Stress Management Interventions: Improving Subjective Psychological Well-Being in the Workplace

           When reviewing the literature, we use a broad definition of subjective psychological well-being which enables us to include studies on burnout, anxiety, and depression, for example, as well as studies focusing directly on positive psychological well-being. Furthermore, as we are interested in psychological well-being we do not include studies that have looked at physical health or job attitudes, or those on well-being related outcomes at the organizational-level such as sickness and absenteeism. However, it is worth noting that these outcomes are related such that improvements in psychological well-being can lead to improvements in health, job satisfaction and absenteeism rates.

Stress Management Interventions: A Typology and Description

           The stress management literature typically classifies interventions according to the ‘focus’ of stress management and according to the ‘level’ at which the intervention takes place (DeFrank & Cooper, 1987; de Jonge & Dollard, 2002).  With regard to the focus of stress management, interventions are categorized as primary, secondary or tertiary.  The aim of primary interventions is to prevent stress from occurring by removing the sources of stress and enhancing the causes of well-being.   Secondary interventions aim to reduce the severity or duration of stress once it has occurred and to prevent the level of stress becoming problematic.  Tertiary interventions seek to rehabilitate and maximize functioning for those who are already experiencing or suffering from psychological ill-health. 

           With regard to the level of an intervention, a common and simple distinction is between individual and organizational levels.  As the name suggests, individual-level interventions focus on helping employees to develop skills to manage, cope with and reduce stress, whereas organizational-level interventions make more systemic changes to organizational practices that either target all employees or a specific group of workers.  A third category, individual-organizational level interventions, is used in some classifications.  Such interventions are thought to differ from others in that they focus on changing the relationship between the individual and organization, e.g., peer support groups.  However, as the distinction between organizational and individual-organizational level interventions is not always clear cut, we use the more parsimonious categorization of individual-level and organizational-level interventions. 

           Classifying stress management interventions (SMI) according to their focus and level implies that both individual and organizational-level interventions can be primary, secondary and tertiary in nature.  This is illustrated in Table 1, along with examples of the different types of intervention in each category. Throughout the remainder of the chapter we describe the different types of SMI according to this classification.

Table 1. A Typology of Stress Management Interventions

Individual-Level Interventions

           The aim of a primary individual-level intervention is to prevent stress from occurring in an employee.  One means of achieving this is through selection and assessment procedures that select applicants who have the skills and abilities to manage the demands of the job and to screen out those who might be susceptible to experiencing stress in the target role, particularly in highly-stressful occupations (Bartone, Roland, Picano & Williams, 2008).  Although such interventions are one way of managing stress and promoting well-being, they are rarely used (Giga, Cooper & Faragher, 2003).  

           Secondary individual-level interventions seek to equip employees with the skills and abilities to manage stress and promote well-being, and to provide employees with opportunities to engage in stress-reducing activities.  They include techniques such as relaxation, meditation, cognitive behavioural therapy, mindfulness training, and exercise programmes, as well as other techniques such as education and interpersonal skill development.  Drawing on the emotion regulation literature (Gross, 1998), these techniques can be understood as promoting antecedent-focused emotion regulation strategies that seek to reduce or remove the causes of stress, or response-focused emotion regulation strategies that seek to reduce the level of stress experienced by individuals.  It can also be noted that some individual-level interventions are multimodal, e.g., a combination of relaxation, CBT and mindfulness exercises.  Such an approach might be used in the expectation that it will increase the likelihood of beneficial outcomes for both the individual and organization, as it can enable employees to develop a wide set of skills that can be used across different circumstances, provide employees with the opportunity to develop both antecedent and response-focused emotion regulation strategies, and increase the chance of meeting employees’ needs.  As secondary individual interventions are one of the more commonly used SMI, and there is a stronger evidence base for their effectiveness (which we outline later), we now detail the more widely used techniques. 

           Relaxation techniques are based on the assumption that states of relaxation and stress (i.e., highly-aroused negative states of well-being such as anxiety, tension and anger) are antithetical (Russell, 1979). In other words, you cannot be both relaxed and stressed at the same time, such that increasing levels of relaxation will necessarily entail lower levels of stress.  Various methods are used to induce relaxation but all can be considered as response-focused emotion regulation strategies as they focus on reducing the symptoms of stress. Progressive muscle relaxation involves tensing and then relaxing different muscles throughout the body in a prescribed order.  For example, a person might start by clenching and unclenching one hand then the other, then focus on the muscles in one forearm and then the other, and so on.  The length, duration and range of muscles covered can vary but with practice employees can induce a deep state of muscle relaxation in minutes (Murphy, 2003).  Meditation is another relaxation technique but one that focuses on a mental rather than physical process.  Techniques typically involve sitting in a quiet place and repeating a word or sound while maintaining a passive mental state that excludes intrusive thoughts.  Although meditation has its roots in religious practice, secular and standardized methods appropriate for clinical trials have been developed (Carrington et al., 1980).  Muscle relaxation and meditation can also involve breathing techniques that focus on the inhalation and exhalation of breath as an additional means of achieving relaxation, although some interventions use breathing techniques alone. 

           Cognitive behavioral therapies (CBT) assume that maladaptive cognitions contribute to the maintenance of psychological distress and problematic behavioral responses to stress.  CBT addresses this by helping the person to identify misconceptions about the nature and causes of stress, to test the validity of existing thoughts and understandings, and to develop new conceptions about stress. In addition, the behavioral element of CBT encourages the person to develop new behavioral responses to stressful events. CBT is thought to work by promoting antecedent-focused emotion strategies that help the person to reappraise and restructure their understanding of stress and stressful events (Hofmann & Asmundson, 2008).  For example, Cecil and Forman (1990) conducted an intervention amongst teachers based on a method of CBT called stress inoculation training (Meichenbaum, 1977) that consists of three phases: education about stress and its causes; acquisition and rehearsal of skills in cognitive restructuring (e.g., identifying irrational thoughts and replacing them with rational thoughts); and developing and practising new cognitive, emotional and behavioral ‘scripts’ for responding to stressful events.  Their evaluation revealed that the CBT intervention decreased stress to a greater extent than a peer support group and a passive listening group. 

           Mindfulness training is an increasingly popular SMI that aims to promote states of mindfulness (i.e., paying attention to the experiences, thoughts and emotions occurring in the present moment in a non-judgmental, compassionate, accepting and non-reactive way) and new adaptive responses to negative thoughts and emotions (Kuyken et al., 2010).  Mindfulness is thought to foster psychological well-being by primarily helping the person to dissociate negative thoughts and emotions from maladaptive behavioral and emotional responses.  Mindfulness training can therefore be understood as a method that develops response-focused emotion regulation strategies.  But mindfulness training has key differences from other methods that also encourage response-focused emotion regulation strategies, such as relaxation and meditation.  In particular, in mindfulness training, negative thoughts and experiences are not avoided or suppressed (which can have negative consequences for well-being, Richards & Gross, 1999) and there is an emphasis on developing new and more adaptive responses.

           Mindfulness therapies include acceptance and commitment therapy (Hayes, Luoma, Bond, Masuda, & Lillis, 2004) and mindfulness-based cognitive therapy (MBCT) (Segal, Williams & Teasdale, 2002).  MBCT, for example, uses a range of methods, often conducted in groups with a trainer, such as body scanning, yoga and meditation exercises. The initial emphasis in MBCT is cultivating states of mindfulness.  At later stages, the emphasis moves towards identifying how negative thoughts trigger maladaptive responses, helping the person to accept those feelings, and encouraging the choice of more adaptive responses.  Although mindfulness training has grown in popularity, few empirical studies have been conducted in organizational settings.  An exception is a study by Hülsheger and colleagues who conducted a two-week, self-directed mindfulness intervention, based on MBCT protocols, with employees from a range of occupations.  Their evaluation, using data collected from daily diaries completed by participants, showed that employees in the mindfulness training condition reported higher levels of daily mindfulness and lower levels of daily emotional exhaustion that those in a control group (Hülsheger, Alberts, Feinholdt & Lang, 2013).

           Other types of secondary individual-level interventions have also been developed. Two popular initiatives include educational programmes that inform employees about stress and how it can be managed, and health promotion initiatives that are based on the assumption that a healthier lifestyle (e.g., more exercise, better diet) will improve well-being (Anger et al., 2015; Rongen, Robroek, van Lenthe, & Burdorf, 2013). In addition, personal skill development training in communication, goal setting and time management all aim to reduce stress by helping employees to prevent stressful situations from occurring, e.g., conflict, high workloads.  

           Lastly, tertiary individual-level interventions focus on individuals who are experiencing high or chronic levels of stress that may be impairing their ability to work.  For example, employee assistance programmes (EAP) provide counselling and advice to those experiencing high-levels of stress or mental health problems, whether these issues are related to work or not (Bhagat, Steverson & Segovis, 2007; Csiernik, 2011; McLeod, 2008).

Organizational-Level Interventions

           Primary organizational-level interventions aim to remove the causes of stress in organizations by changing organizational practices and policies, such as those concerned with leadership, working time, and occupational health and safety (Anger et al., 2015).  The majority of primary organizational-level interventions reported in the literature are job redesign interventions that aim to modify job characteristics (e.g., job discretion, workload, ergonomic design) as a means of enhancing employee well-being, and they have a strong theoretical and empirical grounding in studies of job design that show job characteristics to be key antecedents of employee stress and well-being (Humphrey, Nahrgang & Morgeson, 2007; Demerouti, Bakker, Nachreiner & Schaufeli, 2001).  Job redesign interventions typically differ in scale, with some seeking to change one job characteristic, such as job discretion (Bond & Bunce, 2001), whereas others seek to change multiple job characteristics in the expectation that this will produce larger changes in well-being (Holman & Axtell, 2016). 

           Secondary organizational-level interventions aim to make organizational-wide changes that help employees cope better with stressful experiences.  Such interventions include the introduction of peer support groups that enable employees to discuss the difficulties they face (Peterson, Bergstrom, Samuelsson, Asberg, & Nygren, 2008) and communication skills training courses that seek to improve the ability of all employees to manage stressful situations through, for instance, improved conflict management (Ghazavi, Lohrasbi, & Mehrabi, 2010; Leiter, Laschinger, Day & Oore, 2011). However, in practice, there is not always a clear distinction between primary and secondary organizational level interventions, particularly with regard to social support interventions which can remove a potential source of stress (i.e., a lack of social support) whilst also improving the ability of employees to cope with stressful events. 

           Implementing organizational-level interventions can be difficult and complex.  This means that an important part of organizational-level interventions, particularly primary interventions that seek to change organizational practice, is the efficacy of the implementation process.  In particular, reviews suggest that successful organizational-level interventions involve four key activities: preparation, i.e., securing support; screening, i.e., identifying the psychosocial risks; action planning, i.e., developing change initiatives; and implementation, i.e., embedding change initiatives within the organization (Nielsen, Randall, Holten & González, 2010).  Another important characteristic of organizational-level interventions is the level of employee participation.  For example, in a job redesign intervention conducted by Morgeson et al. (2006), only managers and external consultants were involved in developing and implementing job redesign changes.  In contrast, in a job redesign intervention among knowledge workers in Denmark (Sørensen & Holman, 2014), employees participated throughout the intervention process by identifying, developing and implementing job redesign initiatives to reduce workload and modify work procedures.  Employee participation is thought to be advantageous as it can improve the quality of change initiatives by drawing on employees’ expertise to make them more contextually appropriate, and because it can increase commitment to implementing change initiatives, as employees have a greater sense of ownership of those change initiatives (LaMontagne, Keegel, Louie, Ostry, & Landsbergis, 2007).  Participation can also increase the sense of job control and responsibility, which is beneficial for well-being in its own right (Mikkelsen & Saksvik, 1998).  But participation is not without risk: it can increase complexity by involving a wider range of stakeholders who may have competing ideas and motivations, increase frustration with the intervention process if employees are not experienced or skilled in participation, and raise costs by removing front-line employees from their jobs. 

           Some multimodal SMI combine individual- and organizational-level interventions. For example, ‘total worker health interventions’ integrate organizational-level health and safety programmes with individual level health promotion programmes (see Anger et al., 2015 for a review). An advantage of this approach is that the organizational-level intervention can reduce the causes of stress, while the individual-level intervention can enhance employees’ ability to deal with stress.  Such an approach can be particularly important for employees exposed to higher demands, as they also tend to engage in more high-risk health behaviors (Anger et al., 2015).  Other multimodal interventions occur just at the organizational level.  For example, job redesign interventions have been implemented alongside changes to other organizational practices, in the expectation that this will augment the effects of the job redesign intervention (Daniels, Gedikli, Watson, Semkina, & Vaughn, 2017).  Potential disadvantages of multimodal approaches are an increase in implementation complexity and that one intervention may reduce the effectiveness of the other.

What Works?  Evaluating the Effectiveness of

Stress Management Interventions

           When selecting SMI it is clearly important to consider the evidence for their effectiveness in reducing stress and promoting well-being.  In this section, we review the findings for those SMI that have a more substantial evidence base, namely secondary individual-level and primary organizational-level interventions².  But before we do this, it is important to touch on the issues concerned with how best to evaluate interventions, as this provides a better understanding of the nature and quality of evidence available. 

           Based on the natural science paradigm, randomised control trial (RCT) designs are traditionally considered to be the optimal methodology for evaluating interventions.  RCTs randomly allocate participants to experimental and control groups and then introduce the intervention to the experimental group.  Change in the focal outcome is measured in both groups before and after the intervention.  Randomization is used to control for the effects of context and participant differences, such that when change in the outcome in the experimental group is significantly different to that in the control group, the change is assumed to be caused by the intervention.  RCTs are therefore thought to provide stronger evidence for the causal effects of an intervention than methodological designs that do not randomize individuals to groups, that do not have a control group, and those that do not have pre-intervention measurement. 

           In addition to the evaluation of individual studies, meta-analyses are used to provide a quantitative overview of multiple studies.  A meta-analysis produces an overall effect size based on the individual studies selected for review. More confidence can be placed in meta-analytic results, as they are based on a larger sample that minimizes the likelihood of chance findings, and because the weaknesses of individual studies are balanced out.  To explain this further, the results of one study could be accidental or influenced by something that was not investigated, whereas the results of a meta-analysis that combine data from multiple studies are much more likely to be a true reflection of an intervention’s real impact.  A key issue in a meta-analysis is how to select studies for inclusion (e.g., RCT studies only), since the methodological quality of the studies chosen affects the confidence that can be placed in the findings. 

           However, several authors have questioned whether RCTs provide a sufficient and appropriate methodology for intervention evaluation (Cox, Karanika, Griffiths & Houdmont, 2007; Nielsen & Miraglia, 2017).  Specifically, it is argued that RCTs are not always feasible in organizational settings where, for example, randomisation is not practical or ethical, or when an equivalent control group (consisting of similar participants to the experimental group) is not available.  Randomization can also be considered a change in itself that may affect employee perceptions of the intervention and experiences of work (Nabe-Nielsen et al., 2015).  Furthermore, RCTs tend to focus on changes in outcomes and do not evaluate how the quality of the implementation process or organizational context shapes the effects of an intervention.  In response to these concerns, alternative intervention evaluation frameworks have been proposed.   Cox et al. (2007) suggested a framework that covers both intervention outcomes and implementation processes.  Examples of implementation process variables include managers’ advocacy of the intervention, workers’ engagement with the intervention, and the availability of resources to support implementation. Cox et al. argued that consideration of process variables enables identification of the mechanisms determining intervention outcomes. Indeed, failing to consider process variables could lead to the erroneous conclusion that an intervention is ineffective when in fact the implementation was flawed.  Nielsen and Abildgaard (2013) built on these insights to propose a longitudinal evaluation framework in which each of the intervention ‘stages’ (e.g., preparation, screening etc., as outlined earlier) are assessed separately to identify how each stage may influence subsequent stages and ultimately the outcome of the intervention.  They also suggested that a range of intervention outcomes (e.g., changes in attitudes, working practices and well-being) should be addressed at each stage  on the basis that these changes may become evident at different times.

           Given the importance of intervention processes to the success of an intervention, and the limits of RCTs in organizational settings, it has been argued that assessments of the effectiveness of SMI should draw on systematic quantitative and qualitative evaluations and to include rigorous evaluations of SMI that do not meet the ‘gold standard’ of RCTs. Such an approach is thought to be particularly important in the evaluation of organizational-level SMI because, in comparison with individual SMI that tend to use existing techniques proven in other domains (e.g., CBT), organizational SMI are diverse in design and may be tailored to the organization, increasing the likelihood of outcomes being influenced by process and context variables. Indeed, in light of these issues, process evaluations are now more evident in SMI studies. 

Evidence for Secondary Individual-Level Interventions

           A number of meta-analyses and systematic reviews of the effects of secondary individual-level interventions on stress and well-being have been conducted (see Bhui, Dinos, Stansfeld & White, 2012, for a review of these reviews).  Two of the more comprehensive meta-analyses are by van der Klink, Blonk, Schene and Van Dijk (2001) and Richardson and Rothstein (2008) (Table 2).  Both cover CBT and relaxation interventions across all types of employees.  The Richardson and Rothstein meta-analysis builds on the van der Klink et al. study by covering an extra ten years of research and including unpublished studies (accounting for publication bias towards significant findings) but has stricter inclusion criteria, as it only selects RCT-based studies.  A wide range of individual interventions are also covered by Kuoppala, Lamminpää and Husman  (2008) but their analysis focuses on ‘psychological’ (i.e., education, biofeedback) and health promotion interventions (e.g., exercise), while the meta-analysis by Conn et al. (2009) focuses exclusively on health promotion interventions, specifically exercise and physical activity.  Other meta-analyses focus on SMI in particular groups of workers.  Ruotsalainen, Verbeek, Mariné and Serra’s study (2015) covers CBT and relaxation methods in health care employees, while that by Regehr, Glancy, Pitts and LeBlanc (2014) focuses on individual-level interventions in physicians.

           A conclusion that can be drawn from these meta-analyses is that that there is good evidence that secondary individual-level interventions are effective in reducing stress and promoting well-being (Table 2).  Meta-analytic studies report significant medium and large effect sizes for cognitive-behavioral therapies (d=0.52, van der Klink et al., 2001; d=1.00, Richardson & Rothstein, 2008), small and large effect sizes for relaxation techniques (d=0.31, van der Klink et al., 2001; d=0.83, Richardson & Rothstein, 2008) and small effect sizes for health promotion methods (RR=1.25, Kuoppola et al., 2008; d=0.19, Conn et al., 2009).  These findings suggest that CBT has a comparatively larger effect on stress and well-being and van der Klink et al., (2001) found that effect sizes for CBT were significantly greater than those for relaxation methods.  However, this finding must be treated with caution, as their analysis included effects across all outcomes, including those unrelated to employee well-being.   Furthermore, as there are no comparisons between CBT, relaxation and other interventions, it is difficult to make claims about relative effectiveness of secondary individual-level interventions at this juncture.   It can also be noted that existing meta-analyses of secondary individual-level SMI do not focus on mindfulness training as a separate category, although mindfulness studies are included in some meta-analysis (e.g., Regehr et al., 2014).  This means that we lack a statistical summary of their effects and a comparison of mindfulness and other methods.  Promisingly, most primary studies report positive effects of mindfulness training on psychological well-being (e.g., Cohen-Katz, Wiley, Capuano, Baker, & Shapiro, 2005; Hülsheger et al., 2013), whereas only a few studies have failed to yield supportive evidence (e.g., Malarkey, Jarjoura & Klatt, 2013; van Berkel, Boot, Proper, Bongers, & van der Beek, 2014).

          Meta-analytic studies of multimodal secondary individual-level interventions report significant medium effect sizes (d=0.48, van der Klink et al., 2001; d=0.60 and 0.42, Richardson & Rothstein, 2008).  These effect sizes are not substantially different to studies of single secondary individual-level SMI, meaning that multimodal secondary individual-level SMI do not provide a clear advantage over single secondary individual-level SMI.  But it is worth noting that Richardson and Rothstein (2008) found that the effects of multimodal SMI increase over time and their effects are longer lasting.  One interpretation of this finding is that employees take longer to learn stress management techniques in multimodal SMI but, once learnt, their effects are stronger and more sustainable.  Again, such a conclusion, while plausible, must be treated with caution for present purposes, as the comparison by Richardson and Rothstein included well-being and well-being related outcomes such as absenteeism and physiological responses. 

           Another important feature of research on secondary individual-level SMI is the heterogeneity and direction of effect sizes.  In particular, effect sizes vary significantly and, to the best of our knowledge, almost all effects are positive in direction, even if they are non-significant.  For example, in Van der Klink et al.’s (2001) meta-analysis, effect sizes across CBT and relaxation interventions ranged from 0.0 to 2.2, and 22 of the 35 studies reported non-significant findings.  A similar heterogeneity in effect sizes is reported in other meta-analyses of SMI and in the literature on mindfulness training.  This implies that although secondary individual-level SMI are not harmful to employee well-being, their positive effects are not realised in all contexts.  The heterogeneity in effect sizes therefore suggests that implementation processes, contextual factors or participant differences (e.g., motivation, initial levels of well-being) may impact on the efficacy of secondary individual-level SMI.  However, a lack of research and contradictory evidence prevent firm conclusions being drawn about the effects of context and implementation process on secondary individual-level SMI effectiveness (Van der Klink et al., 2001; Richardson & Rothstein, 2008; Rongen et al., 2013).

Table 2. Meta-Analyses of Stress Management Intervention Effectiveness

Note: RR = Risk Ratio; Cohen’s d effect sizes, 0.2 = small, 0.5 = medium, 0.8 = large. Effect sizes are uncorrected for reliabilities.

¹Authors' contributions were equal.

²For reviews of other SMI, such as tertiary individual-level interventions, see Csiernik (2011) and McLeod (2008).

³Ruotsalainen et al. (2015) provide a meta-analysis of particular types of organisational intervention but the sample sizes are very small and they provide no effect size based on all interventions.