habits, and biological/medical factors. Special attention is paid to potential interaction between the various factors. The study will compromise approximately 2100 cases (1400 men and 700 women) of acute MI (first episode) in the Stockholm county area and the referents are chosen from the general population. Exposure information is collected through questionnaires and biomedical examination.

Several research and clinical units are cooperating in the project – Institute of Environmental Medicine and Department of Social Medicine at Karolinska Institute, the National Institute for Working Life and the National Institute of Psychosocial Factors and Health, Stockholm County units of Environmental Health, Social Medicine, and Occupational Health, Karolinska Hospital and the Departments of Medicine at each of the ten emergency hospitals in the Stockholm County area. Exposure information was collected 1992-1994 and the first results will be reported in 1996.

Financial support by Stockholm County Council, The Social Research Council, and The Medical Research Council.

Thanks to Yrkesmedicin for this information:
Department of Occupational Health, Karolinska Hospital
SE-171 76 Stockholm, Sweden
phone +46 8 517 730 56

The Coeur Project

Summary

In earlier studies the French population has shown a lower incidence of cardiac infarction as compared to the Swedish population.

As a result of the collaboration between Renault and Volvo, scientific research concerning cardiac risk factors has been conducted, especially as related to work situations (with preventive health care as the final objective).

Abstract

In 1993 1000 men 45-50 years of age (in both companies) were divided by professional level into 4 blue collar and 4 white collar categories, and evaluated by a questionnaire and laboratory studies.

During 1996/1997 an in depth hospital study of 90 high risk and 90 low risk individuals from each country will be performed, involving new ultrasonographic methods and special laboratory tests such as insulin, as well as frozen sera for possible future use.

In 1998 all 2000 individuals will be followed up with health examinations; this procedure is planned to be redone after a period of 10 years.

Findings

No large differences have been shown between the two countries with respect to relevance of traditional risk factors (the traditional risk factors also include the Framhingham Risk Index).

The psychosocial questions are currently under analysis. Thus far we have seen that mean Type A-score (Bortner-scale) is higher among French employees and increases with higher worker categories in both countries. There does not seem to be big differences between the white-collar categories in the two countries. Analyses concerning Karasek scores in relation to traditional cardiac risk factors is presently being done with the help of Professor Kristina Orth-Gomer.

We are currently analysing the impact of life events where we will study separate, specific, work related life events.

In addition, analysis is ongoing concerning alcohol consumption, as well as concerning a broad dietary questionnaire involving the hogh and low risk groups mentioned above.

Conclusions

As the project has numerous dimensions, and many analyses are ongoing, we are currently (January 1997) not ready to give any definite conclusions. Several of our sub-studies have been submitted for publication, and are expected to be published soon.

PRIM (Project on Research and Intervention in Monotonous Work)

PRIM is a multidisciplinary study undertaken by the National Institute of Occupational Health; the Departments of Occupational Medicine at Glostrup Hospital, Herning Hospital, Arhus Hospital; Centre for Alternative Social Analysis (CASA); Department of Working Environment, Technical University of Denmark (DWE/TU); and the Danish Institute of Physical Education.

The main objectives of the project are:

To study the musculoskeletal and psychosocial effects of the physical and psychosocial risk factors which occur in monotonous and repetitive work, with special emphasis in the joint and interacting effects of these factors.

To develop knowledge about organizational methods which can lead to viable changes in the companies that introduce more variation in the work, and to evaluate the health effects of such interventions.

To study the effects of dynamic and static muscle training on the perception of musculoskeletal pain among workers with monotonous repetitive work.

The project consists of 6 integrated substudies.

Substudies 1 and 5 are mainly concerned with the musculoskeletal and psychosocial effects of monotonous repetitive work. Substudies 2, 3, and 4 carry out and evaluate interventions in the companies, the aim of which are to introduce more variation in the work. Substudy 6 intervenes against the effect of monotonous repetitive work by subjecting the exposed workers to dynamic and static muscle training.

The projects are closely integrated. Firstly, the persons and companies from substudies 2-6 are also included in substudy 1. Thus, the health effects of the interventions will be closely monitored. Secondly, a number of different scientific methods and fields are involved in the data collection. Since data is collected for individuals, it can be analyzed jointly and thus can give an integrated understanding of the studied phenomena.

Job Strain and the Prevalence of Myocardial Infarction in HES and HANES

The odds of those in High Strain Jobs having had a Myocardial Infarction were: HES: 2.48 (p=.012) and HANES: 3.28 (p=.001)
times that of those Not in High Strain Jobs, controlling for age AJPH, 1988

The Health Examination Survey (HES) 1960-62 is based on an examination of a representative sample of 6,672 persons drawn from the civilian non-institutionalized population of the United States, ages 18 to 79. The detailed component of the Health and Nutrition Examination Survey (HANES), conducted between 1971 and 1975, is similar to the HES in that it is also representative of non institutionalized persons in the United States (6,913 subjects ages 25-74 were examined). Sample weights derived for these surveys are used as appropriate. Both the HES and HANES involve clinical examinations and follow very similar data collection procedures. Our analyses are restricted to males who report current occupation (approximately 17 per cent of the males in each sample are missing occupation data) which results in final sample sizes of 2,409 (2,088 Whites) in the HES, and 2,424 (2,136 Whites) in the HANES.

In the HES data, the presence of myocardial infarction was judged by a clinical panel of four doctors. Myocardial infarctions were classified as definite (93 per cent) or suspect (7 per cent) based on a review of the 12-lead electrocardiograms (EKG), chest X-rays, medical history, and blood chemistry during analyses. The presently available HANES data on coronary heart disease (CHD) are based on the diagnosis made by the field examining physicians after reviewing the medical history, the detailed cardiovascular questionnaire, and the complete physical examination. (Chest X-ray and EKG tracings may or may not have been available. An improved classification based on expert EKG evaluation is still in process.) These criteria have probably led to an underestimation of overall myocardial infarction on the HANES. Experience with examining physicians’ initial diagnosis compared to final diagnosis with the HES shows relatively few false positive CHD cases but a larger number of false negatives. To reduce the unreliability due to misclassification when angina is included, we decided to focus on that subset of CHD which is most reliably assessed, myocardial infarctions (ICD codes, 410-414). The number of cases in the HES is 39 (33 Whites) and 30 in the HANES (28 Whites).

French Psychosocial Stress Research Projects

Aim of the study: This prospective study is designed to assess, in France, the prevalence of hypertension, the one-year incidence of hypertension (unknown today) and the relationship between incidence of hypertension and environmental, occupational and social factors.

Subjects and Methods: Every worker in France, whatever his/her occupation, is affiliated with an occupational medical center and has a systematic clinical check-up every year. In order to ensure a good representation of our sample, we therefore selected from a list of occupational physicians a sample of 50 physicians according to their age and to the typology and location in France, of their occupational center. Each physician then selected randomly 800 workers among his/her assigned population. During the usual systematic annual visit, blood pressure was recorded with a validated semi-automatic device (Omron® > 705 CP) at the 5^th, 6^th and 7^th of rest in the sitting position. When the average of these 3 measurements was 140 / 90 mmHg, a second visit was scheduled one month later for a repeat BP measurement under the same conditions. At each visit, a specifically designed questionnaire was used to record height, body weight, drug treatment, socio-economic status, the level of cardiovascular risks factors as reported by the patient, and an estimation of his/her medical drug consumption over the last year. We anticipated a representative sample of 30,000 French workers. During the second year of the study, a case-control study will be conducted among the cohort of newly identified hypertensive patients and an appropriate cohort of control subjects to assess relationship between psychosocial and occupational factors (Karasek questionnaire) and blood pressure modifications

Results : The study started on January 1^st, 1997. On September 1^st, 20188 subjects (mean age 38.9 ± 9.4 years, proportion of males 56.9%) have been enrolled.. Among the untreated patients, 10% had at the first visit, a mean BP ³ 140 / 90 mmHg, and 90% of them could be seen at the one-month visit.

Conclusion: The inclusion rate is respected in this study and the prevalence data will be available for the ISH meeting.

The Worksite Blood Pressure Study (Cornell University Medical Center)

The Cornell University cohort study of ‘job strain’ and ambulatory blood pressure, begun in 1985, has enrolled 372 initially healthy full-time employees from a wide variety of job titles, aged 30-60, most with at least 3 years tenure with their employer. Of the 372 participants, 80 (22%) are women, and 95 (25%) are members of minority racial/ethnic groups, including 67 Black, 21 Hispanic and 4 Asian participants. An additional 100 participants, nurses and nurses aides from a new worksite (almost all are female and 50% members of minority groups), are being recruited in 1995 and their evaluation will be completed by 11/1/95. Every three years, participants wear an ambulatory (portable) blood pressure monitor for 24 hours on a work day. Subjects also receive medical tests and complete a questionnaire. Every 15 minutes during waking hours (and during hourly sleep) the monitor inflates and records blood pressure. During waking hours, the subject is asked to remain as motionless as possible and then to record his/her activity, location, position, and mood in a diary. The diary information (i.e., whether the subjects reported being at work, home or sleep) has been used to calculate average AmBPs for each location category.

The Cornell study was one of the first work stress studies to use an ambulatory blood pressure monitor. The monitor provides a more reliable measure of blood pressure, since there is no “observer bias” and the number of readings is increased. It also has a more valid (accurate) measure of average blood pressure than causal blood pressure measurements, since blood pressure is measured during a person’s normal daily activities. Studies using an ambulatory monitor, including the Cornell study, have generally found positive associations between ‘job strain’ and blood pressure.

Worksite Surveillance

“In 1998, experts in Japan, Europe, and the United States called for a program of surveillance at workplaces and monitoring at national and regional levels in order to identify the extent of work-stress related health problems and to provide baselines against which to evaluate intervention efforts. They recommended that workplaces measure both workplace stressors and health problems known to result from such stressors [11]. In the United States, occupational and environmental medicine clinics can play a key role in such surveillance efforts. In addition to clinical care, such clinics conduct research, and provide patient education, industrial hygiene and ergonomics services, and social work and support groups [12]. Thus, a team approach is recommended in which epidemiologists work together with clinicians, health educators, ergonomists, psychologists, and other health professionals to identify high-risk workplaces and jobs, facilitate the provision of clinical care, and design and implement workplace interventions.

The surveillance team needs to discover whether the current occupation(s) is high risk; whether workers are exposed to any workplace physical, chemical, work schedule or psychosocial risk factors for disease; and whether any such have been increasing over time [13]. Questionnaires, such as the Job Content Questionnaire [14], the Effort-Reward Imbalance Questionnaire [15], and the Occupational Stress Index [16], can help measure job characteristics and job stressors. Workplace screenings should be conducted for biomedical risk factors, such as high blood pressure [17]. Such surveillance can help to identify clusters of work-related hypertension and help target work sites for primary and secondary prevention programs. Another key part of surveillance is taking an occupational history of workers to see how long they have been facing workplace risk factors for disease, and what types of risk factors [18].

Such surveillance would be an important part of the newly developing field of occupational cardiology, which would link cardiologists, health promotion experts, and occupational health specialists. In addition to efforts described above, they would develop return-to-work guidelines for cardiac patients, including workplace modifications, and expand the use of ambulatory (portable) monitors to measure blood pressure or heart rate [19].”

Emotional Labor

A SHORT DEFINITION by Muntaner

“Learning to manage emotion is essential to forming a mature personality, and is part of all working relationships. The term emotional labour describes jobs that require workers to induce or suppress feelings to sustain the outward countenance that produces the proper state of mind in others.14 For example, airline stewards are responsible for managing situations with customers to create a favourable experience for the customer.14 Other human service jobs that require personal involvement with clients require workers to cede considerable control to patients or clients. Not only has the number of jobs requiring emotional control increased markedly in recent years, but Hochschild 14 has also identified the growing extent to which emotion is actually engineered and managed in these jobs”

C Muntaner, J Benach, W C Hadden, D Gimeno and F G Benavides, A glossary for the social epidemiology of work organisation: Part 1, Terms from social psychology, J. Epidemiol. Community Health 2006;60;914-916 doi:10.1136/jech.2004.032631

14 Hochschild AR. The managed heart. Berkeley: University of California Press, 1983.

ADDITIONALLY…

More recently, researchers have been investigating the effects of emotional labor on worker health. First developed as a concept in the mid-1980s by sociologist Arlie Hochschild, who studied airline attendants, emotional labor is used to describe jobs that require its workers to hide or suppress their actual emotions or call for an inauthentic expression of appropriate emotions for the given situation. These take the form of surface acting, or acting that involves faking or displaying an emotion not felt, and deep acting, wherein workers adjust their actual inner feelings to match those required in the moment. “Front line” workers, human service professionals, sales and service providers are good examples of this (restaurant servers, airline stewards, retail workers, etc). 

The Emotional Labour Scale (ELS) is a 15-item self-report questionnaire that measures six facets of emotional display in the workplace, including the frequency, intensity and variety of emotional display, the duration of interaction, and surface and deep acting. 

 Below are a few items from the Brotheridge and Lee EL scale (1998):

“I just pretend to have the emotions I need to display for my job”.

“I make an effort to actually feel the emotions that I need to display to others”. 

Ten years ago, researchers began finding positive associations between the frequency, intensity, and variety of emotions experienced and burnout [57]. However, since then, other researchers have begun to show that emotional dissonance, a difference between the truly experienced emotion and the emotion that is required to be performed (e.g., “pleasantness”) more consistently leads to negative health outcomes such as burnout [31, 32]. Being required to display positive emotion or suppress negative emotion may be a greater predictor for reduced psychological well-being than just the intensity or variety of emotions experienced on the job.

In this research on human-service workers, Brotheridge and Grandey found that the frequency and duration of service interactions, and the intensity and variety of expression during those interactions, was not associated with emotional exhaustion or depersonalization. In fact, the “amount” of emotional labor performed was related to high personal accomplishment. In other words, having an emotionally demanding job, in and of itself, may actually be a positive experience. It was the requirement to hide negative emotions or display positive emotions which was significantly related to emotional exhaustion and depersonalization. 

In a Swedish study of 2,255 employees of an insurance organization, both quantitative work demands and emotional demands were assessed over a 1-year period [59]. This study found that the emotional demands of the job increased the risk of burnout more than quantitative demands. The authors concluded that it is important to include measures of emotion work when investigating the demand-control model’s effects on burnout in human service workers.

Taken, in part, from: Schnall PL, Dobson M, Rosskam E, Editors Unhealthy Work: Causes, Consequences, Cures. Baywood Publishing, 2009.