Most conscientious HR leaders today understand the importance of wellness programs for employee stress management. Reducing workplace stress contributes to long-term employee happiness, worker retention and productivity, leading to a new trend in more comprehensive mind and body wellness initiatives.
The problem that many HR pros face? Convincing upper management that programs especially focused on stress management are not just “nice to haves” but rather, essential tools to improve both employee health and the health of the business. Stress leads to higher healthcare costs for businesses, lower productivity among employees, and lower employee morale, leading to lower retention and associated hiring costs. In short: Stress in the workplace is a big deal, and it’s worthwhile for the C-suite to take notice.
Higher employee stress leads to higher healthcare costs
A busy CEO or CFO might scoff at team-building exercises or claims that making employees “feel good” is important, but employee wellbeing is much more than that–and the numbers don’t lie: Unhappy, stressed-out workers are often less healthy and more expensive!
There are hundreds of scientific studies out there to support this. Here are some sobering facts that should interest even the most skeptical C-level executive, all directly related to the high costs of employee healthcare.
Preventing high levels of stress at work and facilitating new habits to help employees to better manage stress are goals of a quality stress management program. Thinking ahead and getting such a program in place now can help prevent serious health conditions, prevent employee burnout and lower healthcare costs, saving your company more money in the long run.
Higher employee stress leads to lower productivity
When we feel stressed, it can be hard for employees to get through the day, let alone do their best work. Multiply that foggy or anxious feeling of stress across many workers and departments, and you’ve got a serious barrier to productivity throughout your company.
There’s a lot of scientific evidence out that that makes the connection between stress and productivity crystal clear. Beyond the enormous cost of healthcare, here are some key results we’ve found:
High stress leads to increased HR costs
Last but not least, stress is a huge factor in lowered employee morale and happiness at work. Studies have shown that when employees are burdened with high levels of stress, their level of engagement at work is directly affected. This in turn leads to higher business costs, since disengaged or low-morale employees are not performing at their best, and when they leave, replacing them is often difficult and expensive. We pulled some of the most telling studies on the cost of employee disengagement here:
Including stress reduction in wellness programs can help
The destructive power of stress in the workplace can’t be overstated. From raised healthcare costs to lost productivity, to expensive employee turnover, there are plenty of reasons for HR leaders and C-suite executives to take action.
Stress leads to all manner of workplace ills, such as lowered team perspective, lowered desire to help others or solve problems, decreased motivation and employee disengagement. A King’s College London study indicates that not only are low-morale employees not working to their potential, they are more likely to engage in counterproductive behavior, further spreading that low morale like a virus.
The good news is that the current new trend in comprehensive, holistic mind and body wellness initiatives has meant that there are a great deal more options than even just a couple of years ago. Now, HR leaders have some choices as to what type of stress reduction programs to offer their employees. Investing in employee wellness programs can increase overall employee happiness and workplace harmony, which leads to more engaged workers, less turnover, and many dollars saved.
Do you offer a stress reduction program at your company? We’d love to hear about what has worked for your employees, and share what we’ve seen work. Email us or simply visit our Contact page to continue the discussion. Or, click the button to request a demo of our program.
Population and sample size
Leaders
The target group are leaders of small and medium-sized enterprises (SMEs). We define SMEs as enterprises with up to 500 employees [50, 51] without regarding the annual sales. Leaders of all genders aged between 18 and 65 years are allowed to participate as long as they are not about to retire within the next year. A sample size of n = 100 participants in the intervention group and n = 100 participants in the waiting control group is planned (N = 200 participants).
The sample size calculation was performed using G*Power3.1. Based on a 2 × 3-mixed design (factor 1 (group): intervention or waiting control group; factor 2 (point in time): baseline, 6 months and 12 months later) the required sample size for a MANOVA with measurement repetition was estimated using the following predefined parameters: mean effect size f = .25 (Ω = 0.06), α-error = .05, test strength 1-β = .80. This results in a required sample size of n = 158 participants (n = 59 per group). The calculation of the required sample size was based on the procedure proposed by Prajapati et al. (2010) to calculate the required sample sizes for all possible effects (within, between, interaction) of the present design and to use the largest calculated sample size. Due to the high effort of the outcome criteria to be collected (e.g., saliva samples, 24-h heart rate variability) a high lost-to-follow-up as well as numerous missing values can be expected. We therefore calculate with a total failure rate of 21% (lost-to-follow-up and due to missing values), resulting in a required sample size of N = 200 (n = 100 per group at T0).
Recruitment is carried out in cooperation with AOK Baden-Württemberg, IHK Ulm, IHK Reutlingen, Offensive Mittelstand, Handwerkskammer Ulm, and Südwestmetall. Flyers and advertising texts will be sent via email through suitable channels and presented at suitable events (face-to-face and online). Leaders who want to participate in this study have to be aged between 18 and 64 years, need to sign the informed consent (see appendix), and must have sufficient German language skills. Due to the variety and size of the above mentioned cooperation partners, the achievement of the targeted sample size can be considered as realistic. Moreover, leaders will be reminded by email to take part in the follow-up measures. As the participation is voluntary, consent may be withdrawn without any reason. The protocol was approved by the institutional ethics committee.
Employees
In addition to the leaders participating in the intervention, we also aim to at surveying their employees in order to gather data on the indirect effects of the intervention. Thus, leaders will be informed in advance about the additionally planned and voluntary survey of their employees. Upon prior written consent, leaders will receive information material in order to inform their employees adequately about the study and their ways to participate: (1) qualitative serial interview at T0 and T2 (only intervention group) and/or (2) online survey at T0, T1 and T2. Due to the so far unknown team size, a sample size calculation pertaining to the online survey cannot be performed. On average, we expect three employees per leader to fill in the questionnaires, resulting in a total of N = 300 employees per group. Regarding the qualitative serial interviews, we aim to interview N = 25 employees (intervention group only) at T0. Thus, we consider a conservative drop-out rate of 40% with the aim of being able to interview N = 15 employees at T2.
The randomization of employees is based on the group allocation of their leaders (convenience sampling). Employees of leaders who want to participate in this study have to be aged between 18 and 65 years, need to sign the informed consent, and must have sufficient German language skills in order to complete the online questionnaire or to be interviewed, respectively.
Although leaders can decide on their own whether to involve their employees or not, we will encourage leaders to do so by highlighting the importance of employees’ inclusion. Moreover, we offer anonymous employee summary reports after T2 to each leader, if the following conditions are met: (1) Their employees agree to the anonymous disclosure of results in a summary form (last question in online survey) and (2) ≥ 5 employees per leader provide data.
First, employees will be informed by their leader, preferably in 1-on-1 or small group meetings. Here, employees will also receive written information about the conditions of participation. Second, direct contact is established between employees and study team in order to ask further questions or gain additional information, if necessary. Following verbal and written information, employees can decide voluntarily whether to participate. All participants at T0 are going to be reminded by email to participate again at the follow-up measures. As the participation is voluntary, consent may be withdrawn without reasons that need to be disclosed.
Intervention
The implemented stress management training is based on the successfully conducted intervention of the MAN-GO project [48]. It is going to be adapted based on the results of a previous (and still ongoing) needs analysis concerning format and content. Depending on the results, it will most likely obtain the following modules:
-
Stress: effects of stress, individual stress symptoms, mentalization ability under stress
-
Work stress models: gratification model, demand-control-model, organizational fairness
-
Leadership: stress-preventive, health-oriented, and relational leadership, leaders as role models, new demands on leaders based on COVID-19 and the digitalization
-
Short term relief: brooding strategies, attention guiding
-
Conflict management: dealing with (emotional) conflicts and catch-22 situations at the workplace
-
Relaxation: progressive muscle relaxation, 1-min meditation
-
Resources: sweet spot imagination, social support, 10 fingers
-
Case work: working on cases brought in by participants, approach based on Balint groups and collegial case advise
-
Optional additive online modules: impact of stress on cardiovascular diseases, further stress symptoms
Trainings are going to take place in groups with 8–12 leaders and are held by two trainers. As far as possible in the context of COVID-19, they should be carried out in persona. Concerning the didactical frame, the trainings consist of a mixture of lectures, open discussions as well as single, peer, and group work. Trainings will take place in Ulm and surrounding as well as in Tuebingen and surrounding.
Outcome variables
All measurements will take place in Germany. More specifically, participants are recruited in Ulm and Tuebingen and their surroundings. Participants are allowed to answer all questions from any suitable place, as questionnaires and further items are provided online. For the physiological measures (leaders only), appointments in Ulm and Tuebingen are going to be arranged.
Leaders
Leaders are going to be asked to complete single items, 9 questionnaires and three physiological measures.
Single items
Sociodemographic data. Participants are going to be asked for their gender, age, marital status, children living in the household and the highest educational level attained. Furthermore, the monthly net income can be entered voluntarily and the number of personal dependants.
Work-related data. Regarding the professional situation, it is asked to indicate the branch, current position, need to do shift work and number of employees to be supervised. Furthermore, the average working time and break times as well as unpaid overtime are enquired.
Health-related data. Days of incapacity to work in the last 3 or 12 months, activity in everyday working life, weight and health behaviour (cigarette and alcohol consumption) are going to be recorded.
Questionnaires
All subsequent questionnaires are collected at three measurement times (T0, T1, T2). The target outcome is the change in the intervention group compared to the waiting control group over time.
Perceived Stress-Reactivity-Scale [52]. The PSRS assesses perceived stress reactivity on 5 scales: reactivity to social assessment (5 items, “If I am wrongly criticized by others …”), reactivity to failure (4 items, e.g., “If I did something wrong …”), reactivity to social conflict (5 items, e.g., “When I argued with other people …“), reactivity to work overload (5 items, e.g., “When I have little time for my work …”) and prolonged reactivity (4 items, e.g., “When I have free time after strenuous work …“). In addition, an overall score can also be calculated. All items are answered by agreeing with one of three default answers. Scale reliability is sufficiently high and varies between α = .79 and α = .82 with α = .89 for the total score (Cronbach’s Alpha).
Effort Reward Imbalance Questionnaire (ERI, [53]). The ERI conducts self-assessments of psychosocial workloads on three scales: exertion (6 items, e.g., “I am often interrupted and disturbed in my work”), reward (11 items, e.g., “I get the recognition I deserve from my superiors”) and over-commitment (6 items, e.g., “Those who are closest to me say I sacrifice myself too much for my job”). Reward is a multidimensional concept and consists of the following subscales: recognition, pay/progression and job security. The exertion and reward dimensions are assessed with five-point Likert-scale, ranging from “unexposed to adverse condition” (1) to “exposed to adverse conditions and very distressed” (5). Over-commitment is measured with a four-point Likert-scale, ranging from “Strongly disagree” (1) to “Strongly agree” (4). The internal consistency of the scales is sufficiently high and between α = .62 and α = .90 (Cronbach’s Alpha).
Hospital Anxiety and Depression Scale (HADS, Herrmann, 1997). The HADS is a 14-item questionnaire (7 items each for anxiety and depression) for self-assessment of depressive (e.g., “I still enjoy the things I used to enjoy”) and anxiety symptoms (e.g., “Worrying thoughts go through my mind”). All items are answered by agreeing with one of four default answers. Scale reliability is high with α = .86 for anxiety and α = .91 for depression (Cronbach’s Alpha).
Leader-Member Exchange 7 (LMX-7, Graen & Uhl-Blen, 1995). The LMX-7 questionnaire consists of 7 items assessing the relationship between leaders and employees (e.g., “How well do you understand the work-related problems and needs of your employees?”) on a 5-point Likert-Scale. Cronbach’s Alpha is high: α = .84 [54].
Occupational Self-Efficacy Expectations – short scale (Berufliche Selbstwirksamkeits-erwartung, BSWE, Schyns & von Collani, 2014). The BSWE – short scale consists of 8 items measuring occupational self-efficacy expectations (e.g., “When unexpected situations arise at work, I always know how to act”) with a 6-point Likert-Scale ranging from “completely true” (1) to “not true at all” (6). Scale reliability is high with α = .84 (Cronbach’s Alpha).
Psychosocial Safety Climate Questionnaire 4 (PSC-4, [55]). The PSC-4 is a short questionnaire with only 4 items measuring the Psychosocial Safety Climate at work (e.g., “Senior management show support for stress prevention through involvement and commitment”). All items are answered using a 5-point Likert-Scale, ranging from “strongly disagree” (1) to “strongly agree” (5).
Operationalized psychodynamic diagnosis structure questionnaire (OPD-SF, Ehrental et al., 2012). The OPD-SF assesses the Axis 4 (Structure) from the Operationalized Psychodynamic Diagnostic system (OPD). Only two of the five dimensions are used for the purpose of this study: self-awareness (12 items, e.g., “I often have feelings that I cannot understand”) and self-regulation (13 items, e.g., “When I’m very upset, I often act rashly”). Items are measured on a five-point Likert-scale ranging from “strongly disagree” (1) to “strongly agree” (5). Cronbachs α ranges from .88 to 91 (self-awareness) and .82 to .84 (self-regulation) [56].
Client Sociodemographic and Service Receipt Inventory (CSSRI, [57]). The use and costs of health care resources will be estimated by means of the CSSRI adapted for the application in the target population [57].
EuroQol’s questionnaire for health related quality of life (EQ-5D 5 L,[58]). Health-related quality of life will be assessed by means of the EuroQol EQ-5D 5 L. The EQ-5D 5 L defines subjective health states at the following five dimensions Mobility, Self-care, Usual activities, Pain and discomfort, Anxiety and depression at a five-level scale [58]. The individual health states will be evaluated by means of country-specific representative preference based value sets [59].
Physiological measures
Hair cortisol. As a first biological stress parameter, the content of hair cortisol (determined via a 2 cm long hair strand) as an indicator for the hypothalamic-pituitary-adrenal axis is going to be collected. Several thin hair strands will be cut as close as possible to the scalp from the posterior vertex region of the head. For determination of hair cortisol concentrations, the first scalp-near 2 cm segment will be used which is thought to reflect the cumulative cortisol secretion of the past 2 months [60]. Hair wash and cortisol extraction procedures are based on a laboratory protocol by Stalder et al. [61], with minor modifications. In brief, hair samples will be washed twice for 3 min using 3 mL isopropanol. For cortisol extraction, 10 ± .5 mg whole, finely cut hair will be incubated in 1.8 mL methanol for 18 h at room temperature. After incubation 1.6 mL will be transferred in another glass vial. Then, 1.6 mL of the supernatant is evaporated at 50 °C until samples are completely dried. Finally, the samples will be resuspended with 225 μL HPLC ultra-pure water and vortexed for 20 s. For cortisol determination, a commercially available cortisol luminescence immunoassay will be used (LIA; IBL International, a Tecan Group company, Hamburg, Germany).
Saliva alpha-amylase. As a second biological stress parameter, alpha-amylase activity in saliva is measured as basal activity of the sympathetic nervous system [62, 63]. Using SaliCaps (IBL International, a Tecan Group company, Hamburg, Germany), participants will be asked to not swallow for 2 min and then transfer all accumulated saliva in pre-labelled vials. Furthermore, they will be asked to provide information on meals and drinks 2 h prior to saliva collection in order to control for confounding variables. The saliva samples will be stored as cool as possible in the participants’ fridge or freezer and are then at the earliest convenience stored at − 80 °C in freezers of the University. Salvia alpha-amylase activity will be measured using a kinetic colori metric test (for details, see also [64]) and reagents obtained from DiaSys Diagnostic Systems (Holzheim, Germany). Saliva will be diluted 1:400 using 0.9% saline solution. The reagents contain the enzyme alpha-amylase in a specified amount and alpha glucosidase, which converts the substrate ethylidene nitrophenyl to p-nitrophenol. The rate of formation of p-nitrophenol is directly proportional to the samples’ amylase activity and is detected using an absorbance reader at 405 nm (Biotek Synergy HTX, BioTek Instruments, Winooski, VA, USA).
Heart rate variability. A 24-h measurement of heart rate variability is collected using a Faros 180 ECG monitor with a textile belt. Circadian parameters of 24 HRV measures will serve as an indicator of mental health and stress perception ([65, 66]; Williams et al., 2018).
Satisfaction with the training. Participants are going to be asked to evaluate their trainers (8 questions each, 6-point Likert-scale) and the training (12 questions, 6-point Likert-scale) concerning acceptance and satisfaction.
Employees
Regarding the assessment of employees, seven validated questionnaires and various single-item measures will be used. Employees receive an email with a link that leads to the described questionnaires. Depending on how they wish to proceed or what is most convenient for the employees, they can fill it in at home or at work.
Single items
Socio-demographic data. The data collected includes information about gender, age, marital status, the highest educational level attained, the number of years worked under the current leader, organizational tenure and hours of direct contact with leader per week.
Change-related data. Participants will be asked to rate the statement “I have noticed a change in my leader’s behavior after the training.” This item will be assessed on a 5-point Likert scale ranging from 1 “do not agree” to 5 “fully agree”; respectively there will be another option “My leader will take part in the training in 2022”. This item will only be used at T1 and T2.”.
At the end of the survey participants will have to answer two additional questions. At first, they have to indicate their seriousness during participation [67]. Therefore, we will directly ask participants if questions were answered in a serious manner (“Please indicate whether you have been actively involved in the survey, so that we can use your answers for our scientific analysis, or whether you were just clicking through to gain a first impression?”). Second, employees can decide, whether their answers can be used for a de-identified employee summary report, which will be provided for the leaders after the completion of T2.
Questionnaires
Short questionnaire for work analysis (ger.: Kurzfragebogen zur Arbeitsanalyse, KFZA, [68]). Working conditions will be assessed by the KFZA, which consists of eleven scales derived from 26 items. These scales comprise Decision latitude, Variety of work, Task identity, Social support, Teamwork, Workload (qualitative), Workload (quantitative), Interruption of work, Environmental stress, Information and codetermination, Company benefits. Sample items are: “How much influence do you have on what work is assigned to you?” and “I am often under time pressure”. All items are measured on a 5-point Likert scale ranging from 1 “does not apply at all” to 5 “is completely true”. The internal consistency of the scales is set between .40 and .76 (Cronbach’s α); in studies with larger samples Cronbach’s α ranged from .63 to .80 [69].
Cognitive and emotional irritation scale (IS [70]). The IS measures perceived cognitive and emotional strain. Irritation refers to a psychological state of impairment due to the experience of goal-discrepancy. The IS consist of 8 items that have to be answered on a 7-point Likert scale resulting in two first-order factors: First, Rumination (e.g., “Even when I am on vacation, I sometimes have to think about problems at work”) invokes increased efforts in order to achieve goals, whereas the second factor Irritability (e.g., “I get angry easily”) reflects a tendency to lose the incentive to pursue a certain goal. Internal consistency scores range from .85 to .97 for both subscales.
Subjective psychological well-being (WHO-5, Bech et al., 2003). Employees’ well-being will be measured using the five-item World Health Organization Well-Being Index (WHO-5) representing the most widely used questionnaire for measuring this construct [71]. Items are assessed on a 6-point Likert scale, ranging from 0 “at no time” to 5 “all of the time” concerning the last 2 weeks (e.g., “… I have felt calm and relaxed”). The raw score is multiplied by 4 resulting in a final score ranging from 0 to 100 (representing the best possible well-being). The WHO-5 is suitable for depression screening. The split-half reliability (Guttman’s coefficient) is rtt = .87 (Brähler et al., 2007).
Effort Reward Imbalance (ERI, [53]) and Relationship between employees and their managers (LMX-7, Schyns & Paul, 2014). Both questionnaires are already described in section 2.3.1.
Serial qualitative interviews in the intervention group. We will conduct serial qualitative interviews at T0 and T2 as interviewing participants on multiple occasions is useful in understanding longitudinal change, thus variation in behaviours or circumstances over time [72]. Particularly, semi-structured interviews with employees will be used in order to understand the delivery and impact of indirect outcomes on employees following their leaders‘participation in the intervention. Thus, employees will be asked to describe the leadership behaviour, the leader’s stress level, leadership styles, working conditions and self-perceived health in a narrative style. Following a qualitative content analysis [73] the preliminary category system developed at T0 will be revisited at T2 in order to identify potential changes.
Timeline
The total trial duration is 12 months for the intervention group, and 18 months for the control group. Baseline data of all leaders and employees is going to be collected in the beginning (spring 2021). The intervention group participates in the stress management training in the weeks thereafter (duration: 1.5 days) followed by two booster sessions 3 and 6 months later, each lasting for 180 min. All participants are going to be measured again after 6 and 12 months. Afterwards, participants of the waiting control group are going to participate in the training, followed by the two booster sessions 3 and 6 month thereafter.
Assignment of interventions
Each single participant will randomly be allocated to the intervention or waiting control group. A study staff member uses computer-generated random numbers for the allocation of the participants. It is not possible to blind participants about their group affiliation, due to the different intervention times.
Data collection, management, and analysis
Data collection. Leaders’ sociodemographic and work-related data will be assessed at baseline before the intervention. The self-report measures of health-related indices, perceived stress (PSRS-23 [52]), work satisfaction (ERI [53]), symptoms of depression and anxiety (HADS-D [74]) and well-being (WHO-5 [75]) will be assessed at three different points of time: at baseline, at a 6-month-follow-up post-intervention and a 1-year-follow-up. The questionnaires can be filled in at home or during the clinic visits, where the physiological data will be recorded across the three time points. All questionnaire data will be assessed online via the encrypted German Unipark servers. Physiological measures, i.e. hair cortisol, saliva samples for alpha-amylase activity and blood pressure will be collected by a trained study nurse on site. To assess the 24-h-heart rate variability, participants will wear a belt for 24 h after their appointment at the clinic. Employees will receive online questionnaires that can be filled in during work in order to rate their individual stress levels and their perceived interaction with their leaders. Data will be transmitted electronically across encrypted German servers provided by the online assessment tool Unipark.
Data management
All data will be stored pseudonymously. All identifying information of participants will be removed from the datasets. Data will be stored locally on the clinic networks that underlie a strict security protocol to store sensitive clinical data. If data will need to be transferred electronically, it will be password encrypted. Documents with identifying information, such as signed consent to participation will be stored securely and separately from the study data. Only project staff members will have access to the data collected. The staff members are subject to the obligation of secrecy.
Statistical methods
Data will be analysed using the statistic software IBM SPSS Statistics 27. Descriptive analyses will be conducted after the baseline. A multivariate analysis of variance (MANOVA) with all relevant outcome parameters as specified above will be calculated in order to test the effects of the intervention on different dependent variables. Potential covariates will be included in the analysis. Therefore, if relevant covariates occur, a MANCOVA will be performed instead of the MANOVA. A covariate will be identified as such, if a demographic variable will correlate with the outcome variable or if there is a significant difference between the intervention group and the waitlist control group across this variable.
To evaluate the effects of the stress management intervention more specifically, changes in outcome variables will be analysed by two-factorial ANOVA with repeated measures and group (intervention vs. waiting control group) as a between-person effect. The core hypothesis is that perceived stress and physiological stress parameters of leaders will significantly decrease from baseline to follow-up measurements for participants in the intervention group. Thus, we will test the interaction effect of time x group for the different outcome variables. Where necessary, Bonferroni correction will be applied to avoid inflation of the type one error by multiple comparisons.
Health economic evaluation
Incremental cost-effectiveness analyses (ICEA) by means of the net-benefit approach [76, 77] will be performed from the perspective of SME companies and from the societal perspective. For the SME perspective the incremental costs of the intervention compared to non-intervention incurred by the companies will be related to the incremental reduction in sick-leave days. For the societal perspective the incremental total costs of illness including the direct costs of health care resources and the indirect costs due to productivity losses will be related to the incremental gain in quality adjusted life years (QALYs) [76, 77].
Point estimates of incremental costs effectiveness ratios (ICER) will be computed for both perspectives. Stochastic uncertainty of the ICER will be estimated by means of non-parametric bootstrapping with 4000 replications [76, 78]. ICER will be interpreted on the basis of the cost-effectiveness plane [76]. The significance of the ICER will be estimated on the basis of the cost-effectiveness acceptability curve [76].
Results of the ICEA from the company perspective will provide the necessary maximum willingness to pay for the prevention of one sick leave day at which the intervention is cost-effective at the company level with a probability of 95%. Results of the ICEA from the societal perspective will provide the necessary maximum willingness to pay for the gain of one additional life year in full health (QALY) at which the intervention is cost-effective at the level of the German national economy with a probability of 95% [76].
Handling of missing and spurious data and drop outs
We do not expect item-wise missing values as all questionnaires will be assembled digitally with notifications if items are missing. However, in the case of missing items for questionnaires the missing item score will be substituted by the mean item score of the remaining questionnaire items.
As participating leaders can withdraw from the physiological assessments, we expect some missing data for the physiological stress measures. In physiological stress assessments, measurement errors may occur. Especially for heart rate variability, which will be measured over a period of 24 h without surveillance of clinical personnel, measurement errors may occur. Thus, all data will be checked for plausibility and carefully tested against potential measurement errors. Participants with missing or spurious data will remain in the sample and will only be excluded for those analyses that involve the missing data.
As the study has a longitudinal design that extends over 1 year in time, some dropouts will be inevitable. We will, however, endeavour to retain as many participants of the original sample and will try to maximize compliance with the study’s data assessments. We accounted for some dropouts in our power analyses. Participants who drop out after t1 (the first follow-up measurement) will be included in the data analyses up to this point in time. An intention-to-treat analysis will be conducted to evaluate whether the sub-sample of dropouts significantly differs from the adherent sample.
Monitoring
Data monitoring. A data monitoring committee is usually formed to monitor the data for potential adverse effects the intervention has on the participants. Since the intervention has no known risk, no data monitoring committee is formed. Furthermore, no interim analysis is conducted and no stopping guidelines are formulated.
Harms. The intervention has a low risk of adverse effects. Therefore, no potential harm is recorded. During the intervention sessions participants are monitored by the trainers. If there are unexpected adverse effects, the present trainer takes care of the concerned participant after the training. With regard to the measurements, no risk or complications are expected from measuring blood pressure and recording HRV by using electrodes. If there is a lot of hair on the chest, it may be necessary to remove the hair. However, no permanent damage is to be expected. All experimental procedures used in the study are harmless to health. The only possible risk of becoming aware of sensitive personal data is minimized by pseudonymization and access restrictions.
Auditing. No auditing is planned. An ongoing process evaluation takes place within the respective study sites and in the study network as whole. Corrective action will be taken when necessary.
Ethics and dissemination
Consent or assent
Written informed consent is a prerequisite in order to participate. All employees and leaders interested in participation will receive verbal and written information from the study staff prior to the start of the study.
Confidentiality
In order to preserve confidentiality, all quantitative data of leaders and their employees will be pseudonymized by a self-generated identification code. The survey data will be stored on password-protected data storage devices. The written informed consent form will be locked in cabinets, respectively. Thus, the identification of an individual person by linking a participant’s name to an identification code will be prevented. Only authorized study team members will have access to original data.
Qualitative interviews with employees will be audio recorded, with the absence of mentioning personal data during interviews. After finishing the audio recording, socio-demographic data of the interview partners will be collected (gender, age, the highest educational level attained, the number of years worked under the current leader, organizational tenure, hours of direct contact with leader per week and number of employees working in the company). Following the data collection phase, the audio recordings will be transcribed guaranteeing complete anonymity by removing last personally identifiable information. The audio recordings of the individual interviews will be erased immediately after transcription. After expiry of the obligation to preserve records, the transcripts will also be deleted. Transcripts will be stored in separate data storage devices with password protection, therefore anonymized data cannot be linked to signed consent forms, original audio files (e.g., before deletion), or sociodemographic data.
Statutory provisions will guide the collection, evaluation and use of study related quantitative and qualitative data (e.g., participants‘voluntary written informed consent before participation).
Access to data
During the process of data collection, the database is only accessible to authorized study members bound to secrecy. After database closure, access rights for biometricians regarding the physiological parameters will be granted.
Dissemination policy
Study results will be made available to the scientific community, policy makers, health insurances and the SME sector, pursuing a wide dissemination. If desired, study participants will receive a summary of the general study’s research findings. All study results reported will be anonymous.
