Work-related stress is widespread in modern societies. In Europe, every second employee reports to frequently suffer from stress at work, with every fourth being permanently adversely affected by it [1
]. Work-related stress can be caused by different psychosocial or physical conditions. The focus here will be on work-related psychosocial stress. Chronic psychosocial stress at work has been measured with different models and it has been shown to be an important risk factor for health. The model of effort-reward imbalance (ERI) is one of the best-investigated psychosocial stress measures [2
]. According to this model, occupational stress can be the result of a failed social reciprocity between efforts spent (e.g., work load, responsibility) and monetary as well as non-monetary rewards (esteem, job security, job promotion) received in turn [3
]. ERI has been shown to be associated with common adverse health outcomes [4
]. Recent evidence supports this notion for coronary heart disease [5
], depressive disorders [6
], and burnout [7
Leisure-time physical activity (LTPA) is a further factor that is associated with common physical and mental diseases, and it has been shown to exert protective effects on coronary heart disease [8
] and depression [10
]. There is scientific evidence regarding the stress buffering effect of physical activity on health [12
]. Yet, the effect of LTPA on work-related stress—one of the major stressors of today’s life—is not investigated as well as commonly assumed. There is some epidemiological evidence regarding the long-term beneficial effect of LTPA on work-related stress. One study reports a protective longitudinal effect of light and moderate-to-vigorous physical activity on self-reported stress level, when compared to a sedentary life style in health care and social insurance workers [10
]. Furthermore, LTPA had a beneficial long-term effect on job strain in young adults (imbalance between high job demands and low job control, [13
]) in their mid-years [14
]. To our knowledge, the effect of LTPA on work-related stress measured by ERI over time was not investigated so far.
A negative effect of psychological stress on LTPA has been widely investigated, e.g., in a systematic review including 55 prospective studies [15
], and it has also been observed for work-related stress [16
]. Yet, the reverse association is less well studied [17
With this study, we set out to analyze the effect of LTPA on work-related stress in terms of effort-reward imbalance in a longitudinal study design.
Work-related stress, as measured by the ERI-R at t1, was significantly associated with ERI-R at t2. LTPA was a significant moderator of the association of ERI-R over the two study waves (t1; t2): The higher the amount of weekly LTPA, the weaker the positive association between ERI-R at t1 and t2. The significant moderating effect of LTPA on ERI-R over time persisted after adjustment for socio-demographic variables. Of the socio-demographic covariates, only sex was a significant predictor of ERI-R over time to the disadvantage of women.
The fact that we observed a dose-response relationship of the moderating effect of LTPA on work-related stress over time in the expected direction supports the assumption of a causal association, as discussed by Bradford Hill [22
]. Moreover, these findings are in agreement with earlier studies addressing the beneficial effects of LTPA on work-related stress in general, and with a specific longitudinal study showing a moderating effect of LTPA on chronic work-related stress [23
]. The consistency of independent study findings is a further criterion of epidemiologic causality, as discussed by Bradford Hill [22
] (see also [24
]). Education has the advantage of stability over time in comparison with other indicators of socio-economic status. This also holds true for the age cohorts born in 1959 and 1965 in our study, as there was a low probability of a fundamental change in educational status. In our longitudinal analysis, we could not find a significant association between the educational level and ERI-R.
Our results give support to a preventive effect of LTPA on work-related stress (ERI-R). According to Gerber and Pühse [12
], a preventive effect on health is assumed if exercise contributes to a lower level of stress. This was shown, for instance, in the Copenhagen City Heart study, where increased exercise was followed by decreased perceived stress among participants [25
]. Our finding of a dose-response relationship is in accordance with this result.
While physiological measures of work-related stress, like heart rate variability (HRV), were not included in this study, its results are nevertheless in accordance with findings that pointed to an association of high physical activity with high HRV, a physiological correlate of low level of psychosocial stress [26
]. Along this line of argument, fitness training might enhance the ability of cardiovascular systems to control responses to acute stressors as well as recover faster from stress [26
Besides physiological adaptation to stress, it might also be that LTPA strengthens personal resources, such as self-esteem or social support [28
], which contributes to a more favorable evaluation of one’s effort and reward at work. A study on the association between physical activity and work-related stress by ERI suggests that LTPA might strengthen the personal resilience to effort-reward imbalance [17
]. While this study did not provide longitudinal evidence this could partly explain our findings of a beneficial effect of LTPA on work related stress over time. In the study of Gerber et al. [17
], higher resilience to ERI was associated with less mental health problems. Our study did not test the so-called ‚stress-buffer hypothesis’ that assumes a buffering effect of physical activity on the health damaging effects of stress [28
]. This effect has been also observed in the study of Klaperski et al. [23
]. Further research along these lines is warranted although a direct test was not performed in our investigation.
Last but not least, it can be meaningful to investigate the effect of LTPA on work-related stress over time, independently of a health-related outcome in the occupational context: ERI was also associated with other undesirable work-related effects, like higher intention to retire early [29
], poor work performance, and reduction of work productivity [30
]. Proving that LTPA had a positive effect on ERI over time means that it could also be protective against all of these adverse work-related outcomes associated with ERI.
While the lidA-study has a large number of employees and is a prospective cohort study [19
], there are some limitations of this analysis: the generalizability of our results to all employees might be limited, as only two age cohorts are included in the lidA-study and only those who are paying social security contributions (about 85% of all employees in these age groups). The relatively low response rate of 27.3% in the first study wave has to be seen from the vantage point of the observed high representativeness of our sample to all social secured employees in 16 different socio-demographic variables in comparison with the data of the ‘Integrated Employment Biographies’ (IEB), where the study sample has been drawn from [19
]. The IEB dataset includes all employees subject to social security in Germany [19
]. The loss to follow-up especially occurred in male employees, and missing values in effort-reward imbalance must be identified as clear limitations.
Moreover, additional unmeasured factors may lead to a limited predictability of our results: We only included two study waves; therefore, time trends cannot be recognized in our analysis. Another limitation, the measurement of work-related stress, was restricted to the effort-reward imbalance model. Furthermore, LTPA was only measured with one item. With this measure, an examination of intricacies (e.g., type of activity) of LTPA or an analysis of whether LTPA is more effective in work-related stress reduction by ERI before or after work was not possible. Both aspects should be considered more detailed in future investigations. On balance, as a particular strength of this study, we applied a prospective cohort studies, based on a large sample, and we tested a relevant, not yet well studied, research hypothesis on the link between LTPA and ERI over time.
LTPA might be helpful in reducing work-related stress in older employees suffering from an imbalance between high efforts spent and low rewards received in turn. As work-related stress is a widely prevalent phenomenon in modern working life and it exerts tangible negative effects on different health-related and work-related outcomes, our finding of an association between LTPA and ERI-R over time might be useful in planning sustainable interventions in occupational public health. LTPA might be beneficial in mitigating the health hazards of ERI (e.g., cardiovascular risk, depressive disorders) and their long-term consequences (e.g., loss of work force). LTPA should be promoted, therefore, in preventive activities aiming at stress reduction. Future prospective studies are required to further examine and extend our findings, applying objective data (e.g., measuring physical activity with accelerometer and work-related stress with biological markers, such as cortisol or heart rate variability). Moreover, intervention studies aimed at investigating the effect of LTPA on work-related stress should be implemented to strengthen the current base of evidence.