Abstract
The present study is a review of the effectiveness of preventive stress interventions. First of all, it was tried to determine the effect size of a sample of such interventions. Secondly, it was assessed whether this sample was homogeneous. As a third objective, it was tried to identify some of the working factors, or effect predictors, of the stress interventions. It was hypothesised that the following characteristics would be positively related to effect: carrying out the intervention after working hours, a high-risk target
group, an on site intervention location, a longer intervention duration and the use of multiple methods. Furthermore, a differential effectiveness on the basis of intervention method was expected. Organisational interventions were expected to be equally effective as worker-oriented ones. At follow-up, organisational
interventions were expected to increase in effect relative to worker-oriented ones. Finally, it was expected that the quality of the intervention design would be effect-related. As an additional factor, the effect of publication year was assessed. In order to address the research questions, a meta-analysis was carried
out. Through an extensive literature search, 47 evaluation studies were retrieved. These articles were coded using a specialised coding system. On the basis of their outcomes, a standardised measure of
effect size was calculated. Results showed a medium intervention effect on psychological outcomes. Small effects were found for behavioural outcomes, whereas the effect for cognitive, physiological and organisational measures was very small to zero effects. Due to 5 outlier studies, the sample was heterogeneous. As expected, after work hours programs were (marginally) more effective. Moreover, the quality of the interventions design was positively related to effect. Contrary to the expectations, results
indicated that individual-oriented and off site programs were more effective. Finally, publication year was negatively related to effect. The other results were non-significant. The statistical conclusions were supplemented by a qualitative analysis. This largely confirmed the quantitative findings and suggested a
number of additional factors that appear effect-related. At the end of the study, suggestions for further research are given.
Table of contents
Effectiveness of Preventive Stress Interventions
ABSTRACT
INTRODUCTION
CHAPTER 1: THE STRESS PROCESS
1.1 Introduction
1.2 Strain reactions
1.2.1 Physical reactions
1.2.2 Behavioural reactions
1.2.3 Psychological reactions
1.2.4 Societal impact of stress
1.3 Stressors
1.3.1 Types of stressors
1.3.2 Is the presence of stressors on the work flour increasing?
1.4 Mediator and moderator variables
1.5 An integrative model of the stress process
1.5.1 Introduction
1.5.2 Description of the model
CHAPTER 2: STRESS INTERVENTIONS
2.1 Introduction
2.2 Characteristics of prevention programs
2.2.1 Actor/program provider
2.2.2 Target group
2.2.3 Types of g oals
2.2.4 Risk and protective factors
2.2.5 Instruments
2.2.6 System level
2.2.7 Time aspects
2.2.8 Setting of the intervention
2.3 Methods for stress reduction
2.3.1 Individual methods
2.3.2 Organisational methods
2.4 Implementing the program
CHAPTER 3: METHOD
3.1 Introduction: Why perform a meta-analysis?
3.2 Description of the analysis
3.2.1 Research questions
Effectiveness of Preventive Stress Interventions iii
3.2.2 Literature search and selection of studies
3.2.3 Coding system and procedure
3.2.4 Calculations of effect size
3.2.5 Analysing, reviewing and interpretation of the results
CHAPTER 4: RESULTS
4.1 Research question A: Magnitude of effect size
4.2 Research question B: Homogeneity analysis
4.2 Research question C: Decisions on research hypotheses
4.2.1 Quantitative analysis
4.2.2 Qualitative analysis
CHAPTER 5: DISCUSSION
5.1 Overview of results
5.1.1 Research question A: Magnitude of effect size
5.1.2 Research question B: Homogeneity analysis
5.1.3 Research question C: Decisions on research hypotheses
5.2 Implications for prevention theory
5.2.1 Earlier effect reviews
5.2.2 Implications for effect hypotheses
5.3 Limitations of the study
5.3.1 Ambiguities in the coding procedure
5.3.2 Small sample
5.3.3 Keeping the outliers in the quantitative analysis
5.3.4 Validity of the qualitative analysis
5.4 Recommendations for future theory and practice
5.4.1 Implications for future studies
5.4.2 Implications for stress and intervention theory
5.4.3 Implications for future practice
REFERENCES
LIST OF INCLUDED STUDIES
APPENDIX 1: LIST OF USED PSYCLIT SEARCH TERMS
APPENDIX 2: SHORTENED VERSION OF THE CODING SHEET
AP PENDIX 3: OPERATIONALISATION OF “DESIGN” ITEMS
APPENDIX 4: DESCRIPTION OF OUTLIERS
Abstract
The present study is a review of the effectiveness of preventive stress interventions. First of all, it was tried to determine the effect size of a sample of such interventions. Secondly, it was assessed whether this sample was homogeneous. As a third objective, it was tried to identify some of the working factors, or effect predictors, of the stress interventions. It was hypothesised that the following characteristics would be positively related to effect: carrying out the intervention after working hours, a high-risk target group, an on site intervention location, a longer intervention duration and the use of multiple methods. Furthermore, a differential effectiveness on the basis of intervention method was expected. Organisational interventions were expected to be equally effective as worker-oriented ones. At follow-up, organisational interventions were expected to increase in effect relative to worker-oriented ones. Finally, it was expected that the quality of the intervention design would be effect-related. As an additional factor, the effect of publication year was assessed. In order to address the research questions, a meta-analysis was carried out. Through an extensive literature search, 47 evaluation studies were retrieved. These articles were coded using a specialised coding system. On the basis of their outcomes, a standardised measure of effect size was calculated. Results showed a medium intervention effect on psychological outcomes. Small effects were found for behavioural outcomes, whereas the effect for cognitive, physiological and organisational measures was very small to zero effects. Due to 5 outlier studies, the sample was heterogeneous. As expected, after work hours programs were (marginally) more effective. Moreover, the quality of the interventions design was positively related to effect. Contrary to the expectations, results indicated that individual-oriented and off site programs were more effective. Finally, publication year was negatively related to effect. The other results were non-significant. The statistical conclusions were supplemented by a qualitative analysis. This largely confirmed the quantitative findings and suggested a number of additional factors that appear effect-related. At the end of the study, suggestions for further research are given.
Introduction
Stress is a very complex phenomenon. Take for the example the number of disciplines that occupy themselves with the subject. Medicine is contributing with findings of biochemical correlates of stress and has demonstrated that it can have an effect on the number of immune cells in the human body. Economists are focusing on the effects that changes in production methods have on stress and how the costs of this development can be prevented or reduced. There are also scholars who have pointed at the ideological background of the stress concept (or “discourse”, in Foucauldian terms). And there is of course psychology, a heterogeneous discipline in itself, where schools such as humanistic psychology, cognitive -behavioural psychology and labour psychology all have different things to say about the topic. Maybe this diversity is one of the reasons that there is not yet a widely accepted definition and theory on stress (Gaillard, 1996; Kahn and Byosiere, 1992). Despite of this, academic interest is vastly increasing. Critics have remarked that much of the attention seems to be due to social and scientific trends (Newton, 1995). However, there appear to be very objective and consequential reasons too. As can be seen in later in this essay (chapter 1.2), one of the most important causes for this vast attention lies in the outcomes of stress, which are well described as “real, painful and costly” (Kahn and Byosiere, 1992).
As shall be described in the following chapter, these aversive consequences of stress are large in number and serious in nature. All of the above mentioned disciplines encounter such stress related problems. For the medical profession, stress mainly appears as biological reactions. Such physiological changes belong to the first documented results of stress. Hans Selye for example, one of the founding fathers of stress, studied the responses of rats in reaction to aversive environmental stimuli. The changes he reported included enlargements of the adrenal gland cortex, involution of thymus and lymphatic structures, and the development of ulcers (Murphy, 1996; Newton, 1995). For people, stress can have very serious physiological consequences on cardiac, biochemical and gastrointestinal systems (Kahn and Byosiere, 1992). Moreover, a long-term consequence of stress is the “wear and tear” of the body’s immune system. One reason for this is that cortisol, an important stress hormone, suppresses the body’s immune cells. Another class of responses to stress pertains to behaviour. Stress can cause multiple aversive changes that can negatively impact both the person himself and broader systems, such as his organisation or family. Examples of this include alcoholism, absenteeism and role decrements.
Most important for the preventive psychologist, however, are outcomes relating to psychological health and ill health. Does stress affect people’s minds just as it does their bodies and behaviour? Almost any stress theory would answer this question with yes: there are numerous well-established psychological changes that have been linked to variations in stress levels. In one review by Kahn and Byosiere (1992), as many as 43 possible outcomes are mentioned. Among these outcomes are organisation-related factors as job satisfaction and commitment; emotional states like boredom, confusion and frustration; and self-concept variables such as self-esteem and self-confidence. Moreover, stress has been linked to psychopathological symptomatology such as depression, anxiety and irritability (Murphy, 1996). This connection is supported by a number of sources, which will be elaborated in more detail in the first chapter. Here, research shall be presented that found significant correlations between stress on one side and depression and anxiety on the other. Moreover, the risk factors for depression share a great deal of face-similarity with the stress construct. For example, what to think of “a high level of aversive events”, a risk factor reviewed by Veltman and Hosman (1996), which seems to overlap very clearly with a stressing environment.
Besides reviewing the most important outcomes of stress, an overview of stressors and so-called mediator/moderator variables shall also be given. To begin with the first, a stressor could be defined as a threatening factor in or outside the individual (Lazarus, 1995). One area that has been reported as potentially very stressful is the workplace. The essay shall give a brief overview of the most important ones, such as role ambiguity and job uncertainty. A “stressor” does not always produce the same fixed outcome, however. Its impact is greatly influenced by other important factors, such as the relative gravity of the stressor. Getting new neighbours is probably not going to produce such dramatic changes as the death of a loved one. Factors that have a direct influence on the chain of effects are called mediator
variables 1. Crucial in these mediator variables is the way a person copes (deals with) a stressor. However, choice of an appropriate coping mechanism is in turn influenced by so-called moderator factors2. This choice depends on the person of the worker and his and social support (Gaillard, 1996). The most common types of personality and social support factors shall be reviewed in chapter 1.
When the role of stressors, stress reactions and mediator/moderator factors will h ave been clarified, the essay shall provide a model developed by Hosman (1998) to summarise this information. This Integrative Stresstheoretical (IS) model is an effort to integrate research findings into one coherent whole. It consists of multiple steps such as the existence of a stressor, its perception and appraisal by the individual, and the preparation and execution of the effort to cope with the stressor. The model, which is presented in chapter 1.4, shall include most of the evidence that is reviewed in chapter 1 and shall thus be a very comprehensive and useful theoretical background to understand the stress process. With the help of this framework, it is also easier to understand the rationale behind the numerous efforts to intervene in the stress process. Such a deeper understanding is critical for the future design of more effective and successful programs.
In stress interventions, a number of distinctions can be made. What is the target group for a certain program? Does the program explicitly state its goals or remain vague about them? Was the program carried out with people that were already at risk for stress complaints or was the intervention implemented in a group that did not yet show signs of developing problems? The number of such distinctions is in principle endless. They might sound arbitrary too, if it wasn’t for the fact that many of these factors actually seem to make a difference. For some categorising variables, at least, it can be proven that they are linked with eventual program effectiveness. Variables for which this is the case are called effect predictors, because their presence in the program characteristics predicts more successful results. The present essay shall analyse the data of a number of evaluation studies in order to reach some conclusions on the determinants of such success. For this purpose, a number of effect hypotheses were generated that, when confirmed, result in evidence for the existence of effect predictors.
Determining such effect predictors would be highly useful. For a long time, there did hardly exist effect studies, so the value of preventive efforts was not without doubts (Kahn and Byosiere, 1992). More recent studies did research the effectiveness of stress interventions, though, and this has led to the conclusion that they can indeed contribute to reductions in stress (Murphy, 1996). What remains unclear, however, are the mechanisms that contribute to this change. Moreover, little is known about the so-called efficacy of preventive stress programs. Whereas effectiveness refers to the success of a program within a given context, efficacy pertains to the question whether the program as such would also be effective in other contexts and with other providers. When drawing this distinction to the stress field, it is striking to see that almost all studies are developed and implemented by the researchers themselves. When such programs are shown to be effective, it is a reasonable question whether this has more to do with the
particular characteristics of the program’s providers or implementation context than with the features of the program itself.
A potentially very powerful effect predictor is the method that is used to reduce stress. In selecting a methodology, one can choose between a variety of approaches. In the worker-oriented category, these include cognitive -behavioural skills training, which targets cognitive structures and coping skills of the client. Moreover, worker-oriented interventions include muscle relaxation, exercise and biofeedback. These interventions all involve physical changes that are either beneficial in reducing already built up tension or make the body more “fit” to handle new stressors. Finally, there is meditation, which is based on invoking a state of mental calmness. In the category of environmental interventions, different methods can be most easily characterised according to their target. Here, there are three classes of interventions. Firstly, it is possible that the program targets the amount of social support the worker experiences. Secondly, the degree of control, or decision latitude, can be enhanced. Finally, it is possible to target stress in a very direct way by taking away or reducing the stressor itself (see chapter 2 for a more detailed description of organisational programs).
As mentioned above, the techniques of biofeedback, exercise and relaxation all focus on physical changes. Other methods, such as cognitive-behavioural training, target the psyche of the worker. One hypothesis is that these differences in focus will be reflected in different patterns of effectiveness. Such a prediction is supported by Murphy’s (1996) review on worker-oriented interventions, in which very outcome-specific effects were found (e.g. muscle relaxation and biofeedback were only effective on physiological outcomes). It is hypothesised that this will again be the case : methods are mainly effective on the type of processes at which they aim. Another hypothesis concerns the distinction between worker- and environment-oriented interventions. Here, it is predicted that organisational measures are at least as effective as individual ones. This would be in line with numerous authors (e.g. Daniels, 1996), who state that the two approaches are equally important in reducing stress. A third hypothesis concerning the used method is that the longer the follow-up, the more effective organisational interventions become in relation to individual measures. One of the reasons for this effect could be that environmental programs take away the stressor, whereas worker-oriented ones focus on individual changes. To maintain their results, the latter would require booster sessions but, as these are seldom provided, their effectiveness is expected to diminish.
Often, programs do not stick to a single method, though. It is quite common that an intervention uses more than one method to accomplish a reduction in stress. Suppose for example that highly tensed personnel is first being massaged in order to relieve the worst tension. In a second phase they could then receive cognitive-behavioural skills training to prevent such a build-up in the future. A hypothesis is that such multi -method programs are more effective. The background for this hypothesis roots in the s o-called span of change that such interventions are able of. Suppose that, say, counterproductive cognitions account for 10% of the variance in depressed mood and stressing interpersonal interactions account for another 10%. Programs that have both cognition and interpersonal components thus have a span of 20% of possible change, which is much higher than each separate component would have. Span of change could also be a rationale for another effect hypothesis, namely that programs aimed at a relatively high-risk group are more effective than programmes that do not. High-risk workers probably have a larger number or more powerful risk factors for stress problems, which makes their span of change considerably higher.
In two hypotheses, learning principles play an important role. The first prediction involves the effect of the timing of a program. Participants of programs that are run during the working day are probably still stressed. After all, work is not yet done and demands are still waiting once the intervention has ended. Subjects of after working hours programs are thought to be less stressed, as for them, the working day is over. For these after-hours participants, learning is thought to be easier. This prediction is backed up by the so-called Yerkes -Dodson law of learning, which states that high levels of stress interfere with the learning of new skills (Liebermann, 1999). Therefore, after working hours programs are hypothesised to be more effective. Another hypothesis that roots in a learning principle is the prediction that on site interventions have greater success. The principle that plays a role here is that learning generalises more easily to situations that are similar to the environment where the new knowledge or skill has been
acquired. Because of this, stress management skills that are learned at the workplace are expected to generalise more easily to practical work situations. Thus, more stress reduction can take place.
A final hypothesis has to do with the quality of the methodological design of the intervention study. This is different across studies because researchers use reliable methodological standards, such as control groups and randomisation, whereas other studies are more loosely set up. In this essay, a hypothesis is that the quality of the research design is linked with program effect. Such a link with effectiveness was also suggested by Wolf in his 1986 review of the meta-analysis technique. There are a number of possible reasons for this effect. When the study procedure does not use an equivalent comparison group, for example, it is possible that this leads to higher reported effect. Maybe the experimental group was more motivated to change and therefore more eager to comply with the program’s goals. In this case, methodological design would be negatively related to effect.
In total, there are thus 9 different effect predictors that shall be addressed. Reaching a conclusion on these hypotheses is not as easy as it might seem however. How can it ever become clear that factor X predicts effect in program Y? For this problem, there seem to be several alternative solutions. One alternative is to set up an experiment and see if an experimental group receiving a program with factor X is better of than a control group with the same program minus that factor. This is relatively troublesome, however, because it is very hard to draw hard scientific conclusion on the basis of one study. After all, success can be due to other factors than the program itself, such as the motivation of the provider. Another possibility is the exploration of subject expert opinions. However, these have the disadvantage of being influenced by all kinds of human biases. Thus, valid conclusions regarding the research hypotheses are both threatened by lack of reliability and a high degree of subjectivity. In this essay, it was therefore chosen to perform a meta-analysis. This is a technique that makes it possible to link certain previously coded variables to a measure of effect size. This method has the great advantage that it is not limited to a single research because a greater amount of studies is used. Moreover, it is relatively objective because it uses statistics instead of human opinions (Wolf, 1986).
The present meta-analysis will be performed as follows. First of all, studies that provided evaluations of relevant stress interventions were retrieved. This was done using an extensive search procedure in scientific databases such as PsycLit and references in relevant articles. When this was done, they were coded using a specialised coding system that was designed for classifying prevention programs. The result of this step is a set of quantified variables that can be statistically elaborated. The most important variable in this process is the measure of effect size called Cohen’s d. This is a statistic that is calculated on the basis of the effect of a study compared to its Standard Deviated, which makes possible a standardised comparison across studies. In the statistical elaboration of the data, the different coding variables were then linked to this effect size measure. This will make it possible to see whether one of these coding characteristics is related to study effect. In this case, strong evidence for the existence of an effect predictor shall have been found.
To summarise the structure of the essay, the first chapter shall present some evidence on types of stress reactions, stressors and mediator/moderator variables. Also, a model that tries to integrate these factors (the IS-model by Hosman, 1998) shall be introduced here. After this, programs that aim to prevent such stress-related problems will be described in chapter 2. This chapter shall also include a listing of the most important dimensions of these programs. Moreover, some of the reasons to prevent stress shall be covered, as well as a review of common types of programs. These programs can be described in multiple characterising variables. When such a variable is linked with program effectiveness, it is called an effect predictor (chapter 2). This study is concerned with the identification of some possible effect predictors. The method that shall be used for this purpose is a meta-analysis, of which a full description shall be given in chapter 3. An important part of this procedure is the coding of study characteristics into research variables on the basis of which statistical inferences can be drawn. It is hoped that this makes possible a conclusion on the above-mentioned hypotheses is possible (for a more detailed description of the hypotheses, see chapter 2; for the results of the analysis, see chapter 4). Finally, in the last chapter (chapter 5), the outcomes of the analysis shall be discussed in terms of their implications for the current state of the field and its direction for the future.
Chapter 1: The stress process
1.1 Introduction
When tracing back the history of stress reach, the first great hallmarks originate in the early 20th century. There, scientists like Hans Selye and Walter Canon came with the 0first academic reports. They found out that certain environmental conditions (such as dangerous animals or chemical toxicants) produce a universal pattern of reactions in organisms. For example, Selye injected laboratory rats with various toxic substances and exposed them to extreme temperatures. He observed that this lead to a universal pattern of physiological changes, which included an enlargement of the adrenal gland cortex, involution of thymus and lymphatic structures, and the development of ulcers. He termed this universal pattern the general adaptation syndrome (GAS), which is a precursor of the stress concept. Another “founding father” of stress was Walter Cannon. He also related to the stress concept when he described a general human reaction to dangerous situations. This reaction, clearly rooted in evolutionary history, is either to fight the attacker or to run away from it. To prepare these reaction modes, the subject is physiologically aroused through bodily reactions such as elevated heart rate and blood pressure, increased blood flow to so-called sympathetic regions (such as major muscle groups or the brain) and a decrease in vegetal functions. It were findings such as these that first indicated that man’s reactions to threatening environmental stimuli can influence the body and, as would become clear soon after, also the mind. With these findings, the position of stress as a construct of interest was established. In earlier days, the word stress only referred to some kind of external pressure on an object. Since Selye and Cannon, it became known in the modern meaning of “the non-specific response of the body to any demand” (Murphy's definition, 1986).
This above-outlined process by which a strain reaction results from a threatening stimuli is called the stressor-strain process. In this terminology, stressor stands for some threatening factor inside or outside the individual, whereas strain pertains to the reaction to this (Lazarus, 1995). An example is the development of ulcers as a result of chronic work overload. Here, work overload is the stressor, whereas the ulcers can be thought of as the resulting strain reaction. In a model, this process can be outlined as follows.
Figure 1.1
illustration not visible in this excerpt
The process that is outlined by the model makes clear that there are always at least two aspects involved in stress: a stressor and a stress reaction. As a basis for further elaboration, this definition is very useful. However, many important aspects remain quite general and vague. When a stimulus can be treated as a stressor is not stated, for example. Moreover, how a stressor results in strain is not made explicit. Therefore, this chapter shall further elaborate both sides of the stressor-strain relationship. In the first section (1.2), the strain side shall be discussed. The most important types of strain reactions, which include reactions of physical, behavioural and psychological nature, shall be described. Special attention shall be given to psychological consequences of stress, on which the analysis shall focus. After this, in section 1.3, the most common work-related stressors shall be covered. The final category of stress variables is a number of factors that influence the relationship between stressor and stress. These mediator and moderator variables shall be introduced in the third section of this chapter. When all this is done, a model shall be presented that tries to integrate these interrelated variables into a structural whole. With the help of this so-called Integrative Stress-theoretical model, a more complete picture of the stress process will be achieved. This can serve as a background for reading and understanding the rest of the chapters.
1.2 Strain reactions
As stated in the introduction, defining stress is not an easy task. Especially, it is often difficult to specify the precise role that each of its different determinants has in relation to the others. This can for example be seen in the definition of Kahn and Byosiere (1992), which, focused at defining a stressor, is able to elaborate strain as well.
“A stimulus is defined as stressful (i.e., a stressor) because its effect is assumed, hypothesised and (cumulatively) demonstrated to have certain undesirable effects
-adverse physiological changes of the kind described by Selye […], decrements in role performance, emotional tensions, onset of physical symptoms such as sleep disorders and gastric disturbances and so forth- in the population about which we prognose to generalise.” (p. 575)
In this section, these reactions shall further be elaborated. For this purpose, responses to stressors shall be divided into three broad categories: physiological, behavioural and psychological reactions. Because the present study was written in the framework of clinical psychology and the prevention of mental illness, the accent shall lay on psychological effects. After the review of these three classes of responses, some broader consequences of stress shall be described. As shall be outlined here, stress not only results in biological, behavioural and psychological reactions, but also in very significant medical, societal and economic costs.
1.2.1 Physical reactions
As can be read back, the first “discoverers” of stress focused mainly on bodily reactions to external threats. After them, much additional research has followed and this has lead to the establishment of a great deal of other outcomes. Following a distinction by Kahn and Byosiere (1992), three categories of bodily reactions can be distinguished. First, there are cardiovascular symptoms, such as blood pressure, cholesterol level, cardiac activation, etc. Secondly, there are biochemical outcomes, which include alterations in catecholamines (such as epinephrine, norepinephrine and dopamine), corticosteroids (cortisol) and uric acid. Thirdly, there are gastrointestinal symptoms, which can for example exist of peptic ulcers. Although much of these outcomes are relatively short -term, there also exist serious long- term consequences. After a longer period of arousal, the body can become exhausted and show long- term physiological responses. This process is also known as wear and tear. In this process, stress can lead to chronic health effects. Although much of the mechanisms in this process are still unclear, the following factors have been shown to play a role (all from Gaillard, 1996).
-Overactivity
People under stress react more intense than necessary to external demands (Gaillard, 1996). Many researchers (such as Karasek and Theorell (1990) in Gaillard, 1996) see in this reaction the main cause of pathophysiological processes.
-Insufficient recovery
Central in this process is the fact that stress reactions leave behind a residue. When these residues are not removed, serious consequences can eventually accumulate. Insufficient recovery is often stated in combination with overactivity, which is because overactivity results in greater residues. The result of this is an accelerated accumulation.
-Higher resting values
Stressed individuals are known to have higher physiological activation while at rest (e.g. during sleep). These higher values have shown to be a good predictor of health damage, especially in the domain of cardiovascular activity.
-Sustained activation
Normally, acute reactions to stressors (e.g. adrenaline) are reduced to normal values when the threat is over. However, higher order mechanisms in the central nervous system, which are being controlled by psychological factors (e.g. anxiety), can block these deactivation systems. In this way, stress can lead to an enduring arousal of the nervous system.
-Disrupted set points
For certain kinds of physiological activation processes, the body has set points (normal values of activation). These set points are part of certain feedback mechanisms that regulate bodily activation. When these mechanisms are disrupted, this means that the psychobiological equilibrium is thrown off balance and physiological functions do not return to their starting values.
-Suppression of the immune system
It is known that sustained sympathetic activation and elevated cortisol levels lead to a decrease in immune system functioning. For example, people under stress produce less so-called killer-T cells. These cells are responsible for warding of alien cells, thereby protecting the organism for disease. This causes stressed individuals to be more susceptible to infections or viral threats.
1.2.2 Behavioural reactions
In the above section (1.2.1), bodily reactions to stress were described. Not only the body reacts to a stressor, however. In a second category of reactions, stress can also penetrate behaviour. This might not be surprising, because tension in the body produces parallel psychological stress. Sometimes, it stays there (as will be elaborated further when discussing the third class of psychological stress responses) and sometimes is be “acted out” in behaviour. There are many instances of such stress reactive behaviours. Because the current analysis shall also focus on behavioural reactions to worksite stressors, especially in the context of the organisation, it is interesting to review some the most important organisational effects. They shall be listed according to a division made by Kahn and Byosiere (1992), who grouped stressor-evoked behaviours along five broad categories. Although most of these categories pertain to labour/organisational life, they can be easily expanded to other life settings. What is important to realise is that stress can disrupt important life patterns or roles. For example, the relational equivalent of flight from job could be the termination of the relationship. Aggressive behaviours could focus on child abuse instead of on stealing on the job. In principle, stress can affect many realms of behaviour, but as stated above, the essay shall predominately focus on organisational outcomes. Below, the most important of these reactions are listed:
-Degradation/disruption of the work role itself
This consists of decrements in job performance, accidents and errors, and alcohol and drug use at work. These are very costly behaviours for employers because they lower productivity and effectiveness.
-Aggressive behaviour at work
This includes counterproductive acts, such as stealing, purposeful damage, spreading rumours, etc.
-Flight from job
When work becomes too stressful, workers often try to “escape” these aversive environments by absenteeism and turnover (leaving the organisation for other employment). Other escapist behaviours include early retirement and even strikes (Kahn and Byosiere, 1992).
-Degradation/disruption of other life roles
In this category, stressful work spills over into private and social life. Life roles that are been shown to suffer from stress include the role of spouse (e.g. spouse abuse), friend and citizen.
-Self-damaging behaviours
These behaviours can cause high medical costs. Excessive smoking, for example, can lead to lung cancer, a disease with often-fatal consequences and costly treatments. Other examples of self-damaging behaviours include alcohol and drug use, caffeine use and accidents.
1.2.3 Psychological reactions
Introduction
The third broad category of reactions to stressors consists of psychological reactions. Such psychological factors are highly important in the experience of stress. In modern industrial life, work is increasingly dominated by cognitive demands and it is therefore not surprising that the problems associated with adverse working conditions are psychological as well (Gaillard, 1996). Psychological effects have been reasonably well studied and are considered as painful and costly (Kahn and Byosiere, 1992). There exist many types of these responses. These include organisational/job outcomes such as job satisfaction, organisational commitment and turnover intent. Psychological reactions also include emotional states, such as boredom, confusion, frustration, irritation, etc., and self-concept variables like depersonalisation, self-confidence and self-esteem. Moreover, a common psychological strain reaction to these demands is burnout, a stress syndrome that is characterised by three symptoms: (1) the feeling of being extremely tired and empty (“burned out”), (2) a decreased involvement in others, especially in patients, students or clients, (3) and a lowered job competency (Gaillard, 1996). As a final type of responses, there is even evidence that stress is linked with psychopathological outcomes such as depression and anxiety.
Because it shall constitute the core of the present essay and its analysis, the stress-psychopathology link needs to be elaborated further. For this purpose, numerous sources were used. Firstly, some individual correlational studies shall be explored that researched the relationship between stressors and multiple outcomes of psychiatric distress. Secondly, a review of risk factors for depression by Veltman and Hosman (1996) shall be presented. In the following section, both of these sources shall be described in more detail.
Evidence from correlational research
There are a substantial number of correlational studies that support the conclusion that stress can be linked with pathology. Caplan and Jones (1975), for example, investigated the effect of a computer shutdown on psychological stress. 73 male users of a university computer system were used as subjects. Because these people were highly dependent of the university computer (note that in 1975 people did not have personal computers yet), an increase in stress level was expected. It was indeed found that workload and role ambiguity, two common measures of stress, were higher during the shutdown than in a control period. Moreover, workload correlated significant with anxiety, whereas role ambiguity was related to both anxiety and depression. Thus, although the study lacked a control group, it strongly indicated that stress can be linked to psychopathology.
In another study, Beehr (1976) held a large survey among 651 respondents. These included employees from a printing company, a research and development company, two automotive supply companies and four service departments of a hospital. The survey measured the relationship between role ambiguity and a large number of strain variables. All the included psychological variables (i.e. job satisfaction, life satisfaction, low self-esteem and depressed mood) correlated significantly with the role strain measure. Similarly, Ganster, Fusilier and Mayes (1986), collected data from 326 contracting firm employees. Measures included 6 common stressors (role, conflict, role ambiguity, work underload, lack of variability, underutilisation and responsibility) and 4 outcome variables (life dissatisfaction, job dissatisfaction, somatic complaints and depression). Of these, life satisfaction was associated with role ambiguity and underutilisation. Somatic complaints was linked to role conflict, role ambiguity and underutilisation. For depression and job dissatisfaction, connections we re even more powerful. Depression correlated significantly with all stressors but lack of variability and job satisfaction with all stressors except responsibility.
Evidence from epidemiological research
Other evidence that points to a connection between stressor and mental illness comes from research that addresses the risk factors for mental illness. In the section below, the example of depression shall be elaborated in detail. For this purpose, an overview shall be presented of the variables that are associated with the onset of a depressive disorder. When inspecting these variables, some clear points of overlap with the stress concept arise. The risk factors that are reviewed are based on a project designing a prevention program for depression among adolescents (Veltman, Ruiter and Hosman, 1996).
-A first risk factor that was found is negative self-valuation. Because one of the psychological outcomes of stress is lowered self-esteem and self-confidence (Kahn and Byosiere, 1992), stress could lead to a greater risk for depression.
-Veltman and Hosman (1996) also found a high level of aversive events to be related with the onset of depression. Because stress can be thought of as aversive environmental pressure, this risk factor has considerable overlap with the concept of stress itself. Again, this provides a link by which stress could be a cause of psychopathology.
-Internal, stable and global attributions in uncontrollable, negative situations were another risk factor. Because stressful situations are often uncontrollable (e.g. a company reorganisation or chronic poverty), stressed people with these thinking patterns have a considerable risk of developing depressive patterns.
-A fourth risk factor was having interpersonal problems. Because social interactions at work can be a considerable source of stress, a problematic and stressful social climate (e.g. conflicts with boss or quarrels with spouse) can be a risk factor for depression.
-Finally, life events and chronic daily difficulties in combination with vulnerability factors and a lack of social support were also related with depression. This calls to mind the fact that chronic stress can have a more serious influence than do life events (Kessler, Price and Wortman, 1985). Work related stress seems to fall more in the category of chronic stress and has therefore a high potential of triggering depressive patterns.
Summary
To summarise the chapter so far, some possible outcomes of the stressor-strain process were covered. These outcomes could be divided into three broad categories, namely physiological, psychological and behavioural consequences of stress. Physiological outcomes included aversive changes in cardiovascular symptoms, biochemical outcomes and gastrointestinal symptoms. It was shown that stress can result in chronic conditions, and a number of mechanisms in this “wear and tear” process were reviewed. The second category of reactions was made up by behavioural changes. These included the following 5 groups of problems: (1) degradation/disruption of the work role itself, (2) aggressive behaviour at work, (3) flight from job, (4) degradation/disruption of other life roles and (5) self-damaging behaviours. Examples in this category include alcoholism and absenteeism, and can result in very costly consequences. The third main group of stress outcomes was constituted by psychological variables. Stress at work can result in organisational/job outcomes such as job satisfaction; emotional states, such as confusion and frustration; and self-concept variables like depersonalisation and self-esteem. Finally, stress is linked with outcomes of psychopathology such as anxiety, irritability and depression. In the case of depression, an analysis of risk factors was presented that indicated an important role that stress can play in the development of this mental illness.
1.2.4 Societal impact of stress
When considering the outcomes of the stress process, it is clear that stress can give rise to very serious problems in the individual. These outcomes have been reviewed in the previous sections 1.2.1 – 1.2.3). Undesirable changes in the spheres of biology, psychology and behaviour do not exist in a vacuum, however. Not only stressors are located in an environment; so are stress outcomes. Individual outcomes, such as discussed above, have broader societal-economical consequences. Below, the most important ones are reviewed.
Societal costs
In the form of societal costs, stress can have a very destructive impact on the larger framework of society. When people bec ome depressed, for example, they can end up in clinics, or even commit suicide. Also, when people have to quit their job (because they are unable to perform any longer), they are robbed of a very important source of meaning and satisfaction. These people lack the access to many resources and do not enjoy the socialising experience that work can be. For the working population that remains, jobs are becoming increasingly stressed (for reasons for this development, see 1.3.2). Time constraints seriously cut back some very important possibilities to share experiences and activities with relatives and friends. “Jobs”, such as doing the garden, preparing food, or fixing the car are more and more changing from important ties to “simple life” to frustrating obstacles in a tight time- management scheme. The family is one area that especially suffers from these developments. Here, a s o-called spill over effect of work stress is possible. In this case, a worker “brings home” the problems he experiences in his work. When both partners vent out their problems on each other, serious problems for the relationship can result (Dipboye, Smith and Howell, 1994). This process is especially striking in the case of dual career stress. As more and more couples both have ambitious career plans, it is becoming incredibly difficult to harmonise work and private life. When one partner is being promoted to another city, it is not that obvious anymore that the rest of the family will simply join the “breadwinner”. The full impact of the “modern dilemma” is not yet clear, but it could well be that we are facing one of the biggest challenges of the new millennium (Cooper, 1983).
Economic costs
In terms of money, behavioural, physiological and psychological outcomes of stress can become very costly. In the case of behavioural problems, stress can result in drops in sales because of strikes, damaged company property, etc. Moreover, stress-related illness results in costs for sickness leave and medical treatment. When stress results in psychopathological outcomes, such as depression, he will have to rely on a disability pension. Already, 34% of the current influx in the Netherlands is being caused by working stress (LISV (2000), telephonical account). This picture is basically the same for all industrialised countries. Macro- economically it is estimated that in the UK the costs of working stress and mental disorders are 15 billion a year. In the US the total amount of money spent on stress-related disorders is estimated on 100-150 billion US$, or even 300 billion when costs for sickness leave are included (Matteson and Ivancevich, 1987). This includes the costs for a so-called worker compensation system (see also 2.1), which amounts to 50-60 billion US$ per year. On the company level, the money that is spent on disability claims can range from 3 to 6% of the payroll, although in some industries, extremes of 40% exist (Schwarz, Watson, Galvin and Lipoff (1989) in Kohler Moran, Wolff and Green, 1995).
1.3 Stressors
1.3.1 Types of stressors
Now that the consequences of stress have been reviewed (1.2), the focus shall shift to the causes of stress problems. When looking again at the definition of Kahn and B., a stressor is defined as follows:
“A stimulus is defined as stressful (i.e., a stressor) because its effect is assumed, hypothesised and (cumulatively) demonstrated to have certain undesirable effects [...]” (p. 575)
As noted in the definition, a stressor is something that leads to very serious biological, psychological, behavioural and societal outcomes. Unanswered however, remains the question into the origins of stress. Already briefly mentioned as such factors are stressful life events: events with a very far-reaching impact for the person involved. However, recent research has shown that chronic stressors -and workplace stress is often chronic- can have a much higher impact on individual functioning than episodic “event stressors” (Hosman, 1998; Kessler, Price and Wortman, 1985). Work can be a very important facet of people’s life. It can be a major source of satisfaction, social interaction, confirmation and resources. By the same token, however, stressing work can pose a threat to these valuable experiences. Moreover,
work can posses many stressful circumstances. Because people spend a high proportion of their lifetime working, these stressors are extremely influential. In the following section, the most important sources of workplace stress shall be presented. They shall be grouped along categories used by Cary Cooper (1983) in her review of stressors at work.
Factors intrinsic to the job
To begin with, there seems to be very large differences in the amount of stress experienced by different professions. It has been found that occupations with a very high workload but relatively low decision latitude are especially troublesome (Gaillard, 1996). Jobs in the helping professions, such as police officers, nurses and social workers, are also considered as extra stressful. One of the reasons lies in the emotional nature of these roles: sometimes the emotional pressure is so high that the worker threatens to become exhausted or “burned out” (Gaillard, 1996). Besides high work load/low control and emotional exhaustion, six other job stressors can be distinguished (Cooper, 1983):
-Poor physical conditions, such as noise, fume and heat (Dipboye et al., 1994). These are mainly found in more traditional blue-collar jobs.
-Shift work. Working on irregular times is more tiring, whereas sleeping irregularly reduces sleep quality. Moreover, there often has to be worked under chronic sleep deficiency (Cooper, 1983), which is physically very demanding. Another common complaint from shift workers is that working and resting at unusual times produces the feeling of being “excluded from society”.
-Job over load. This happens when there is too much work to do, or when the job is too difficult. In any case, the worker feels overtaxed (Gaillard, 1996; Dipboye et al., 1994).
-Job underload, which occurs when the job is too repetitive or boring. Sometimes, as in the case of surveillance personnel, periods of boredom are suddenly disrupted by brief moments of fierce activity. These extreme contrasts are reported to be highly stressing.
-In some occupations, the risk of physical danger is especially high (Cooper, 1983). This is for example the case with fire fighters and police personnel, or with nurses working in a HIV clinic.
-A bad person/environment fit , which occurs when work does not match the personality and expectations of the worker.
Organisational role
A second class of stressors is related to a person’s role in the organisation. Here, the two stressors that have been most frequently studied are role conflict and role ambiguity. Role conflict happens when somebody has two or more roles that conflict with each other (e.g. mother and career woman). Role ambiguity occurs when role demands are unclear or unknown. Another role characteristic that can be sometimes perceived as stressful is responsibility (Cooper, 1983; Dipboye et al., 1994). A possible explanation is that responsibility is often paralleled by risks of failure.
Career development
Career development can serve as a stressor in the following instances: overpromotion, underpromotion, status incongruence (the impossibility of job advancement), lack of job security and thwarted ambition. Some of these career blockages are especially common among women managers, who often face a so called “glass ceiling”, the phenomenon that it is very difficult for women to reach jobs beyond certain levels in the organisational hierarchy, despite a “publicly” communicated message of equality (Brehm and Kassin, 1996).
Relationships at work
These can include problems with supervisors and co-workers. Besides being a stressor in itself, interpersonal problems also lead to ineffective communication, which further increases role ambiguity and other problems (Dipboye et al., 1994).
Organisational climate
An example of a stressful organisational climate is personnel policy (Kahn et Byosiere, 1992). Are job promotions guaranteed or do they have to be “conquered” through fierce internal competition? Similarly, inadequate minority practices can have a very stressful impact for women and people of other colour and religion. Another organisational feature that can produce strain is lack of participation in decision-making (Cooper, 1983; Jackson, 1983). In this case, the worker has too little influence over his job.
1.3.2 Is the presence of stressors on the work flour increasing?
From the previous section, it can be concluded that the work flour can give rise to a great variety of stressors. In recent times, there are indications that the presence of these stressors is increasing. Stress seems to be on the rise and it does not seem that it will disappear from economic life that easily. “Stress is here to stay”, as one author (Murphy, 1996) puts it. As shall be outlined in this section, increases in stressfulness primarily root in macro-economical developments. For a very long time, economic life was marked by the Industrial Revolution, which brought machines and factories. This development signalled an era of centralisation, specialisation and standardisation of capital and labour. Out of the first little entrepreneurs, huge corporations emerged. In the first phase of this development, there were many competitors fighting for a piece of the market. This resulted in the necessity to produce more effectively and profitable, and stimulated an immense growth of the production powers; not only by scientific innovations, but also by a different deployment of labour forces itself. Replacing handicraft, mass production reached a dominant position, and manufacturing became highly standardised and centralised. The joint result of these developments was that the emphasis came to rest on routine jobs,
i.e. jobs that required relatively little initiative or education from the worker. As a result of this, work became increasingly predictable. When could be spoken of stress, it was predominately physical, which is still the case in more underdeveloped regions of the world (De Gier, 1995).
On the verge of the new millennium, western economies changed dramatically. In his article in Job Stress Interventions E. de Gier (1995) describes how industry, due to the rise of new information- technologies such as the mobile phone and the Internet, increased its pace. The concept of mass production, which had reigned for almost a century, was no longer competitive on the world market. The higher pace of changes enhanced the need for so-called flexible specialisation, i.e. “a strategy of permanent innovation: accommodation to ceaseless change rather than an effort to control it (De Gier, 1995). While in the past, companies had often tried to spread risk and build up reserves, dynamics and challenge were once again being embraced. These changing demands asked for a “structural adjustment” of western economy, or a reappraisal of free market principles in the form of flexibilisation, deregulation, privatisation and adjustment of economic structures. (De Gier, 1995).
This structural adjustment had tremendous consequences for the quality of work. Following the ongoing flexibilisation process, the number of routine jobs went down in favour of person-to-person and symbolic- analytic services (Reich (1992) in de Gier, 1995) Moreover, a transformation of big companies into smaller networks of core organisations and suppliers took place. In these dynamic corporations, new production concepts arose. At least in some industries, modes like “just-in-time” and “lean” production more and more replaced the dominant position of traditional mass production. Of course this had a revolutionary impact on the labour process. Instead of the traditional assembly line, where every segment of the production process was chronologically done by individual workers, so-called autonomous working groups appeared. These working groups demanded new skills of workers, not only because they combined formerly fragmented tasks, but also while a substantial amount of social and communication skills is needed to perform effectively. All these economic changes had one other consequence, however. Accelerated work pace, increased responsibility and social interaction contributed to heightened unpredictability and thereby to a very serious by-product: stress.
The assertion that stress is increasing is backed up by numerous sources. In an American example, the percentage of stress-related cases from the Northwestern Mutual Life Insurance Company more than doubled from 1982 to 1988 (Murphy, 1996). Other evidence for the increase in stressfulness in the US comes from the skyrocket increases in the amount of money that is paid to workers that sue their companies for stress complaints. From 1985 to 1990, claim costs have more than doubled, placing an increasing burden on many companies (Kohler Moran et al., 1995). Another source are the recent figures
from the Netherlands. Here, it is estimated that 34,23% of the new influx comes into the occupational disability pension from psychological grounds. The total amount of occupationally disabled, however, is 31,17% (LISV Amsterdam (2000), telephonical account). Because people that become occupationally disabled on psychological grounds still comprise a larger percentage of the influx than of the total amount of cases, this is indicative of an increase in stress, since a high proportion of these cases is stress related. Figures from the German Federal Republic point in the same direction. Here, the total number of occupational sickness cases dropped with 1.2 percent from 1998 to 1991. In the same period, the percentages of cases due to psychiatric conditions rose with 0.7 percent, however. When the number of sickness day is considered, the contrast becomes even more striking. Whereas the total number of sickness days went down with 2.3 percent, missed days because of a psychiatric disease increased 2.0% (Zoike, 1999). Figures such as these are clearly supportive of the conclusion that stress in the modern economy is indeed on the increase.
1.4 Mediator and moderator variables
Introduction
The first epidemiological studies into the relationship between stress and psychopathology were focused on so-called life events. Initially, very high correlations with psychopathology were found. However, although the influence of stressful life events is still being estimated on 9%, later research failed to confirm these high correlations. Apparently, the link between stressful life events and illness is not a mono-causal one as researchers had originally assumed. Whether someone develops an illness as a result of a life event turned out to be linked to mediator and moderating variables. And indeed: when social support and coping skills are added as moderating respectively mediating factor, the predictive power of life events increased dramatically (Kessler, Price and Wortman, 1985.) The same is true for working stress: the impact of a stressor cannot be explained without accounting for intervening constructs and variables. In the following section, the most important of these factors shall be elaborated.
Coping
The central role in these intermediate processes is played by the coping construct. A key figure in the development of this construct is Richard Lazarus, who outlined the process of cognitive appraisal, an intra-psychic process that translates objective events into perceptual experiences. Lazarus (1995) divides cognitive appraisal into two steps. The taxation of the amount of threat in a situation is labelled primary appraisal. The evaluation of ones response possibilities as either adequate or insufficient is termed secondary appraisal. After the appraisal phase, individual will develop some way to deal with the stressor. This way of reacting to these external threats Lazarus calls coping, which can be distinguished along two dimensions: problem focused vs. emotion focused and active vs. passive coping. Combined, these dimensions give rise to the following four categories: active problem coping (e.g. trying to alter the stressor), passive problem coping (e.g. trying to ignore the problem), active emotion coping (e.g. trying to talk with friends about difficulties) and passive emotion coping (e.g. taking drugs to “escape” from misery). Certainly, each form of coping can have some adaptive value in certain situations, and how much success a certain way of coping has is dependent on the nature of the problem. Nevertheless, when faced with solvable problems, it is in general thought better to employ active and problem-oriented coping styles.
Which coping mechanism a person chooses is not solely dependent on the amount of success this way of reacting would have, however. The choice of a coping strategy is also determined by the personality and social support of the worker (Gaillard 1996). The influence of a person’s social network on his reaction to stress has been the topic of much research. Here, research has conformed that this is a highly powerful determinant (e.g. Kessler, Price and Wortman, 1985; Kahn and Byosiere, 1992; Ogden, 1996). Apparently, having friends or colleagues, who can act as role models and who are available for advice and discussing problems, is very important.
Personality factors
The second factor that influences a person’s cope style is formed by personality factors. Although the list of personality factors that have an influence on stress processes is in principle endless, the following ones are widely accepted and have been the topic of considerable research (Kahn and Byosiere, 1992).
- Locus of control
This refers to the extend in which a person places the impetus for events within himself or onto external influences. Whether a person is internally or externally oriented thus influences the appraisal process, where people have to judge whether an event is uncontrollable or to be managed.
-Type A personality
People with this personality structure are characterised by excessive competitiveness, impatience, hostility and vigorous speech (Ogden, 1996). Almost by nature, these people will opt for active coping styles. Although such an active style is usually adaptive, there are important exceptions. It is for example not very useful to get overly excited when stuck in a traffic jam: it won’t change anything. This is one of the reasons why type-A is an important risk factor for heart problems (Ogden, 1996).
-Hardines s (Ogden, 1996)
Individuals with this personality structure share personal characteristics of a) personal feelings of control, b) a desire to accept challenges and c) commitment. Hardiness could be involved in the primary appraisal phase of Lazarus coping model. Here, “hardy” people are more apt to see stressing circumstances as challenges and under their control. This can lead to less stress reactions.
Summary
In sum, the relationship between stressors and strains has been shown to be a complex process, where the presence of stressor A cannot always predict the development of symptom B. Whether this actually happens depends on how people appraise the stressor and how they eventually cope with it. The outcome of the appraisal process and the choice of a coping strategy are in turn dependent on a person’s personality and social support. People with a good social network have been shown to deal with stress better. With regards to personality, there is evidence that the factors of locus of control, type A personality structure and hardiness have an influence on the stressor-strain relationship.
1.5 An integrative model of the stress process
1.5.1 Introduction
As was stated above, the aim of this last section is to integrate the previously covered research into a unifying model. Because stress is a very complex phenomenon, a model is needed that captures this complexity, as well as clarifies the variables and the relationships of the depicted process. The choice of such a model is not an easy task, however. There exist quite a variety of approaches and some of them fit large aspects of the data quite well. However, most of these models seem to neglect important aspects. For example, the so-called Michigan model (Gaillard, 1986) highlights the objective and the subjective environment, as well as outlining the role of personality and social support. However, it is not made very clear by what processes the objective environment leads to a subjective representation. This process is more clearly conceptualised by Lazarus (Gaillard, 1996), who distinguishes between primary and secondary appraisal and different ways of coping. However, in the latter’s model, the role of social support and personality factors are not clearly distinguished.
Because, until now, single models that depict the stress process in all its important aspects are rarely to never found, an integrative approach was used. Such integration has the important advantage that it reduces the amount of “blind spots” that each theoretical model is doomed to have. The model that was chosen for his purpose is Hosman’s IS-model and is presented in the following section. The IS-model integrates numerous other theoretical orientations, but its cornerstone is Lazarus’ model of stress and
coping, which emphasises the role of the subjective processing of objective environmental characteristics. As shall be seen below, the model goes beyond this, however. It specifies constructs more clearly, builds in feedback loops and accounts for multiple system levels (Hosman, 1998). To cover the model in this chapter, a description shall first be given. Within this description, links shall be made to the evidence that was reviewed in the previous chapter. This description is best read with the model itself at hand, since at first sight, it has a fairly complex structure. At the end of the chapter, when the model has been presented and explained, it is hoped that the reader has a clearer understanding of the role of each of the reviewed aspects of the stress process. Also, with a comprehension of the structure and the relationships within the stress process, it shall be easier to understand the rationale behind the study’s hypotheses (which will be presented in chapter 2).
1.5.2 Description of the model
The integrative stress theoretical model describes a stepwise process through which stressors in the environment eventually result in adverse or pathological outcomes. In this model, many of the influences are reciprocal, in that one factor exerts an influence on the other and vice versa. For example, goals and needs have an influence on the perception of a stressor, but this goes the other way around too. Furthermore, the IS -model accounts for multiple system levels. This is depicted by the “onion shells” of the meso and macro environment. Thus, it should be kept in mind that the stress process, with all its phases and elements, is also influenced by a person’s different contexts and vice versa. In the section below, the IS-model shall be explained in detail by reviewing each of its individual steps: (1) appearance of stressors/compensating factors, (2) perception, interpretation and valuation of the stressor, (3) problem experience (occurrence of crisis), (4) anticipation of problem reaction, (5) performance of problem reaction and (6) consequences of this reaction (feedback loop).
Appearance of a stressors and protecting factors
The process depicted in the model begins with the existence of certain stressors in the environment. Indeed, this is the same basic notion as in the general “stressor-strain” concept that was reviewed previously. The most important kinds of stressors, such as understimulation and threat of unemployment, were already covered in section 1.3. The influence of stressors, however, is buffered by the availability of protective factors. These are resources in the environment that are beneficial to people’s health and functioning. On this topic, not much research has yet appeared. For this reason, only an example of a protective factor can be listed, without wanting to claim that there are no others. This example is physical work, which was mentioned by Kahn and Byosiere (1992). The two factors of risk and protective factors together constitute the environmental input of the IS-model. After the occurrence of these risk and protective factors, the model continues with the subject that reacts to these influences.
Perception of the environment
The core process of this step is that a person perceives the environmental factors from step 1 and judges whether they form a threat to important resources. This process is basically similar to Lazarus’ conception of primary appraisal that was covered in the previous section (1.4). In the model, it is dependent on a number of factors whether someone perceives his environment as threatening or not. Firstly, this phase is influenced by a person’s goals and needs. Thus, an event such as a major company reorganisation will be perceived as stressful because it threatens to take away highly valued resources of money, prestige and satisfaction. This process is reciprocal, because the threat of losing important resources also triggers the goals and needs that are at stake, making them more salient (Brehm and Kassin, 1996). Secondly, the amount of threat an individual perceives will depend on his appraisal of how difficult he expects it to be to overcome the environmental threat. This will vary according to his problem solving skills because if the person possesses a very powerful and broad range of coping skills, fewer situations will be perceived as stressful. This description is very similar to Lazarus’ secondary appraisal construct, in which a person judges to what degree his coping resources are sufficient to solve a problem. Again, this process goes the other way around too because perc eption has an influence on a person’s perceived coping skills. For example, if the situation is highly dangerous and overwhelming, the person might lose his confidence in his available coping resources. A third moderator of the appraisal and coping process is the expected social support. If this is available, stressors become less threatening because the social system can help a person cope with it. Fourthly and finally, the process of perception is influenced by the subsequent phase of crisis perception. If the individual already
sees the situation as highly threatening, that will lead to a focusing on the stressors and other kinds of perceptual biases. This will further strengthen the appraisal of the environment as stressing.
Crisis experience
In the case of a sufficiently serious stressor, the result of the previous step is an attribution of the situation as dangerous and a threat to important resources such as money. This leads to an emotional and physical “alarm” (Hosman, 1998) that arouses the person to invest great energy in solving the problem. Such an alarm state bears close resemblance to Cannon’s “fight/flight” concept, which was introduced at the beginning of the first chapter. Also, this agitated condition calls to mind Selye’s general adaptation syndrome, which included arousal symptoms such as an enlargement of the adrenal (arousal hormone) gland cortex. In case of a crisis, this emotional reaction can be so fierce that the problem overwhelms the person. This process is influenced by a number of factors. The influences of this crisis phase follow more or less the same lines as the previous step, as crisis experience is influenced by problem solving skills and social support. Furthermore, there exist a two-way relationship with the preparatory reaction. If the person prepares a reaction of which he anticipates an insufficient result, the experience of crisis will only be sharpened.
Preparatory reaction
In this process, the individual selects and anticipates a way of responding to the environmental threat. In other words, the person prepares the way by which he is going to cope with the problem. Here, he can choose between different coping types. From different possibilities, the person selects the way he is going to react to the experienced crisis. This resulting process is influenced by the problem solving skills of the person. There might exist a preference for certain styles, or some forms might be more sophisticated and trained. This process is reciprocal, however, because preparing and using a coping skill will also affect competency with this skill. The preparatory reaction is further influenced by the available social support because certain modes of coping can only be successfully deployed if a stimulating environment supports them.
Stress response
In this phase, the person performs the selected response from step 4. As stated above, there are many different ways with which a person can cope with a stressful situation. Coping is essentially an effort to adapt to the stressing environment. The result of this adaptation determines whether an individual succeeds in preventing future problems or becomes involved in a feedback loop of distress and adversity (see below). The exact success of the way a person copes with his environment depends on the particular characteristics of the situation. There are, however, a number of factors that play an important role. In the model, the way this response is carried out is influenced by the person’s social support. People can attract advice and resources from their social environment, which will make efforts to cope more effective. Also, the eventual effect of a response depends on how competent the person is in performing this response. This is an important reason why many stress interventions include some kind of (coping) skills training.
Consequences of the reaction
In it’s final phase, the IS-model contains a feedback mechanism that loops back the results of the stress response to the person himself and the other factors in the model (depicted by the block arrows). In this chapter (1.2), many types of such consequences have been reviewed. One of these categories is constituted by physiological health problems, which can take the form of cardiovascular symptoms, biochemical outcomes and gastrointestinal problems. Another type of outcomes, which is of particular interest to the current writing, is mental health problems. When a person fails to cope with a threatening stressor, psychological illness such as depression and anxiety might be the result. Finally, besides accounting for such individual consequences, the model feeds back to the environment as well. Effects can include financial costs to companies and governments and social costs to families and society as a whole (see also 1.2.4). In the figure, these outcomes are specified by special boxes with the header “CONSEQUENCES”. These boxes should be regarded as illustrative and are not meant as specific elements or phases of the model.
Other factors that receive feedback from the model are goals and needs, problem solving skills, the social support system and the existing risk and protective factors. For example, the stress response of alcoholism can alienate a person from his environment so that his social support system weakens.
Furthermore, his coping skills will also be reduced because alcoholism makes the use of certain ways of coping (such as active problem solving coping) less probable. A coping response can also result in actually altering the stressing environment. Through coping, a person can decrease stressors and increase the presence of protecting factors. A person might, say, complain about certain harmful aspects of his working environment. When successful, such actions lead to a reduction in the amount of stress.
Chapter 2: Stress Interventions
2.1 Introduction
In the previous chapter, it was pointed out that stress can lead to numerous adverse consequences that manifest themselves on the level of physiology, psychology and behaviour. Moreover, these individual outcomes affect broader system levels such as society, family life and economy. As noted in the introduction, there are a lot of disciplines that are confronted with these stress-associated problems. They are also faced with the problem of reducing these problems. “Doing something about stress” is not as simple as it seems, though. What exactly needs to be done? On first thought, the answer seems simple: when someone is experiencing high blood pressure as a result of stress, try to lower the blood pressure (e.g. by means of anti-hypertensive medication). This strategy evaluates a problem in terms of its symptoms and sees its solution in reducing these symptoms. An important argument for this approach is that often the underlying causes of a problem are not known, so reducing symptoms is the only possibility. Moreover, some underlying causes, such as genetic dispositions, are simply not (yet) amenable to change. Finally, in some instances where they could be changed, this is not wanted because underlying mechanisms serve highly valued other purposes. It might for example be very stress reducing to abolish the use of deadlines in a newspaper agency, but when this would result in a reduction of quality, it is not likely to be implemented.
Besides this symptom approach, there is another possibility. This other approach tends to see symptoms as a part of a deeper underlying problem. In this view, symptoms are caused by pathogenic factors, so in order to reduce the symptoms one should eliminate these factors. Although the value of the arguments for the symptom-oriented side are recognised, this essay starts from the second perspective. It does so for a number of reasons. Firstly, focusing on symptoms cannot completely eliminate a disease. Because the underlying causes are still present, the curing of symptoms has to be repeated over and over again, whereas taking away the causes would solve the problem once and for all. This is the more attractive because, as noted in chapter 1, curative care is socially and economically very costly. A second reason for the use of preventive interventions is that it is unethical to let people suffer the symptoms of a condition when intervening in its causes could have prevented this. Thirdly, a symptom-oriented perspective focuses mainly on the individual. This ignores the fact that many conditions (and not in the last place stress) have their roots in socio-economic circumstances. These factors are out of the scope of curative interventions, but are possibly changed by preventive efforts.
Besides the above-described arguments for the preventionist position, there exist legal incentives too. In Europe, the European Union has implemented laws that have highlighted the need for healthier work environments. In 1989, the Union issued the “framework directive on health and safety at work”, which became to resemble a kind of European work environment act. It included quite specific employer directives regarding the quality of the job content, well being of workers at work, medical examination, training, information and worker participation. The consequence of these legal developments was that many member states implemented penalties to make sure that companies would guarantee a healthy working environment in line with the EU directives. In the Netherlands, for example, the so-called “Pemba law” obliges enterprises to directly pay for each worker that becomes occupationally disabled. In the US, too, industries are facing financial responsibilities for their occupationally disabled, although the legal situation there is markedly different. Through a workers compensation system, employees have the option to sue their employers for missed income due to occupational disability (Kohler Moran et al., 1995). Compared to a Social Security pension, very high sums can be the result. In both the US and the EU, therefore, the legal situation causes a shift in focus, so that companies are becoming more and more responsible for their stress problems.
2.2 Characteristics of prevention programs
In categorising a preventive intervention, a number of distinctions can be made. Following a distinction by Hosman (1998), the following 8 dimensions can be distinguished: (1) type of actor, (2) target group, (3) goals, (4) determinants, (5) instruments, (6) system level, (7) time and (8) setting. These dimensions are not only important in describing and categorising different interventions but also in answering the question of which aspects of an intervention are crucial for its effectiveness. For this latter purpose, an analysis is needed that links program characteristics to effect. To prepare this, however, such characteristics shall first have to be coded using a classification system in which many of the here described distinctions are reflected. In this chapter, each of the 8 research dimensions shall be elaborated in more detail:
2.2.1 Actor/program provider
The dimension of actors pertains to the people that provide the stress intervention. Because stress can be a very costly economic factor, it are often organisations that initiate a stress intervention. Sometimes, the manager of the human resources department notices problems and decides to take action (e.g. in the study of Malkinson, Kashmir and Weinberg, 1997). However, besides the Personnel Development office, it are often (mental) health professionals or external consultants that provide the intervention.
2.2.2 Target group
This pertains to the question on what kind of group the intervention shall focus. Regarding this factor, a number of aspects are important. One distinction, which also returns in the coding system, is the difference between ultimate and intermediate target group. Here, the group in which the intervention is carried out in order to prevent future problems is called the ultimate target group. In some cases, however, programs use a so-called intermediate target group. In this case, the program uses an additional group that has an important influence on the “problem group”. In the case of organisations, for example, stress is often caused by organisational constraints, such as unsupportive supervisors. Because the problem group might not be able to alter these constraints, interventions sometimes target other groups that do have this possibility. In the example of the unsupportive supervisors, they could be given a sensitivity training. Here, the workers that are stressed are the ultimate target group, whereas the unsupportive supervisors constitute the intermediate target group.
Another important factor is whether the target group is characterised by a high amount of risk. It is quite common that intervention programs are targeted at high-risk groups. In the present coding system, this was the case when the target group possesses a lot of risk factors3, or when there are already many stress related cases4. When the program was targeted at a population for which no higher risk was known or stated, this was coded as normal risk5 (see appendix 3 for an operationalisation of this variable). The question of high-risk is important because this affects some very important strategic decisions. Budgets for prevention are limited, so it would be very valuable information whether it is more efficient to spend the efforts on a smaller group of higher risk people or to spread the intervention over a larger population. A hypothesis that shall be researched is that programs aimed at high-risk groups are more effective than programs that do not. There are a number of reasons why this could be the case. Firstly, people that are at high-risk for a problem could well be more motiva ted to participate in an intervention. Many authors, such as Bunce (1997), suggest that programs are more effective when they have more motivated participants. Another reason why high-risk programs could be more effective is related to the so-called span of change. People that are at risk for a problem will often be characterised
by a greater number of risk factors. According to the span of change concept, such people therefore have more room for improvement.
[...]
1 For a definition see Bunce (1997): “A mediator is “the generative mechanism through which the focal independent variable is able to influence the dependent variable of interest.”
2 A moderator is a “variable that affects the direction and/or strength of the relation between independent or predictor variable and a dependent or criterion variable.” (Bunce, 1997)
3 Selective preventive intervention.
4 Indicative preventive intervention.
5 Universal preventive intervention.