A study protocol for a cluster randomised controlled trial on mindfulness-based stress reduction: studying effects of mindfulness-based stress reduction and an additional organisational health intervention on mental health and work-related perceptions of teachers in Dutch secondary vocational schools

Dutch society needs healthy teachers to maintain and improve the quality of the education sector and to enhance student performance [1]. In all sectors, however, work-related stress has become an inherent feature of the employment relationship in industrialised countries such as the Netherlands [2]. Work-related stress is an increasingly important cause of workers’ mental health problems, such as stress symptoms, overstrain, and burnout, which can decrease work performance [3, 4]. In 2017, almost one in six Dutch employees reported stress or burnout complaints. In the educational sector, this figure was more than one in five employees [5]. More than 30% of teachers have reported that major changes in the work context are an important cause of work-related stress. Teachers are expected to meet higher job demands (e.g., high workload, emotional strain) with fewer job resources, especially less professional autonomy [5].

Job demands can be defined as the physical, social, or organisational aspects of the job that require sustained physical or psychological effort [6]. The increase in teachers’ workload is caused by numerous administrative tasks and school reforms. The growing needs of students also generate emotional strain [7]. Job resources can be defined as the physical, social, or organisational aspects that may help teachers to achieve goals and to stimulate learning and development. As such, job resources can buffer the influence of job demands [6, 8].

Work-related stress is associated with several negative organisational outcomes, such as increased absenteeism and early retirement [5]. In comparison with the agriculture, information, and communication sectors, the absenteeism rate in the educational sector is relatively high: 5.3% in the latter versus less than 3% in the former sectors in 2017 [5]. In the Netherlands, the costs of work-related stress absenteeism for the total workforce is €1.8 billion, of which €275 million involves the costs in the educational sector. Work stress-related absenteeism costs are the highest in the educational sector: almost €6000 (number of days × costs per day) for each employee who is absent [9]. When a teacher is absent, organisations in the educational sector strongly rely on the (mostly serendipitous) availability of substitutable colleagues to cover for the absent worker. Consequently, colleagues are overloaded (i.e., a job demand), while the job resources they can draw from remain the same at best. This pattern creates an imbalance between these colleagues’ job demands and resources, which can jeopardise their well-being [7]. This imbalance between job demands and resources and its associated risk of negative effects on one’s well-being may be an important reason that many novice teachers leave the educational sector within the first 5 years of their career [10] and that many experienced teachers retire early. In fact, 45–70% of early retirements in the educational sector can be attributed to psychosomatic and psychological problems [7]. Therefore, it is extremely important to reduce and prevent stress and absenteeism in the occupational sector and to develop effective mental health management interventions, which can be both person-focused and organisation-focused.

A high percentage of Dutch employees (57%), especially in the educational sector, ask for interventions to address work-related stress problems [5]. Many employers in this sector (48%) also recognise the risk of stress [5]. Preventive interventions can be classified as primary, secondary, or tertiary. Primary interventions, which are oriented to the organisational level, aim to change the sources of work-related stress. Secondary and tertiary interventions, both of which are focused on the individual employee, aim to decrease stress symptoms before they cause mental health problems and to treat mental health problems (e.g., burnout), respectively [11]. Mental health interventions in the educational sector are mostly secondary preventive and targeted at the individual level, with the goal of enhancing the ability of teachers to cope with stressors in the workplace [12‐18]. Examples are workshops on stress management skills and mindfulness-based stress reduction (MBSR) programmes. MBSR has been shown to be partly effective in influencing mental health outcomes [19].

From a health perspective, primary prevention - when possible - is preferable to secondary and tertiary prevention. In their review of occupational stress interventions in Australia, Caulfield et al. [20] suggested that primary interventions generate more positive changes in comparison with individual-focused secondary or tertiary interventions. However, two meta-analyses on work-related stress interventions [21, 22] found no substantial differences between organisational-level and individual-level interventions. One explanation is the complexity of organisational-level interventions, which might hinder the implementation and measurement of outcomes [7, 23]. In view of this, an appropriate (i.e., mixed-methods) evaluation of an organisational health intervention may require consideration of multiple process outcomes to monitor the implementation process and to investigate the outcomes of the intervention in depth [7, 24]. We agree with Van der Klink et al. [25] that there is a need for an integrated approach that combines both an individual-focused intervention and an organisation-focused intervention.

Two systematic reviews have shown that an MBSR intervention programme in the workplace can significantly affect deficit-based outcomes, such as emotional exhaustion (one of the three dimensions of occupational burnout), (occupational) stress, psychological distress, anxiety, and depression [19, 26]. Three systematic reviews also identified significant improvements in asset-based outcomes, such as mindfulness, personal accomplishment (a dimension of burnout), (occupational) self-compassion, quality of sleep, relaxation, and job performance [19, 26, 27]. The systematic review by Donaldson-Feilder et al. [28] reported positive effects on the well-being, resilience, and leadership capability of leaders/managers. Slutsky et al. [29] conducted a randomised controlled trial (RCT) and suggested that small doses of mindfulness training (half-day training) are sufficient to increase job productivity, but that larger doses (6-week training) are needed to improve attentional focus at work, job satisfaction, and work-life balance. The systematic review by Donald et al. [30] identified a positive relationship between mindfulness (both operationalised as a personality variable and as an intervention) and prosocial behaviour.

In a meta-analysis, Klingbeil and Renshaw [31] mentioned that mindfulness-based interventions with teachers are promising for increasing their mindfulness and psychological well-being and for decreasing psychological distress. Overall, they concluded that their findings were similar to the outcomes found in other meta-analyses of the effects of such interventions on employees’ mental health.

Research on mindfulness is often criticised for its poor methodological quality [32, 33]. However, it is impossible to conduct such research using a double-blind placebo-controlled design, which is often applied in medical interventions [34]. It is obvious that participants cannot be kept blind to the fact that they are (or are not) assigned to an MBSR training programme. This raises questions about which methodological features should be included to improve the research design. Goldberg et al. [32] highlighted six features: (1) active control conditions to consider the amount of non-specific attention participants receive, called the Hawthorne effect [35]; (2) larger sample sizes; (3) longer follow-up assessment to measure the sustainability of training effects; (4) evaluation of treatment fidelity; (5) reporting of instructors’ skill levels; and (6) intention-to-treat (ITT) analysis. Three other important features are assessing a diversity of outcomes (negative and positive, process and effect measures, mental health and work performance); using a mixed-methods approach that combines quantitative and qualitative data; and combining an individual-focused intervention, such as MBSR, with an additional organisational intervention (i.e., taking an integrated approach) [19].

The key points of participatory action research (PAR) [36] are the effective ingredients for organisational interventions: having a bottom-up approach; composing a participatory group; fostering active participation by stakeholders (e.g., employees) and collaboration between researchers and stakeholders; using stakeholders’ knowledge, skills, and perceptions; and creating joint ownership of problems and solutions [7]. Solutions from stakeholders appeared more effective than solutions adopted by others [37]. The belief that one is the master of one’s own behaviour and is able to influence others and the environment (i.e., an internal locus of control) is crucial [38]. In other words, the organisational health intervention should target individuals’ occupational self-efficacy: the belief in one’s own ability in a specific domain of work. The most effective way to enhance one’s self-efficacy is through mastery of experiences [7, 39]. By taking part in the organisational intervention or even by experiencing its effects, we assume that occupational self-efficacy can be influenced to decrease burnout. Indeed, Consiglio et al. [40] found a negative relationship between occupational self-efficacy and burnout, which appeared to be partially mediated by job demands and job resources.

The proposed study aims to contribute to the debate on prevention of work-related stress and burnout, and improvement of work engagement and work performance by evaluating the short-term and long-term effectiveness of MBSR, as an individual-focused intervention, on teachers in secondary vocational schools, as an example of a possible application area. It will look at the effects of the intervention on teachers’ mental health (mindfulness as the primary outcome), work performance, work-related perceptions (job demands and job resources), and personal competencies. In addition, it will investigate the effects of a participatory, preventive, organisational health intervention (i.e., a participatory action approach) that targets and engages teachers in a specific course. We hypothesise that participating in the organisational health intervention will positively influence occupational self-efficacy. The application of the organisational health intervention will generate tailored work solutions that may positively influence the balance between job demands (work pressure, work-life balance) and job resources (autonomy, feedback, relationships) for all teachers in schools.

For this study, we will use a conceptual model (see Fig. 1) inspired by the job demands-resources (JD-R) model [6, 8] and the literature on mindfulness [19, 41‐43]. The JD-R model and the literature on mindfulness present two different but complementary points of view on work stress. The original JD-R model has been expanded to include personal resources, aspects of the self, referring to one’s ability to successfully influence the environment. Examples are self-efficacy, emotional stability, extraversion, and resilience [44, 45]. Both the original and expanded model suggest that job characteristics (i.e., job demands and job resources) can influence work stress via two processes. The first process was referred to by Demerouti et al. [6] as the health impairment process, in which high job demands exhaust workers’ mental and physical resources and may therefore lead to a depletion of energy, exhaustion, health problems, and, eventually, premature retirement from their profession. The second process implies a motivational process: job resources have motivational potential that is either intrinsic (because they foster growth, learning, and development) or extrinsic (because they are instrumental in achieving work goals) and lead to positive work outcomes [6]. Job resources and personal resources can buffer the effects of the job demands [6].

Mindfulness, the primary outcome in our study, can reduce stress by separating work characteristics from employees’ reactions to them. This enables the individual to become aware of the difference between observation and interpretation [46].

Figure 1 depicts the conceptual model, including the processes mentioned above and how these relate to mental health outcomes. The right-hand side of the conceptual model presents the study’s secondary outcome variables: mental health outcomes (mindfulness, burnout, stress, sleep quality complaints, positive and negative emotions at work, work engagement, perceived general health, organisational commitment) and work performance outcomes (work performance and work behaviour, absenteeism) [19].

The relationships between the two interventions (MBSR and MBSR with an additional organisational health intervention), on the one hand, and the two clusters of outcome variables, on the other hand, are mediated by two clusters of process variables, which are presented in the middle of the model. The first cluster contains personal competencies that represent the personal resource outcomes resulting from the interventions (occupational self-efficacy, taking distance, as the opposite of worry) [41‐43]. The second cluster contains secondary outcome variables: work-related perceptions that refer to how an individual worker experiences work characteristics. In line with the JD-R model, we make a distinction between job demands (work pressure, emotional demands, work-life balance) and job resources (autonomy, feedback from colleagues and superiors, relationship with colleagues, relationship with superiors, relationship with students).

We assume that the five-factor-model of personality, that is the Big Five [47], especially the factors of extraversion and openness, can be expected to positively moderate mental health and work performance. The Big Five consists of five personality characteristics or traits, that are fixed and cannot be developed, in contrast to personal competencies.

The proposed study is a cluster randomised controlled trial (CRCT) that uses a mixed-methods design (quantitative and qualitative, online questionnaire, telephone and face-to-face interviews) and contains four measurement time points (see Fig. 2). The Ethics Committee Practice based Research of het HAN University of Applied Sciences (ECPR) and the Medical Ethics Committee (METC) of Radboud University Medical Centre, both located in Nijmegen, the Netherlands, approved the research proposal (Registration number ACPO 07.12/15; File number CMO: 2019–5266). Both committees stated that the research complied with the requirements of ethical conduct of research as set out in the national Code of Conduct for Scientific Integrity in the Netherlands and that it fulfilled the criteria of the Declaration of Helsinki on Ethical Principles for Medical Research Involving Human Subjects. The study will be carried out in the Netherlands in full compliance with the applicable rules concerning the review of research ethics committees. Participation is voluntary and participants can withdraw at any moment with no consequences. The study title given to the potential participants and other stakeholders is “Mindfulness and job satisfaction of teachers in secondary vocational schools”. Participants will sign informed consent forms before participating in this study. They will be asked if they agree to use of their data should they choose to withdraw from the trial. This trial does not involve collecting biological specimens for storage.

Significant deviations from the protocol will be documented using a breach report form and will be sent to the funder NWO and to the ethics committees. The protocol in the trial register will be updated. A Standard Protocol Items Recommendations for Interventional Trials (SPIRIT 2013) checklist (see Additional file 1) and figure (see Table 1) are provided.

Study participants will be recruited from the teaching staff at no fewer than three secondary vocational schools. When a secondary vocational school agrees to participate, we will recruit potential participants from three programmes (Care, Technology, and Economy) using e-mail, posters, flyers, and each school’s intranet. The researchers, Human Resources (HR) consultants, and supervisors will inform potential participants about the research project.

Respondents who are willing to participate will be screened in terms of the eligibility criteria by the first author (MJ) (see Table 2). Eligible candidates will receive an information letter about the project. This letter includes the information as approved by the ECPR and the METC and the informed consent letter. One week before the start of the interventions, cluster randomisation will be conducted and the participants will be informed about their assignment to one of the intervention groups or the waiting list group.

A CRCT is a randomised controlled trial in which groups of subjects (i.e., schools) are randomised rather than individual subjects (i.e., teachers) [48]. Cluster randomisation will be performed at the school level. This will provide the researchers with the opportunity to study the effects of an additional organisational health intervention that cannot be directed towards selected individuals (i.e., teachers) and to control for “contamination” across individuals (i.e., the effects on one teacher may influence the effects on another teacher in the same course) [49]. In the first secondary vocational school (known as an MBO in Dutch), participating teachers from one course (Care, Technology, or Economy) will be assigned to intervention group 1 (IG 1: MBSR), teachers from another course will be assigned to intervention group 2 (IG 2: MBSR and an additional organisational health intervention), and teachers from the third course will be assigned to the waiting list group (WG). The allocation will be different at each school (see Table 3). A researcher who is not involved in assigning courses/participants to the groups will prepare concealed, consecutively numbered, sealed opaque envelopes. Every envelope will contain a paper indicating the treatment assignment at school level (type 1, 2 or 3). The MBO schools will receive their envelopes from a researcher who is unaware of the randomisation sequence. The MBO schools can open the envelope in the presence of the researcher and the researcher will be informed about the treatment assignment.

There is little evidence of harmful effects of MBSR [19]. Also, we will recruit participants from a healthy target population, therefore, there will be no special criteria for discontinuing or modifying the allocated interventions. In the case of (serious) adverse events and harms from the intervention, the participant concerned will be referred to an occupational health professional and the project management group, consisting of YH, BVdH, PP and JE, the funder NWO and the ethics committees will be informed. A final decision to terminate the trial will be made by the project management group, deliberating at least every 6 weeks or more, when necessary.

The trial conduct will be audited by an annual evaluation report for the funding organisation NWO. The report is also available for the ethics committees and the project management group, which will discuss the progress of the trial every 6 weeks. YH and JE will be responsible for the daily supervision of the trial. The implementation of the interventions and the data collection will be strictly separated.

All study participants will be asked to complete an online questionnaire on a secured website before the start of the intervention(s) (the starting date of the study is different for each school) (T₀). After completing the questionnaire, the participating schools will be randomly assigned (type 1, 2, or 3; see Table 3), meaning that participants will take part in IG 1, IG 2, or WG depending on the course where they are working. At T₀, the first author will conduct 10-min telephone interviews with at least 12 participants from IG 1 and 12 participants from IG 2 about their expectations of the interventions. All participants will receive the other three follow-up questionnaires on a secured website after the MBSR training (T₁), 3 months later (T₂), and 9 months after the MBSR training (T₃). The first author will conduct face-to-face interviews with at least 12 participants from IG 1 and 12 participants from IG 2 at T₁ and T_3. At T₁, some members of the participatory group that will be involved in the organisational health intervention - excluding the teachers participating in IG 2 (e.g., a superior, an HR consultant, and the director of the programme) - will be interviewed about the process and effects of the organisational health intervention.

Participants in IG 1 and IG 2 should attend at least four of the nine MBSR sessions, because Bear et al. [50] revealed that structural changes in perceived stress did not occur until after four MBSR sessions [19]. Participants in WG will attend a MBSR programme one year later.

The collected data will be stored on a secure disk to ensure confidentiality. Not the researcher (MJ) but an independent external organisation, assigning encrypted numbers to the participants, will collect the data. The researcher (MJ) cannot link the numbers and the participants. Only the researcher (MJ), the members of the project management group and a methodologist (HK) will have access to the data.

Measures of the baseline characteristics of the participants are gender, age (years), family situation, level of education, nature of employment (course Care, Technology, or Economy), years of work experience, and number of working days and hours.

A process evaluation will be conducted to explore working mechanisms and possible barriers to MBSR in this population. The process evaluation of MBSR will be conducted using both quantitative (online questionnaire, primarily questions about experiences with the MBSR training) and qualitative measurements (semi-structured interviews). All participants will receive the online questionnaire at T₀, T₁, T₂, and T₃. A selection of the participants in IG 1 and IG 2 will be interviewed at T₀, T₁, and T₃. The interview at T₀, lasting 10 min, will be conducted by telephone and will be focused on expectations about MBSR. The face-to-face interview at T₁, lasting 25– 35 min, will be about experiences during the MBSR training and its short-term effects. The face-to-face interview at T₃, lasting approximately 25–35 min, will be focused on long-term effects. All interviews will be recorded, fully transcribed, and anonymised. A deductive qualitative analysis will be performed, because of the availability of a focused main research question and a conceptual model [75, 76]. The interviews can provide valuable information about the working mechanisms and possible barriers of the MBSR training.

A process evaluation of the additional organisational health intervention will be performed to assess the requirements/conditions for successful implementation, based on a simplified version of the theoretical framework presented by Nielsen and Randall [57, 77]. These researchers indicate that a process evaluation is important because the implementation process can moderate or mediate the potential effects of the intervention on health and well-being [57, 77]. Successful implementation is a prerequisite for exposure to the intervention and therefore for entailing possible health effects. The framework, which enables us to link intervention processes to intervention outcomes, will be applied to qualitatively appraise 3 themes of process components: (1) intervention design and implementation, determining the maximum level of intervention exposure; (2) intervention context; and (3) participants’ mental models [57, 77]. The process components of themes 2 and 3 may mediate or moderate the link between any intervention exposure and intervention effects [77]. Table 6 lists the themes and requirements/process components for successful implementation that will be assessed in the semi-structured interviews (T₁). Applying the framework will help us to understand why the implementation process was successful or not [57].

The process evaluation will be conducted using semi-structured interviews. A selection of participants in IG 2 (taking into account participating in the participatory group or not) and of other participatory group members who are not participating in the MBSR training (e.g., teachers not participating in the MBSR training, or supervisor, director, HR consultant, work council member, trade union member) will be interviewed at T₀, T₁, and T₃.

A power analysis (G*Power; version 3.9.1.4) revealed that a sample size of 22 participants in each group (IG 1; IG 2; WG), with at least two repeated measurements would enable detection of a medium effect size (d = 0.50) [78], with power of 0.95 and alpha of 0.05. A total sample size of 66 participants is therefore required.

Participants, trainers, facilitator, and researchers cannot be blinded to their assigned intervention after cluster randomisation. All participants have to fill in the online questionnaire at home or at work, excluding the influence of the researcher. The developer of the online questionnaire will collect the data and provide the anonymous data to the researcher. The researcher will analyse the data blinded to the assigned intervention.

Baseline characteristics of participants will be presented as means and standard deviations (SDs) for metric variables, and as frequencies and percentages for categorical variables. The outcomes of the questionnaires will be compared at baseline (T₀), immediately after the intervention(s) (T₁), 3 months later (T₂), and 9 months after the intervention(s) (T₃). All analyses will be conducted according to the intention-to-treat (ITT) principle. ITT analysis, based on the initial treatment allocation and not on the treatment eventually received, will avoid the effects of drop-out, and as such we prevent breaking the random allocation to the intervention groups [79]. Per-protocol (PP) analyses in the treatment-adherent sample (i.e., participants in IG 1 and IG 2 have to attend at least four of the nine MBSR sessions, and participants in WG cannot attend a MBSR programme or stress reduction training) will also be performed. The aim of PP analysis is to assess the effects of MBSR and the additional organisational health intervention under optimal conditions: what is the effect if participants are fully compliant [80]? Therefore, drop-outs need to be excluded from any PP analysis.

The quantitative short-term and long-term effects of MBSR and the additional organisational health intervention (differences between T_0, T₁, T₂, and T₃) will be examined using longitudinal regression analysis (generalised estimating equations, GEE, or mixed models), which is fit to analyse longitudinal/clustered data in clinical trials [81] or repeated-measures designs (general linear model, GLM) [82]. The baseline values of outcomes (T₀) of the three groups (IG 1; IG 2; WG) will be defined as independent variables, while the outcomes on the follow-up measurements (T₁, T₂, T₃) will be treated as dependent variables. Correction of confounding variables will be applied.

All statistical analyses will be conducted using IBM SPSS Statistics, version 25. The level of significance will be set at 0.05. The analysis of the qualitative data, collected by the semi-structured interviews at T₀, T₁, and T₃, will be deductive [− 76], and will be conducted by means of using ATLAS.ti [83].

Results of the trial will be communicated by scientific articles in open access journals, management letters for participants and non-participants of Dutch secondary vocational schools, and articles for professional magazines intended for occupational health professionals.

The trial is funded for 5 years. The first participants were randomly assigned in September 2016. Final outcome assessments will be completed in June 2020. This is the first protocol version (31 August 2015).