1. Long term consequences of stress


1.1 IPD – work, large pooled data studies on psychosocial factors, health and life style

Psychosocial working conditions have been implicated in the etiology of many of the most common diseases, but the results have been challenged by employers, physicians, and researchers. One of the reasons is that most studies have been relatively under-powered and used weak designs such as cross-sectional observations based on self-reports only. In order to provide better evidence regarding the role of psychosocial factors, SSC (E+I+S) has participated in a unique international effort to pool individual-participant data (IPD) from many different studies with a total of >100.000 participants, the “IPD-work” consortium. As a first step, measurements have been harmonized across studies. Subsequent analyses in IPD-work has demonstrated a clearly significant, but modest, effect of job strain on the risk of cardiovascular disease. Other studies in the same consortium have shown moderate associations between job strain and Framingham risk score, diabetes life style, both weight gain and weight loss, number of cigarettes smoked among current smokers, high alcohol consumption (as well as abstinence), and risk of reduced physical activity. In contrast, it was also shown that job strain is unlikely to be an important risk factor for major cancer types or inflammatory bowel diseases and that the association with asthma exacerbation is probably due to confounding. The risk of coronary artery disease was highest among those who reported job strain and an unhealthy lifestyle; those with job strain and a healthy lifestyle had half the risk of disease. A healthy lifestyle may thus substantially reduce disease risk among people with job strain. The IPD-work publications provide the best evidence to date to understand the role of psychosocial conditions in the etiology of health problems. Because of the importance of such evidence, the IPD-work papers have triggered an international scientific debate regarding the magnitude and practical significance of this association.

1.2 Psychosocial factors and disability pension

Disability pension is an important outcome that combines the effects of many different diseases and circumstances on work ability. In several collaborative studies (I+W) we have studied associations between different psychosocial factors and future disability pension. Thus, using nationwide survey data for 11,000 women we found that social relations and weak connections to working life involved higher risk of disability pension in young women, also after control for socioeconomic conditions and self-reported ill health. A study of 54,000 people, followed for 12 years suggests that social isolation and low societal participation are predictors of future disability pension, especially regarding mental disability pension among younger individuals. Combinations of frequent self-rated sickness presenteeism and absence are associated with negative values in four measured aspects of self-reported health one year later. Psychosocial conditions could also impact on desired and expected retirement age.

1.3 Psychosocial factors and sleep

We have been instrumental in showing the role of sleep in stress, and how the impact of stress exposures in e.g. working life may to large extents be determined by the ability to gain restitution from stress. Very few prospective studies investigating the importance of sleep to support individual resilience were available at the start of SSC. In several collaborative (S+E & S+W) prospective epidemiolog- ical studies we have therefore investigated the stress/sleep link. We have demonstrated that stress predicts new cases of impaired sleep across periods of several years. Importantly, in the last study we found that reverse causality was more important than the effects of stress on sleep. Thus, disturbed sleep affects (the perception of) stress. With a similar epidemiological approach, presence of sleep disturbances was coupled to a fivefold increase in mortality in younger males. This also links to two studies of the “Whole Sweden” cohort in which we have shown that a diagnosis of insomnia leads to an increased risk of mortality and disability pension. The studies advance the knowledge on the connection between stress, sleep and health in a working life perspective. We are currently starting studies (S+P+E) to disseminate knowledge to decrease sleep problems among the employees in work organizations.

1.4 Hearing and Stress

Work is frequently related to hearing problems and studies based on SLOSH have shown hearing problems to be present in a third of the working population and that tinnitus severity was reduced with the reduction of depressive mood. Another study found
a clear and mostly linear relationship between higher prevalence of hearing problems (tinnitus or hearing loss or both) and different stressors, e.g. occupational, poorer self-rated health, long-term illness, poorer sleep quality, and higher burnout scores. A third study showed that women with high levels of emotional exhaustion became more sensitive to sound after an acute stress task. A further study indicated that exposure to a combination of noise exposure and job strain substantially increased the risk of myocardial infarction. A review concluded that socioeconomic status and long-term stress exposure increase the risk of hearing problems.

1.6 Psychosocial factors and exhaustion syndrome

Maladaptive stress reaction and exhaustion syndrome (“burnout”) have been major factors in sickness absence statistics and is a focus of much attention of society. The predictors of this outcome has been looked at in several studies. Thus, in a sample of 25 000 Swedish twins we showed that burnout is moderately heritable, that stressful and traumatic life events are of importance in the burnout process, that lack of sleep seems to be a precipitating factor of exhaustion syndrome, and that the associations between events and burnout were evident also after adjusting for familial (genetic and shared environment primarily during childhood) factors. We also found that familial factors are of importance in the association between support and burnout for both women and men and between home-work conflict and burnout among women, but not between job demands or work-home conflict and burnout. In other studies of twins we explored the effect of familial factors for associations between health, work situation, sickness absence, disability pension, and mortality.


Planned research

1.1 Assessing the global impact of different stressors.

Studies of associations between single stressors, for instance in working life, and single outcomes may improve our understanding of mechanisms, but provide relatively little help in terms of priorities between targets of intervention since one stressor usually impacts on many outcomes, some of which are linked in complex ways (e.g. lower back pain, neck pain and sickness absence), and may be moderated by contextual factors. The aim of this project is to utilize existing knowledge and existing infrastructure (e.g. prospective cohort studies from several countries) to assess the effects of modifiable stressors and risk factors with global measures of impact, particularly total life expectancy, healthy life expectancy, DALYs and working life expectancy at different ages. This will be done in such a way that differential effects in subgroups can be analysed, for instance between men and women, and between social classes. The project builds upon two existing international collaborations, the IDEAR (Integrated Datasets across Europe for Ageing Research) network partially funded by the ER-AGE-2 Joint Call for Research on Ageing (Forte #2012-1661) and led by E in collaboration with I, focussing on the impact of working life on healthy life expectancy, and the IPD-work consortium mentioned above. In contrast to existing projects, the new project will add an explicit life course perspective based on the realisation that exposures to different stressors are not independent of each other and that stress exposures are neither discrete events nor completely stable over time, and that both sensitive periods and chains of risk may exist. We will therefore explicitly explore the interrelationship between exposures over time and test models of accu- mulation, chains of risk and sensitive periods. Finally, the project also builds on the realisation that the direction of causality between “exposure” and “outcome” is often not straightforward in psychosocial epidemiology. For instance, while unemployment may impact negatively on mental health, poor mental health also increases the risk of (especially sustained) unemployment, potentially creating vicious (or virtuous) spirals. Novel, advanced longitudinal modelling will be used to account for such complex sequences of causation. The competence currently being built in IDEAR will be extended to allow complex patterns of association and causation to be presented in easily understood form, for instance through interactive, web-based visualization tools. This project may in turn be the foundation of a work package in a European Horizon 2020 consortium which has recently been invited to the second application stage.

1.2 Identifying targets of intervention through improved understanding of mechanisms.

Information about associations between different measures of exposure and outcomes in terms of health or healthy life expectancy may help identify targets of intervention. However, changing exposure can be difficult in reality, and may also potentially be ineffective if the causal mechanisms by which the exposures influence the outcome are poorly understood. This project therefore focuses on understanding underlying social, psychological, behavioral, and biological mechanisms through which different exposures impact on health. In order to do this, nested in-depth studies will be conducted within the large epidemiological cohort studies used by the Stress Research Institute in collaboration within SSC (E+S+W+I) as well as with our international partners. SSC funding will be used for coordination of the collaboration and facilitation for partners to use the cohorts updating data bases. The methodologies used will be any of the following, by themselves or in combination with each other, but always utilizing the synergy between in-depth data with longitudinal data available from repeat questionnaires and administrative registers:

  • Qualitative studies to explore and understand how people interpret, experience and act in relation to the studied stressors (E+W)
  • Field studies where more detailed and objective assessments of stressors, behaviors, and physiological responses in relation to (changes in) environmental circumstances, e.g. major life transitions or job changes (E+S)).
  • Experimental studies, where exposures are manipulated under controlled conditions in strategically sampled individuals. Examples of manipulated variables could include time in bed, light and sound exposure during work, and working hours (E+S+P).
  • Quasi-experimental studies where instrumental variables, i.e. variables which are known to influence an exposure or mechanism, but which are likely to be unrelated to the outcome except through the exposure/mechanism under study. For instance, changes or regional differences in retirement schemes will be used as instruments to study the impact of retirement on health (E+I).
  • Medical add-ons in sub-cohorts to assess the potential confounding or mediating effect of intermediate risk factors (e.g. hypertension and dyslipidemia) as well as to study early stages of (pre)clinical disease (e.g. diabetes or atherosclerosis) (E+I+P).

1.3 Work and sleep

The collaborative (S+I) epidemiological work on work and sleep will continue, using the All-Sweden database, using work stress indicators from the ULF data base (>100.000) to study effects on health and use of hypnotics, as well as sleep problems as predictors of use of hypnotics. Also sleep apnea diagnoses and health/mortality will be studied, as will sleep and later dementia (We will use the Swedish Twin Registry, The Swedish Patient Registry and the Swedish Medical Products Agency).

1.4 Combinations of stress and sleep in the prediction of health

One important hypothesis in the SSC work is that restitution through good sleep may counteract negative health effects of stress and poor sleep may exacerbate negative health effects. The SLOSH cohort now contains sufficient number of waves and follow-up register data to permit a comparison of the health effects of high stress vs good sleep and low stress vs poor sleep (as well as other combination). These possibilities will be investigated in collaborative studies (S+E)