Does Psychotherapy work in treating workrelated stress?
If so – what works in psychotherapy treating workrelated stress?
By Carsten Thomas Hansen
Introduction
Research in psychotherapeutic effectivenessi has been ongoing since the beginning of the 19th century. It is now generally accepted that psychotherapy is effective. When attempting to answer the question of psychotherapeutic effectiveness one is also attempting to answer the question of absolute efficacy of psychotherapy which is: does a treatment work better than no treatment (Wampold et al. 2015). The purpose of this scholarly review is to ”translate” the empirical quantitative evidence of ,what is typically named as, SMI (Stress Management Intervention) effectiveness into an understanding of what part of the SMI qualify as psychotherapy. In reading the quantitative research on workrelated stress (sometimes also named ”occupational stress management”) SMI do not always qualify as psychotherapy. So by distinguishing between SMI and psychotherapy the aim is first of all to shed light on if psychotherapy in treating workrelated stress is better than no treatment? According to Wampold et al.(ibid. p. 112) the approx. 40 years old controversy about whether psychotherapy produced outcomes that were better than the rate of spontaneuous remission is today still put to sleep because it is difficult to find an example of an empirically tested psychotherapy offered to a treatment-seeking client that is not effective. Taken this in account I hypothesis that research will show that psychotherapy is effective in treating workrelated stress.
If treatment is better than no treatment what does the empirical evidence tell us regarding relative efficacy i.e. do the various psychotherapies vary in efficacy in treating workrelated stress (ibid.). By example and with reference to meta-analysis´ by Benish et al (2007) and Wampold et al (2010) research on the effectiveness of PTSD psychotherapeutic treatment fails to reject the null hypothesis that PTSD treatments are similarly effective. Wampold et al (2015) concludes that from the 1980s exemplary studies and methodologically sound meta-analyses produced evidence that demonstrated that there were small, if not, zero, differences among treatments.
Work-related stress & Psychotherapeutic interventions in the litterature
The phenomenon of work-related stress is according to the litterature at least researched on four continents of the world: Europe, Australia, Asia and America.
According to Greiner (2008) the European Agency for Safety and Health at Work, work-related stress affects 28% of workers in the EU and according to Richardson et al. (2008) The American Institute of Stress reported that stress is a factor in up to 80% of all work-related injuries and 40% of workplace turnover. The confederation of British Industry reported stress as the second highest cause of absenteeism among nonmanual workers in the UK and in Australia states report an increasing number of workers´compensation claims resulting from workplace stress (ibid.). The most recent study in Denmark done with the sample size of 5000 employees in 39 organisations was in 2002 and found that in 23,5% of the incidents stress was documentet as the reason for a sickleave period lasting more than 2 weeks (Dasam 2012).
Work-related stress is according to WHO (2016) defined as ”the response people may have when presented with work demands and pressures that are not matched to their knowledge and abilities and which challenge their ability to cope”. WHO´s definition point to a shared understanding in the literature of work-related stress as a result of organizational factors and an unevennes of demands, personal skills & ressources and social support at work (Alin et al 2011, Richardson et al. 2008, Dasam 2012, Routsalainen et al 2014, NRC 2014, Netterstrøm 2015, Karasek 1990). Work- related stress is associated with psychological and physiological conditions that depart from the persons normal functioning (Ibid.). In Denmark the ICD-10´s definition of adjustment disorder and ”belastningsreaktion” (ICD10 code F43.0 og F43.2) is commenly used in the clinical context to classify the condition of work-related stress and document the clinical need for referal to treatment by a psychologist or other professional (Dasam 2012).
Research history on treatment effect in the field of work-related stress is characterized by threats to both external –and internal validity of the RCT´s included. Including RCT´s that do not make use of control groups becomes a problem related to the internal validity of the study when claiming positive effects (see Murphy 1984 & DeFrank and Cooper 1987). If the population in the RCT are not representative of people experiencing work-related stress then the problem is related to the external validity of the study. Nicholson et al. (1988) claimed positive effects but only a third of the studies in Nicholson´s review included clients from occupational settings (Richardson et al. 2008).
Richardson defines the classification of interventions on behalf of Murphy & Sauter (2003) as primary, secondary and tertiary (Richardson et al 2008 Bunce et al. 2000). Primary interventions attempt to alter the sources of stress at work. Secondary interventions are preventive and attempt to reduce the severity of stress symptoms before they lead to serious health problems. Tertiary interventions are designed to treat the employee´s health condition. According to van der Klink (2001) interventions are categorized as either 1. Aiming to increase individual psychological ressources and responses or 2. Aiming to change the occupational context. In terms of what kind of interventions can qualify as psychotherapy Wampold defines psychotherapy as the following: ”Psychotherapy is a primarily interpersonal treatment that is a) based on psychological principles; b) involves a trained therapist and a client who is seeking help for a mental disorder; problem, or complaint; c) is intended by the therapist to be remedial for the client disorder, problem, or complaint: and d) is adapted or individualized for the particular client and his or her disorder, problem, or complaint” (Wampold et al p.37 2015). Based on Wampolds definition Richardsons secondary & tertiary interventions and van der Klinks first categorization would qualify as psychotherapeutic treatment and dependent on the exact study in focus include both individualized and/or group treatment. Organizational development and job redesign is excluded from this paper.
Methodological considerations concerning criteria for evaluating included studies
The following section will briefly consider two research designs that are used for collecting and utilizing data to establish absolute efficacy and relative efficacy (ibid.)
Absolute Efficacy
Randomised controlled studies are best qualified to answer the question if a specific treatment is working according to intentions or not (Dasam 2012). Absolute efficacy refers to the effects of treatment vis-á-vis no treatment and is best adressed by a research design where treated clients are contrasted to untreated clients (Wampold et al. 2015).
In the prototypical design to test for efficacy, clients meeting the study criteria would be randomly selected from a population and then randomly assigned to one of two groups: a treatment group and a no-treatment control group. If the post-treatment scores of the treated sample are significantly superior to those of the control group, the treatment is considered efficacious (Ibid).
The other paramount research design is meta-analysis. Meta-analysis is a statistical method for combining evidence from the results of many independent RCT´s that address a related set of hypotheses with different measures (Wampold et al. 2015). The critical step in meta-analysis is to aggregate effects from studies investigating the same hypothesis. If there have been ten RCT´s that investigated the efficacy of the treatments for work-related stress, then the effects from each RCT can be ”averaged” to form an aggregate effect d, providing an estimate of the treatment effect, which is more precise than the estimate provided by a single RCT.ii (Wampold et al p. 69 2015).
Relative Efficacy
The central design for establishing relative efficacy is the comparative outcome strategy.
Meta-analysis can address issues raised by the interpretation of primary studies and can be used to estimate a effect size for relative efficacy thus testing the null hypothesis that treatments are uniformly effective vs. that the treatments vary in effectiveness (Wampold et al 2010, Wampold et al 2015).
This paper will present two methods used to establish relative efficacy.
The first method is treatment package design utilizing no-treatment control groups. By example Bisson et al (2007) used the treatment package design in their meta- analysis of psychological treatments for chronic post-traumatic stress disorder. For further discussion see Wampold et al 2010 & Ehlers et al 2010. According to the treatment package design a. treatments examined in studies are classified into categories (e.g., CBT, dynamic etc.) b. the effect size is computed for each treatment vis á vis the no treatment control group c. the effect size within a category are averaged and d. the mean effect sizes for the categories are compared (Bisson et al 2007 & Wampold et al 2015).
According to Wampold it is problematic to make inferences based on no-treatment control group designs because the studies in a given category (CBT compared with no-treatment control group) may be differenct from the studies of treatments in other categories (dynamic treatment with no-treatment control group) on such factors as severity of disorder treated, treatment standardization, treatment length, allegiance of researcher, comorbidity of clients and outcome measures employed.
The treatment package design involves threats to validity that the other method, the direct comparison design, avoid. In the direct comparison design only studies that directly compare two types of treatments would be examined. This design avoids the mix up that are created by comparing classes of treatments compared to controls because factors such as problem treated, severity of disorder, setting and outcome measures employed would be comparable for each direct comparison due to random assignment e.g. each direct comparison of cognitive behavioral therapy and solution focused therapy would use the same outcome measure (Wampold et al 2015, Benish et al 2007).
Inclusion criteria
a. Be a meta-analysis which aim is to determine the effectiveness of psychoterapeutic treatment for work-related stress. b. Qualifiy as psychotherapy by definition mentioned on page 2
3. Include either secondary -or tertiery intervention.
Search strategi:
Preferring Meta-analysis above Reviews and primary studies (RCT´s).
Computerized literature searches were conducted in oct. 2015 using PudMed/psycLIT and Rex.
Keywords employed in the ”PudMed/psycLIT & Rex”: Meta-analysis, occupational, stress management, psychotherapy, work-related stress, evaluation, job-related stress. Following articles was found and included: Virgili (2013), Alin et al. (2011), Richardson et al. (2008), van der Klink et al. (2001), Ruotsalainen et al. (2014), MTV
(2012). The excluded studies is present in the list of references and is not discussed in this scholarly review.
In extension I performed a network search which gave knowledge of one dutch PhD dissertation ( Arends 2014) and a Danish MTV (Medical Technological Evaluation) (Dasam 2012).
Reviewing the evidence for absolute-and relative efficacy in treating work- related stress
Results of the four included meta-analysis show in terms of absolute efficacy that there is evidence that psychotherapy in treating workrelated stress is better than no treatment.
The van der Klink et al. (2001) meta-analysis conclude that stress management interventions are effective revealing an effect size across all studies (d=0.34 (Ibid. p. 272); which can be interpreted as indicating that the average treated person will be better of than 63% of the population (Wampold et al. p. 70)). The study hypothesis was to determine if stress interventions is effective and if so, which type of stress intervention is most effective?
The meta-analysis included forty-eight experimental and quasi-experimental studies (n=3736) utilizing the treatment package design categorizing four types of interventions distinguishing between Cognitive-behavioural, relaxations techniques, multimodal programs. The included RCT´s should be specifically designed to prevent or reduce psychological complaints related to occupational stress, target population was working healthy individuals with already manifested stressrelated psychological problems.
The Richardson et. al. meta-analysis (2008) shares the same hypothesis as van der Klink to ”identify what works & how well it works” (ibid. p. 73) and concludes that the weigthed average effect size from each individual intervention, yielded a significant effect size (d=0.526, 95% CI=0.364,0.687).
The total sample size was 2.847 healthy people from a wide range of occupations, 36 RCT´s were included representing 55 interventions, reporting sample sizes, means and SD for treatment –and no treatment groups. The treatment package design was utilized and treatments were categorised as cognitive-behavioural, relaxation, multimodal and alternative.
In 2011 Alin et al. finds that the overall mean effect size for CBT treatments indicates a strong effect (d=0.81). According to their own conclusion ”results demonstrate the effectiveness of CBT-based intervention programs ” (Alin et al. 2011 p. 221).
The Meta-analysis focuses on CBT distinguishing between treatments based on Rational Emotional Behavioural Therapy (REBT) and other CBT treatments. The study hypothesis is to 1. determine the effectivenes of CBT-based interventions in reducing emotional distress in occupational settings 2. calculate a quantitative estimate of the overall effect size of REBT-based interventions in reducing emotional distress. The meta-analysis utilized the treatment package design but only included CBT-based treatments on emotional distress (23 studies), a population sample (1282 participants) without clinical symptoms, studies including a control group and studies reporting data on means, SD and inferential statistics.
Instead of a singular focus on CBT Virgili´s et al. (2013) choice of focus is on mindfulness-based interventions (MBI) and conclude that MBIs were shown to have a medium-large effect on psychological distress in working adults revealing an overall effect size for the combined measures of psychological distress = Hedges´s g=0.68.
The study hypothesis was to assess 1. the effectiveness of MBI for reducing psychological distress in working adults 2. the effectiveness of MBIs in comparison to active control conditions.
The meta-analysis included 19 intervention studies utilizing both the treatment package design and the direct comparison design in establishing relative efficacy. Studies that evaluated the effects of MBSR or a similarly structered interventions were included. With a total of 1.139 participants mostly self-selected with a target population of working healthy individuals using validated scales of measurement at inclusion.
The purpose of the Dasam review is to document the empirical evidence for a specific treatment program (including a clinical population sample) used by the Occupational medicine clinics in Denmark (arbejdsmedicinske klinikker). The conclusion from Dasam is that the empirical evidence on the treatment of work- related stress measured as according to ICD-10´s adjustment disorder and ”belastningsreaktion” (ICD10 code F43.0 og F43.2) is that ”solutionsoriented” psychotherapy gives symptom reduction compared to control groups. Dasam included 19 international RCT´s with a total of 1.674 healthy participants from the working popluation, using validated outcome measures and psychotherapy should build on CBT, psychological education and mindfullness. Dasam defines it as ”solutionsoriented psychotherapy. Standardized protocols are available and described in Dasam 2012.
The review by Ruotsalainen et al. (2014) for the Cochrane Collaboration found that there was low-quality evidence that CBT in treating work-related stress in healthcare workers with or without relaxation techniques at one to six months follow-up reduced stress more than no intervention (SMD -0.27; 332 participants).
The study hypothesis was to evaluate the effectiveness of work- and person-directed interventions compared to no intervention or alternative interventions. Interventions was aimed at preventing or reducing stress arising from work categorizing three types of interventions distinguishing between cognitive-behavioural training (CBT), mental and physical relaxation and combined CBT and relaxation. 45 studies used a no- intervention control group and 7 RCT´s utilized direct comparison between one or more active stress treatments. The target population was healthy individuals working in the healthcare setting including 58 RCT´s with 7188 participants categorizing outcomes according to validated self-report questionnaires. There were used multiple outcome measures and for that reason the review transformed the means into standardised mean differences (SMDs).
In terms of relative efficacy three meta-analysis finds difference in effect and one find no difference as did one review.
The van der Klink finds that cognitive-behavioural interventions are more effective than other intervention types. A moderate effect was found for cognitive-behavioural interventions (d=0.68; no. of par.=858), multimodal interventions (d=0.51; no. of par.=470) and a small effect was found for relaxation techniques (d=0.35; no. of par.=982).
Larger effect sizes were found for remedial programs (n=4, d=0.59) than for preventive programs (n=44, d=0.32).
Likewise Richardson finds that treatments vary in effect and concludes that cognitive-behavioural treatments consistently produced larger effects than other types and Alin finds that CBT treatments vary in effects because when separating the studies in two catagories the effect size for REBT-based interventions increased while the effect size for the rest of the CBT treatments slightly decreased. Alin conclude that ”REBT based intervention programs seems to be the most efficient, as supported by a mean effect size situated in the ”strong” range of Cohen´s criteria”. (Alin et al. p. 232).
Contrary to the former three meta-analysis Virgili finds that the effect sizes were similar across intervention type (i.e. standard MBSR, brief MBSR and other MBIs). One study demonstrated MBI to be superior to an active control. Four studies of direct comparison between MBI and relaxation training and between MBI and yoga showed no difference in effectiveness between interventions. Virgili´s results are in line with Sedlmeijer et al. (2012) meta-analysis which concludes that differenct types of meditations showed equal effect sizes on healthy people experiencing stress.
The Dasam Review do not contribute in answering the question of relative efficacy (Wampold 2015). It is limited by its own inclusion criteria when only searching for studies that do not apply a research design that are fit to utilize and establish relative efficacy.
The Routsalainen review found that in 7 RCT´s doing direct comparison treatments were uniformly effective. Comparisons of CBT with other active treatments there were found no considerable differences and with relaxation interventions the reductions in stress levels were comparable with those of CBT.
Threats validity
First: threats to the external validity: In the van der Klink meta-analysis population characteristics such as sex, age, years of employment, occupational status was not available in the included RCT´s and none of the 48 RCT´s included a population consisting of people seeking help. They were volunteerly selected not clinical refered. The meta-analysis by Alin and Virgili suffers same limitation in that it opt-out population from clinical settings. Alin dont provide information on demographics. In the Virgili study particpants were largely self-selected, older, predominantly female, more higly educated and exhibiting mild to moderate elevations of stress at baseline thereby limiting the generalisability of these findings to populations with similar profiles.
Although the Richardson meta-analysis speaks of reducing selection bias, reporting on attrition and increasing the internal validity of the study by including new studies, eliminating studies of varying study designs and methodological quality it is faulty one important issue: the population sample do not represent clients seeking help but are volunteers as in the van der Klink, Alin & Virgili studies.
Second: As Richardson speaks of increasing the internal validity of their study it is in comparison to the van der Klink meta-analysis where nine included studies used quasi-experimental designs, dont inform on blinding, others dont include a no- treatment control group, seven randomized experiements dont report sufficient statistics to calculate an effect size. 44 studies was considered preventive and no participant selection had taken place in regard to stress level at inclusion. Alin dont provide information on attrition, blinding and dont inform on outcome measures.
Third: All four meta-analysis utilizes the treatment package design vis á vis controls. Only the Virgili meta-analysis & the Routsalainen review includes RCT´s doing direct comparison.
Discussion & Main results
The aim of this paper was to answer two basic questions that of absolute efficacy: does psychotherapy in treating work-related stress work better than no treatment? And that of relative efficiacy: do the various psychotherapies vary in efficacy in treating workrelated stress?
The paper found evidence showing that psychotherapy in treating work-related stress works better than no treatment.
The included studies show a skew in the research population of different psychological treatments with an overrepresention of CBTs. This could be interpreted as if CBT is the most effective treatment but that would be faulty. The results indicate that CBT works but in terms of absolute efficiacy CBT is not the only treatment that is shown to be efficacious.
It is argued that all 4 meta-analysis have problems concerning both external & internal validity. According to the research history on work-related stress this is not a new problem. Refering to Wampolds definition of psychotherapy the population for whom the treatments should be intended to help is people who is seeking help. Most worrying is that the Dasam review is the only one that included a population sample who were clinical referred seeking help. The remainder of the studies included volunteerly selected healthy people. For the majority of the included research the absolute effect found represents populations with similar profiles.
In regard to the internal validity the quality of majority of the included research is very fluctuating which in sum leaves the question open to the possibility that occured change also could be the result of other extraneous factors.
By example the Routsalainen review was the only included study that considered the reporting of stress by questionnaire as an outcome assessment that could be biased by knowledge of the intervention. And very few authors in the Routsalainen review mentioned that blinding could be an issue because they argue it is impossible here. But as Routsalainen this paper agree that this does not eliminate the risk of bias anyway.
Even though the absolute effect size according to Routsalainen is small comparing treatment to no-treatment this can according to Bunce et al. (2000) be explained by the fact that the review is a preventive study. When the reports of the participants level of strain is low then little change in outcome variables can be expected (Ibid. p. 198, 2000). This is in line with the van der Klink study showing that Larger effect sizes were found for remedial programs (n=4, d=0.59) than for preventive programs (n=44, d=0.32).
In regard to the hypothesis of relative efficacy three meta-analysis finds difference in effect and one find no difference as did one review.
The present paper adresses a primary criticism of the included meta-analysis when claiming the superiority of CBT on behalf of the application of research design that
are not fit for utilizing and establishing relative efficacy between different treatments. Van der Klink, Richardson and Alins choice of design becomes a problem concerning the internal validity of the meta-analysis when they conclude that CBT is the most effective. It is this papers interpretation that is seems premature to conclude that CBT is more effective than other treatments in treating work-related stress.
Utilizing the treatment package design introduces significant threats to validity. As shown studies vary in regard to participants, outcome measures used, treatment standardization, measure reactivity, blinding procedures, treatment length, severity of disorder and mulitiple unmeasured variables. Consequently, any differences between categories of treatments could be attributed to differences among the categories relative to these variables, creating any number confounds. For example no uniform outcome measure was used for any of the RCT´s, so that an apparent superiority actually reflects the difference between measure used or one class of treatments may have used more reactive measures generally and the apparent superiority could reflect the ease with which varibles change. One solution to this problem is to meta- analytically examine potential mediators and moderators of effect size. This strategy introduces a number of additional problems including omitting important variables which is the case amongst the included meta-analysis and reviews estimating relative efficacy (Benish et al 2007, Wampold et al 2015). The applied research designs do not give conceptual attention to include or control for confounds such as non-specific factors or specific therapeutic components which would have been possible if they had used either a direct comparison strategy or a dismantling research design (ibid). The direct comparison strategy controls for potential confounds involved in the comparison of pscyhotherapies because the purported confounding variables are constant. For example, reactivity of measurement is no longer a confound in a direct comparison study because both treatments are assessed with same measures (ibid.)
Contrary to the van der Klink, Richardson & Alin meta-analysis, Virgili and Routsalainen includes RCT´s doing direct comparison. Both looks for difference but finds uniform effect. Virgili concludes that this leaves the question open to the possiblity that the effectiveness of MBIs may be due non-specific factors or non- mindfulness components (Virgili 2013). Which this paper agrees because both studies omitt focus on therapueutic factors in general.
Many other factors or therapeutic components could have contributed in finding relative efficacy between treatments such as researcher allegiance, alteration of legitimate protocols, therapist effects, client factors, reactivity of the recording instrument or other non-specific CBT components.
In terms of answering the question of relative efficiacy this paper concludes that it is not possible to answer the question regarding relative efficacy: do the various psychotherapies vary in efficacy in treating workrelated stress. There is not made research that can answer the question.. In terms om relative efficiacy current meta- analysis and reviews do not contribute in answering the question if there is specific factors in treating work-related stress that makes one type of psychotherapy better than the other.
What is new and which there is no likewise readings? is the focus of distinguishing between absolute & -relative effect when treating work-related stress. This is an important contribution. It makes it possible for researchers and clinicians treating work-related stress to distinguish between which therapuetic factors do we know of and which can we talk of that we do not yet know of that makes treatment work. Both when comprehending treatment efficacy in regard to research on absolute efficiacy and when comprehending treatment efficiacy in regard of research on relative efficiacy.
References
Alin, R. D., & Szamoskozi, S. (2011): A Meta-Analytical Study On The Effects Of Cognitive Behavioral Techniques For Reducing Distress In Organisations. Journal of Cognitive and Behavioral Psychotherapies. Vol. 11, No. 2, 221-236.
Arends I, Bruinvels DJ, Rebergen DS, Nieuwenhuijsen K, Madan I, Neumeyer- Gromen A, Bültmann U, Verbeek JH (2012): Interventions to facilitate return to work in adults with adjustment disorders (Review). The Cochrane Collaboration. The Cochrane Library 2012, Issue 12.
Arends, I. (2013): Prevention of Recurrent Sickness Absence in Workers with Common Mental Disorders. Rijksuniversiteit Groningen.
Barkham, M, & Shapiro, D. A. (1990): Brief psychotherapeutic interventions for job- related distress: A pilot study of prescriptive and exploratory therapy. Counselling Psychology Quarterly, Vol. 3 Issue 2, p133.
Benish, S., Imel, Z. E., & Wampold, B. E. (2007). The relative efficacy of bona fide psychotherapies of post-traumatic stress disorder: A meta-analysis of direct comparisons. Clinical Psychology Review, 28, 746−758.
Bisson, J. I., Ehlers, A, Matthews, R., Pilling, S., Richard, D., & Turner, S., (2007). Psychological Treatments for chronic post-traumatic stress disorder. British Journal of Psychiatry. Systematic review and meta-analysis. 190, 97-104.
Bhui, K.S., Dinos, S., Stansfeld, S. A., and White, P. D. (2012): A Synthesis of the Evidence for Managing Stress at Work: A Review of the Reviews Reporting on Anxiety, Depression,
and Absenteeism. Journal of Environmental and Public Health Volume 2012, Article ID 515874, 21 pages.
Bunce D., & Stephenson K. (2000): Statistical considerations in the interpretation of research on occupational stress management interventions. Work & Stress. Vol. 14, No. 3. 197-212.
Dasam – Dansk Selskab for Arbejdsmiljø- og Miljømedicin (2012): Medicinsk Teknologi Vurdering af Behandling af arbejdsrelateret Stress.
DeFrank, R. S., & Cooper, C. L. (1987). Worksite stress management interventions: Their effectiveness and conceptualizations. Journal of Occupational Health Psychology, 5, 309-320.
Ehlers, A., Bisson, J., Clark, D. M., Creamer, M., Pilling, S., Richards, D., Schnurr, P. P., Turner, S., Yule, W., (2010). Do all psychological treatments really work the same in posttraumatic stress disorder? Clinical Psychology Review. Vol. 30, 269-276.
Greiner, A. (2008). An economic model of work related distress. Journal of Economic Behavior & Organiztion, 66, 335-346.
Karasek, R., Theorell Töres (1990): Healthy Work. Stress, Productivity and the Reconstruction of Working life. Basic Books
Murphy, L. R. (1984): Occupational Stress Management: A review and appraisal. Journal of Occupational Psychology, 57, 1-15.
National Research Centre (NRC) for the Working Enviroment (2014): Arbejdsmiljø og Helbred i Danmark 2012-2020.
Netterstrøm, Bo (2015). Stress og arbejde. Hans Reitzels Forlag.
Nicholson, T., Duncan, D. F., Hawkins, W., Belcastro, P. A., Gold, R. (1988): Stress Treatment: Two Aspirins, fluids, and one more workshop. Professinal Psychology: Research and Practice, 19, 637-641.
Virgili, Mario (2013): Mindfulness-Based interventions Reduce Psychological Distress in Working Adults: A Meta-Analysis of Interventions Studies. Springer Science+Business Media New York 2013.
Reynolds, S., Emma, T., Shapiro, D., (1993): Session Impact and Outcome in Stress Management Training. Journal of Community & Applied Social Psychology, Vol. 3. 325-337.
Reynolds, S. (1994): Stress Management At Work: With Whom, For Whom And To What Ends?. British Journal of Guidance & Counselling. Feb94, Vol. 22, Issue 1, p75. 15p.
Richardson, K. M., & Rothstein, H. R. (2008): Effect of Occupational Stress Management Intervention Programs: A Meta-Analysis. Journal of Occupational Health Psychology. Vol. 13, No. 1, 69-93.
Ruotsalainen, J. H., Verbeek J. H., Mariné, A., Serra, C. (2014): Preventing occupational stress in healthcare workers (Review). The Cochrane Collaboration. Issue 12. 1-150. Wiley.
Sedlmeier, P., Eberth, J., Schwarz, M., Zimmerman, D., Haarig, F., Jaeger, S., Kunze. (2012). The psychological effects of meditation: a meta-analysis. Psychological Bulletin. Nov;138(6):1139-71.
Shapiro, D. A., & Firth Jenny (1986): An evaluation of psychotherapy for job-related distress. Journal of Occupational Psychology, 59, 111-119.
Van der Klink, Jac J. L., Roland, W. B. B., Schene, A. H., van Dijk, F. J. H. (2001): The Benefits of Interventions for Work-Related Stress. American Journal of Public Health. Vol. 91, No. 2. 270-276.
Wampold, B. E., Imel, Z. E., Laska, K. M., Benish, S., Miller, S. D., Flűckiger, C., Del Re, A. C., Baardseth, T. P., Budge, S. (2010): Determining what works in the treatment of PTSD. Clinical Psychological Review. Dec;30(8):923-33.
Wampold, B. E., & Imel, Z. E. (2015): The Great Psyhotherapy Debate. The Evidence for What Makes Psychotherapy Work. Second Edition. Routledge.
World Health Organization (2016): http://www.who.int/occupational_health/topics/stressatwp/en/
i I use the terms efficacy and effectiveness as according to the definition of Wampold et al. 2015 p. 97. As Wampold et al. points out the distinction between the context of clinical trials and naturalistic setting is not always clear. In this review I do not make at strict distinction between the terms.
ii The null hypothesis tested is that the aggregate effect size is zero (the treatment is not efficacious) vs. the alternative that the effect is different from zero .With meta-analysis sample sizes grow and thereby increases the ability of the study to detect change in outcome variables. This is defined as statistical power. Statistical power is the proability that a statistical test will correctly reject the null hypothesis when it is false. So as statistical power increases as a function of sample size, larger samples are required to detect smaller effect sizes (Bunce et al. 2000). But there is a flipside to that coin. With sufficient statistical power a metaanalysis will almost surely result in rejecting the null hypothesis and thereby reflect true effects very close to zero but even if the null is true five percent of the studies could be false positives – falsely reject the null. So no study kan rule out all threats to validity and no study will provide complete refutation (Wampold et al 2015).