Self-perceived stress reactivity is an indicator of psychosocial impairment at the workplace

Organisational restructuring, downsizing, and increasing pressure on management in keeping up with performance figures have radically transformed the nature of work [1]. Chronic job stress can be associated with a range of debilitating health outcomes, including cardiovascular diseases [2, 3], depression and anxiety [4–7], musculoskeletal problems [8, 9], or alcohol dependence [10, 11]. Stress-associated health problems at the workplace lead to approximately 50-60% of total sick leave days [12]. Within the European Union, 350 million work days are lost because of stress-related ill-health, resulting in an overall cost of at least 20 billion € each year [13]. In the United States alone, businesses loose 300 billion $ annually as a result of lowered productivity, days absent, and health-care costs due to stress [14]. Increasingly, experts call for the implementation of workplace health promotion programmes, as psychological problems, stress-related physical complaints [3, 14], perceived work stress [3, 14–18], and obesity [19] have negative effects on workers' health as well as overall productivity [14, 20–23]. In particular, employees who suffer from chronic job stress should be identified at an early stage so that preventive action can be taken.

So far, no unanimously accepted psychometric or biological indicator exists for the early identification and evaluation of individuals suffering from chronic job stress [24–26]. In human stress theories, it is assumed that stress arises from person-environment interactions. Thus, in addition to situational factors representing stressful demands, the ability to cope with job stress is assumed to be influenced by subjective emotional and physiological reactions in different stressful situations, which has been conceptualised as stress reactivity [27]. Somatic parameters for physiological stress reactivity, like salivary cortisol, heart rate, cardiovascular responses, and arterial catecholamine reactivity are well known. But there is still a lack of self-report questionnaires focussing on subjective, self-perceived reactions to stressful demands and situations. The stress reactivity scale (SRS) [28] is a self-report questionnaire which assesses on multiple levels typical subjective reactions to different stressful situations. It refers to the disposition of a person to answer stressors with immediate, intense, and long lasting stress reactions. By focussing on the frequency of recent stress events this concept specifically reflects on environmental components with regard to the person-environment interaction in the stress process [27]. In contrast to more traditional stress concepts based on personality traits this specific focus makes the SRS more sensitive to the measurement of change. In a laboratory stress study self-perceived stress reactivity has recently been shown to be related to personality traits, bodily complaints, sleep-related variables and cortisol reactions in a dermatological patient sample. Thus, on a psychological level, perceived stress reactivity may be a valid marker for current as well as prospective adverse health outcome [27–29]. However, this hypothesis has not yet been verified within a workplace sample.

Given the very small empirical basis of self-perceived stress reactivity and the complete lack of studies in workplace samples, there is a great need to better understand the correlations and interactions between self-perceived stress reactivity and different measures of chronic stress, job stress, somatic parameters for stress, cardiovascular risk factors, and psychosocial health outcomes. Industrial employees in a middle management ("sandwich") position seem to be an especially interesting group to test for this concept [1, 3].

The present study therefore presents data of a work-health promotion study in a large sample of middle-aged industrial employees with leadership responsibility in a typical middle management position of a large international manufacturing company. The primary aim of this study was to analyse the relationship of dispositional self-perceived stress reactivity with various psychological as well as biological indicators of stress and health. Therefore, we hypothesised that high stress reactive employees have a significantly worse status of health and stress-related phenomena than for low stress reactive employees (H1). The secondary aim of this study was to identify potential key variables that explain part of the observed variance in the self-perceived stress reactivity. We hypothesised that heightened stress reactivity can be predicted by psychosocial as well as biological parameters of stress and health (H2).

Referring to the main hypothesis (H1), our study shows that in industrial employees holding a middle management ("sandwich") position in manufacturing, increased individual stress reactivity is indeed associated with heightened psychosocial work stress, a higher rate of psychosomatic and physical complaints, poorer psychological health, and negative health behaviour. In addition, high stress-reactive employees experience an increased effort-reward imbalance (especially less job security and less esteem), show higher levels of depression and anxiety, and suffer more often from musculoskeletal and cardiovascular complaints. Compared to the normal population, our industrial workplace sample showed poorer results in terms of self-rated chronic stress, effort-reward ratio, anxiety, a variety of other related psychological problems, and obesity. These findings are of considerable interest since a study on work-related stress in a population of employed Swedish women showed an association between work-related stress, illness symptoms and sick-leave [43]. Furthermore a recent meta-analysis [44] reported psychosomatic complaints and/or suffering from psychological problems as significant predictors to sickness absence. Our findings may reflect the overall growing economical and organisational pressure of increased productivity which our sample is dealing with. Indeed, like in many other industries and places worldwide, production in our sample increased by 20% between 2003 and 2007 while keeping the workforce numbers nearly at the same level.

In respect to H2, our data revealed that education, cardiovascular complaints, chronic stress, and effort-reward imbalance were moderate predictors for perceived stress reactivity. Contrary to our expectations and to the few findings available of somatic correlates of stress phenomena in working place samples [26, 45], we did not find any relationship between perceived stress reactivity on one hand and specific biological indicators of stress or generic measures of somatic stress reaction, e.g. allostatic load index, on the other. Thus, our preliminary findings do not support data which suggest that prolonged psychosocial job stress [14] affects the biological stress axis and may consecutively cause severe illnesses and negative health outcomes. The following reasons may explain our results:

1. Unlike the only workplace study investigating the interaction of job profile and allostatic load [26], our participants were selected primarily from a middle management ("sandwich") position within the production line of an international manufacturer. Better work position and specific job demands of our middle-aged target group may have led to better health and coping capabilities in comparison to a sample of predominantly blue collar workers. 2. Especially with regard to biological parameters, a healthy worker effect may be assumed i.e., those in employment are known to be healthier than those not working [46].

Some more general factors may also explain our cross sectional finding: lack of longitudinal data, age of the participants, and a selection bias due to voluntary participation in the study.

We do recognise that there are some additional limitations of our study. First, our results cannot be generalised to other working populations. As participation was voluntary it can be assumed that all participants in the current study were in at least some positive stage of change [47]. Based on the clinical impression due to the interviews at baseline, a selection bias can be assumed because most of the participants were highly self-motivated with respect to their attitude towards demands of work and health. Second, as self-reported data on stress may lead to underestimation or overestimation of the real work stress situation, we are not able to objectively describe the work load of our sample. Third, our study faces a well known dilemma in stress research of bridging the gap between biological and psychosocial concepts and methods of stress assessment. The dilemma becomes even clearer as there is still a lack of standardised norm and cut-off values for job stress. However, here we present cross sectional findings only, but especially longitudinal data might be useful for a better understanding of the relationship and pathways between psychosocial and biological stress parameters and their impact on disability, diseases, and workers' productivity.

Finally, we conclude that the concept of stress reactivity as measured by the SRS correlates with various important aspects of psychosocial impairment and maladaptive health behaviour and thus indeed may be a valid tool to identify employees potentially prone to suffer from chronic job stress.

The results of this cross sectional study in a sample of middle management employees in the production line of an international plant demonstrate that in a relevant subgroup of the whole sample psychosocial stress parameters and well known health risk factors are significantly increased. Self-report data on stress reactivity allow to identify employees who are under stress according to established models and who suffer from stress-associated complaints. They may thus be at risk for severe stress sequelae. However, this cross sectional study failed to demonstrate a significant relationship between psychosocial and biological stress indicators. Longitudinal data may further elucidate the pathways of perceived stress and health status, and thus may prove if the measurement of dispositional stress reactivity may indeed serve as a predictor for adverse long term stress-related health outcomes.