Post on 25-Apr-2018
Workshop Public Health EnglandManchester, 23 March 2015
Work stress and health: Do we need more evidence to act?
Johannes Siegrist, Ph.D.
Senior Professor, Faculty of MedicineUniversity of Duesseldorf. Germany
Work …
§ provides a source of regular income and related opportunities
§ provides a source of personal growth and training of capabilities/competencies
§ provides social identity, social status and related rewards
§ enables access to social networks beyond primary groups
§ Impacts on personal health and well being by exposure to material and psychosocial stressors
Importance of work for health
§ Increase of work pressure and competition, including ‘high performance work organization’ (impact of economic globalization)
§ High demand for flexibility, mobility, and adaption of workers to new tasks/technologies
§ Fragmentation of occupational careers, de-standardized or atypical work, and growing job instability/insecurity
§ Increase of service and IT professions/occupations with high psychomental/emotional workload
§ Segmentation of labor market; social inequalities in quality of work and employment
Significant changes in the nature of work and labor market
Increased pressure of rationalisation(mainly due to wage competition)
Downsizing, Merging, Outsourcing
Work Job Low wage / intensification insecurity salary
Effects of economic globalisation: Labour market consequences in developed countries
Increase in work intensity 2004-2010: European Social Survey, 19 EU countries
Source: Gallie D (Ed.) (2013) ESS Topline Results Series 3, European Social Survey
Job insecurity 2004-2010European Social Survey, 19 EU countries
Source: Gallie D (Ed.) (2013) ESS Topline Results Series 3, European Social Survey
Job loss and risk of acute myocardial infarction (HRS Study, USA; n = 1.351)
Source: Dupre, ME et al. 2012: Arch Intern Med, 172(22): 1731-1737, (p. 1734).
High work pressure (e.g. overtime work) and job instability (e.g. downsizing) are unhealthy!
Examples of recent evidence from UK and Finland:
Overtime work (>11 hrs/day):risk of severe depression: HR 2.4risk of incident CHD: HR 1.7(Virtanen M et al. PLoS One 2012, Eur Heart J 2010)
‚Surviving‘ severe downsizing:risk of all-cause mortality: HR 1.4risk of CHD mortality: HR 2.0(Vahtera J et al. BMJ 2004)
Psychosocial stress at work
Stress occurs if a person is exposed to a threatening demand(stressor) that taxes or exceeds her/his capacity of successful response à risk of loss of control and reward
Dimensions of stress reactions: • Cognitive evaluation of threat• Negative emotions (anxiety, anger)• Activation of stress axes in organism (SAM, HPA)• Behavioural reaction (fight or flight) (restricted option!)
Critical for health:• Chronic stressors requiring active coping à allostatic load;
à risk of stress-related disorders (depression, CHD)• Adverse work is a major chronic stressor in adult life
Progress in defining and measuring a health-adverse psychosocial work environment
Development and test of theoretical models of stressful work
Given the complexities and varieties of modern work it is not easy to identify common
features that account for elevated health risks among people exposed to these conditions.
Yet, identifying such common features and demonstrating their contribution to the
explanation of workers’ health is the task of a stress-theoretical model:
– A theoretical model of work stress is delineated at a level of generalization or
abstraction that allows for the identification of health-adverse work in a wide
range of different occupations and contexts.
– A theoretical model of work stress is defined by a selective reduction of
complexity to some core analytical dimensions rooted in human stress theory (e.g.
loss of control, frustration of reward, violation of equity)
Theoretical models of work stress and evidence of adverse health effects
• Demand-control model (R. Karasek, 1979; R. Karasek & T. Theorell, 1990)
• Effort-reward imbalance model(J. Siegrist, 1996; J. Siegrist et al., 2004)
• Organizational injustice model(J. Greenberg et al.,1982; M. Elovainio et al., 2002)
Ø Focus on job task profile: high demand/low control
Ø Focus on work contract: high
effort/low reward
Ø Focus on unfair procedures and
interactions
Chronic psychosocial stress at work: Complementary stress-theoretical models
Active
Passive
Low-strain
High-strain
Psychological Demands
Deci
sion
latit
ude
(con
trol
)low high
low
high
The demand-control model(R. Karasek 1979, R. Karasek, T. Theorell 1990)
Source: Karasek: Healthy work, New York: Basic Books, 1990, p. 32.
• Procedural justicePerceptions of consistent, accurate, unbiased and ethical rules of procedures
• Relational justicePerceptions of polite, fair interactions from supervisors
• Distributive justicePerceptions of appropriate distribution of job tasks and gains among employees
So far, mainly procedural and relational justice were measured with relevance to health and performance.
The Organizational Justice Model(J. Greenberg et al. 1982; M. Elovainio et al. 2002)
effort
reward
demands / obligations
- labor income- career mobility / job security
- esteem, respect
motivation(‘overcommitment‘)
motivation(‘overcommitment‘)
Extrinsic components
Intrinsic component
The model of effort-reward imbalance (J. Siegrist 1996)
Source: Based on Siegrist, J (1996): J Occup Health Psychol, 1: 27-41.
§ DependencyThe working person has no alternative choice in the labour market: accepting contractual unfairness is preferred to job loss.
§ Strategic choiceThe working person accepts imbalance in order to improve future career development (anticipatory investment).
§ OvercommitmentThe working person exhibits a motivational patternof excessive work-related commitment where investments often exceed gains. Overcommitment is either due topersonality or due to pressure at work.
Why do people continue to work in ‚high cost – low gain‘ conditions?
§ Both models are measured by a standardized self-assessed questionnaire which can be applied to a variety of different occupational groups:
- Job Content Questionnaire (JCQ) (R. A. Karasek)http://www.jcqcenter.org/
- Effort-Reward Imbalance Questionnaire (ERI) (J. Siegrist)www.uniklinik-duesseldorf.de/Med-Soziologie
§ Both questionnaires fulfill criteria of psychometric quality (factorial structure of scales, reliability, discriminant and predictive validity etc.).
§ Both questionnaires are available in a number of languages and have been used in comparative international studies.
Progress in measuring models of stressful work: Demand-Control and Effort-Reward Imbalance
Confirmatory factor analysis: Replication of the theoretical model (effort-reward imbalance)
.65-.33work stress
over-commitment
jobinsecurity
esteem
reward
effort
salary / promotion
ERI1
ERI2
ERI3
ERI4
ERI6
ERI7
ERI8
ERI9
ERI12
ERI13
ERI11
ERI14
ERI16
ERI15
ERI10
ERI17
OC1
OC2
OC3
OC4
OC5
OC6
.75(1)
.95
.82(1)
.49
.49
.64
.67(1)
.66(1).51
.58
.59
.55
.65
.54(1)
.80
.62
.80(1).76
.51(1)
.19
.74
.64
.86
.69
.75
.75
N=666 employed men and women
χ2/dfGFIAGFI CFIRMSEA
Source: A. Rödel et al. (2004) Z diff diagn Psychol 25: 227-238
2,99.91.89.90.06
Mean level of work stress in 17 European countries(SHARE, ELSA, n = 14 254, aged 50-64)
Source: T. Lunau et al. (2013): Unpublished results
.75 1 1.25MeanERI
Hungary
Portugal
Czechia
Poland
Italy
Estonia
Slovenia
England
Spain
France
Germany
Austria
Belgium
Denmark
Netherlands
Sweden
Switzerland
3.5 4 4.5 5MeanLowControl
Poland
Hungary
Italy
Czechia
Spain
Austria
Estonia
England
France
Germany
Belgium
Slovenia
Portugal
Switzerland
Netherlands
Sweden
Denmark
PsychosocialWorkingConditions
Social gradient of work stress
0
5
10
15
20
25
30
35
40
Effort-Reward Imbalance Low control
Perc
ent h
igh
stre
ssed
Very low LowMediumHighVery high
Source: Wahrendorf M et al. (2013) European Sociological Review 29: 792-802
The social gradient of work stress in the European workforce (age 50-64): SHARE-study
Methodological approaches: epidemiological and experimental
Epidemiological research:è prospective observational cohort study (gold standard)è cross-sectional and case-control-study (weaker
evidence)è intervention study (limited options)
Experimental research:è laboratory experiments (limited ecological validity)è ambulatory monitoring at work (limited control)
What is the scientific evidence of a direct association of work stress with disease?
• Prospective Study (exposition à incident disease; statistical control of confounders)
• Strength of association
• Dose – response relationship
• Evidence on biological pathways from exposure to disease
• Consistency of results across contexts and study designs
• Risk reduction through exposure reduction (experimental evidence, e.g. RCT)
Cave:
• Specificity of association of exposure with disease (often not applicable due to multiple health effects of exposure)
Bradford Hill criteria of a causal association in epidemiological research
• Depression:
• ~ 30 studies (Europe, USA, Canada, Japan):People exposed to stress at work: mean increase of relative risk: 80% = OR 1.8 (95% CI 1.1-3.1)
• Coronary heart disease:
• ~ 20 studies (Europe, USA): People exposed to stress at work: mean increase of relative . risk: 40% = OR: 1.4 (95% CI 1.2-1.6)
• Additional evidence of elevated health risks:
Metabolic syndrome / type II diabetesAlcohol dependenceMusculoskeletal disorders
Scientific evidence from prospective cohort studies: Demand-control and effort-reward-imbalance models
Source: Steptoe A, Kivimäki M 2012. Nat Rev Cardiol.9 ; Stansfeld SA ,Candy B 2006 Scand J WEH 32: 443
Source: Based on Bosma, H, et al. (1998), Amer J Publ Health, 88: 68-74
0,5
1
1,5
2
2,5
3
No workstress
Intermediatejob control
Low jobcontrol
0,5
1
1,5
2
2,5
3
No workstress
High effort orlow reward
High effortand lowreward
adjusted for age, sex, length of follow-up+ alternative work stress model+ grade, coronary risk factors, negative affect * p < .05
* *
Work stress (effort reward imbalance/job control) and CHD incidence, men and women: Whitehall II-Study
Decreases risk Increases risk
4/9
Source: Kivimaki, M, et al. Scand J Work Environ Health (2006): 32: 431-442, (p. 436).
Consistency: Meta-analysis of cohort studies on relative risks of coronary heart disease due to ’job strain’
Biological pathways: Control at work and blood pressure
Mean ambulatory blood pressure (low control vs.
high control).
N = 227 men and women (47-59 years); Whitehall
Cohort Study
Low control
High control
Low control
High control
Systolic BP
Diastolic BP
Source: Based on Steptoe, A, et al. (2004), Journal of Hypertension, 22(5): 915-920.
Cardiovascular monitoring over 3 days in healthy male computer employees and work stress
CRP change#
(μg/ml) as function of effort-reward imbalance
# adjusted for age, BMI, baseline levels
Effort-Reward Imbalancelow medium high
p < .050.12
0.10
0.08
0.06
0.04
0.02
0.00
Inflammatory response (CRP) during experimentally induced mental stress according to level of effort-
reward imbalance (N=92)
0
1
2
3
4
5
6
Effort-Reward Imbalance
lowmiddlehigh
Adjusted for age, and sex; Additionally adjusted for hypertension, diabetes mellitus, smoking, BMI, CHD family history, educational level, and marital status; *p<0.05; **p<0.01; ***p<0.001
Generalization: Psychosocial stress at work in Chinese male coronary patients vs. healthy controls (N=388)
Source: Based on Xu, W. et al (2009) J Occup Health 51: 107-113
Depression
• Serious public health problem worldwide
• Estimated life time prevalence: 13-16 %
• Severity due to high co-morbidity (esp. cardiovascular diseases) and risk of suicide
• Manifestation in early adult life, compromised work ability (sickness absence, disability pension)
• Massive direct and indirect costs
• Genetic, early life and other personal determinants, but also role of work stress
Consistency of associations of work stress (effort-reward imbalance) with depressive symptoms in cohort studies
Autor (Jahr) OR 95% KI
Stansfeld (1999) Männer: 2.6Frauen: 1.7
1.8 – 3.61.0 – 2.9
Godin (2005) Männer: 2.8Frauen: 4.6
1.3 – 5.72.3 – 9.0
Kivimäki (2007a) 1.5 1.2 – 1.8
Kivimäki (2007b) 1.6 0.9 – 2.7
Dragano (2011) 1.5 1.3 – 1.8
Rugulies (2013) 2.2 1.1 – 4.2
Wang (2012) 2.3 1.1 – 4.7
Juvaini (2014) 1.9 1.5 – 2.4
Additional consistent results based on linear regression analyses
Source: Juvani A et al. (2014): Scand J Work Environ Health, 40: 266-277.
Dose-response relation: Cumulative hazard curves of disability pension due to depression by quartiles of work
stress (ERI) (n =51.874)
Biological pathways: Work stress (ERI) and natural killer cells in 347 Japanese employees
Source: Nakata A et al (2011), Biol Psychol 88:270-279, (p. 277).
Psychosocial stress at work and depressive symptoms: 13.128 employed men and women 50-64 yrs. from 17 countries in three
continents (SHARE, ELSA, HRS, JSTAR)
0
0,5
1
1,5
2
2,5
USA (N=1560) Europa (N=10342) Japan (N=1226)
ERILow control
Source: J. Siegrist et al (2012) Globalization and Health 8:27.
* **
* *
Reduced fatigue and depression is associated with labour market exit (GAZEL-study, France)
Source: Westerlund H et al (2010) BMJ 341:c6149.
Variable
DemandControlSocial supportRewardEffort-reward imbal.Work-rel. burnout
Means at t2 adj. for t0
experimental - control hospital p
11.970.023.731.21.0
43.2
12.668.723.030.21.1
48.3
.008
.051
.011
.003
.001
.003
Source: R. Bourbonnais et al. (2011), Occup Environ Med, 68: 479-486.
Organizational intervention in a Canadian hospital vs. control hospital*
*36 month-follow-up, two Canadian hospitals, N=248 (intervention) vs. 240 (control hospital) (ANCOVA, adj. for baseline values)
Summary
• Robust scientific evidence that unhealthy work (high demand/low control; high effort/low reward) is associated with increased risk of stress-related disorders.
• Overall, every fifth working person is exposed to unhealthy work, with higher levels among socially disadvantaged, less qualified groups of employees.
• Additional negative effects of stressful work due to sickness absence, reduced productivity and early exit from labour market > Important direct and indirect costs due to non-activity, lack of preventive efforts.
§ Personal level: Stress prevention programs§ Interpersonal level: Leadership training;
communication skills; § Structural level: Organizational/personnel
development (based on work stress models)§ Job enrichment/ enlargement (autonomy, control, responsibility)§ Skill utilization / active learning§ Participation / team work and social support § Culture of recognition§ Fair wages/ gain-sharing§ Continued qualification/ promotion prospects
What can be done to reduce unhealthy work and promote workers‘ health?
Personal level: Effects of worksite stress prevention programs: Meta-analysis
Source: D Montano et al. (2014) Scand J Work Environ Health, doi: 10.5271/sjweh.3412 .
Frequency of worksite health promotion RCTs according to occupational class
Frequencies of the occupational class of samples and outcomes of RCTs: EGP classes I-III: managerial, professional and clerical occupations, and classes VI-VII: manual skilled and semi-
skilled manual occupations
Source: D Montano et al. (2014) Scand J Work Environ Health, doi: 10.5271/sjweh.3412 .
Structural – level: Organizational and personnel development
• Provide evidence of a business case• Provide shared commitment from management
and employees• Provide available expertise (e.g. occupational
safety & health) and equipment • Monitor working conditions and employees‘ health• Develop and implement programmes, invest in
improvements, consult models of best practice• Ensure continuity, evaluate outcomes, build
networks
Initiatives at national and international level
• Apply legal frameworks (national, international)• Enforce fair work standards in trade contracts• Promote voluntary agreements between employers and
trade unions• Support social movements (e.g. EU: Corporate Social
Responsibility)• Maintain and develop distinct national labor and social
policiesØ Welfare regimes securing basic social and economic protection,
access to health care, further education, etc.
Mean score of stressful work and extent of implementation of national labour market integration policy (SHARE-Study)
Quelle: Wahrendorf M, Siegrist J. (2014) BMC Public Health 14: 849
Social gradient of stressful work according to extent of implementation ofdistinct labour market policies (SHARE)
Lunau T. et al. (2015) PLoS One (in press)
Odds ratios of depressive symptoms by work stress: mitigation of effect by distinct labour policies?
Source: Lunau T. et al. (2013), BMC Public Health, doi:10.1186/1471-2458-13-1086
Based on SHARE, HRS,
ELSA; n = 5650, m/w
aged 50-64. 13 countries
While additional research is needed priority should be put on implementing available knowledge in practice and on evaluating its outcomes, both at the level of organizations and of national labor and social policies!
Conclusion
Thank you!