This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Mental Health, is properly cited. The complete bibliographic information, a link to the original publication on http://mental.jmir.org/, as well as this copyright and license information must be included.
To counteract the negative impact of mental health problems on business, organizations are increasingly investing in mental health intervention measures. However, those services are often underused, which, to a great extent, can be attributed to fear of stigmatization. Nevertheless, so far only a few workplace interventions have specifically targeted stigma, and evidence on their effectiveness is limited.
The objective of this study was to develop and evaluate a digital game-based training program for managers to promote employee mental health and reduce mental illness stigma at work.
We describe the empirical development of Leadership Training in Mental Health Promotion (LMHP), a digital game-based training program for leaders. A 1-group pre-post design and a 3-month follow-up were used for training evaluation. We applied multilevel growth models to investigate change over time in the dependent variables knowledge, attitudes, self-efficacy, and intentions to promote employee mental health in 48 managers of a global enterprise in the United Kingdom. Participants were mainly male (44/48, 92%) and ranged in age from 32 to 58 (mean 46.0, SD 7.2) years.
We found a positive impact of the Web-based training program on managers’ knowledge of mental health and mental illness (
Results provide first evidence of the effectiveness of LMHP to positively affect managers’ skills to promote employee mental health at work. Furthermore, the high rate of participation in LMHP (48/54, 89%) supports the use of digital game-based interventions to increase user engagement and user experience in mental health programs at work.
Due to their high prevalence (1 in 4) [
To counteract the negative impact of mental health problems on business, organizations are increasingly investing in mental health promotion, prevention, and intervention efforts [
However, there are a few drawbacks worth discussing with regard to the current practice of workplace mental health promotion. First, most interventions aiming to promote employee mental health focus on the employee level (such as in stress management) while neglecting the organizational level (working conditions) [
Stigma is defined as (1) the lack of knowledge of mental health problems and treatments, (2) prejudicial attitudes, and (3) the lack of supportive behavior, or anticipated or real acts of discrimination against people with mental health problems [
While the majority of stigma reduction programs targeted the general population—for example, in public health campaigns—there is growing interest in the effectiveness of workplace antistigma interventions [
The dissemination of digital interventions, however, could be a powerful strategy to facilitate widespread behavioral and cultural change in organizations [
Specifically, we hypothesized that our digital game-based intervention, called Leadership Training in Mental Health Promotion (LMHP), would lead to (1) improved mental health knowledge, (2) increased positive attitudes toward people with mental health problems, (3) increased self-efficacy to deal with mental health situations at work, and (4) improved intentions to promote employee mental health at work in managers undertaking the training.
The intervention was developed in a collaborative effort between the department of psychosocial health and well-being of a large global private sector company, which employed around 348,000 employees in more than 100 countries in 2015, and the Chair for Public Health and Health Services Research of Ludwig-Maximilians-University (LMU) in Munich, Germany.
In developing LMHP, we followed a systematic approach similar to intervention mapping [
We developed training content based on a review of workplace training programs on mental health [
While e-learning is well established in larger enterprises, Web-based training in its most common form, animated slidecasts, is losing more and more in attractiveness and acceptance [
To facilitate an innovative and engaging learning experience [
The goal of this pilot study was to evaluate the effectiveness of a digital game-based training program for managers, which we developed to promote employee mental health and reduce mental health-related stigma at work, using a 1-group pre-post design and a 3-month follow-up. The pilot study was carried out at a defined site of the participating organization near Oxford, United Kingdom.
All managers of this site were invited to take part in LMHP and its associated research study. To be included, participants had to be of working age (between 18 and 65 years) and be managing at least one employee at the time of the training. Informed consent was obtained from all individual participants included in the study.
Invitations to participate in LMHP were sent out by email approximately a week in advance of the scheduled Web-based training. This invitation notified participants about the study’s objectives, potential risks, data protection, etc.
Participants were then sent a personal link that allowed (1) participants to give their informed consent to participate in this study, (2) participants to access the training program for a limited time period of 3 weeks, (3) participants to access the pre- and postquestionnaire immediately before (T1) and after (T2) completion of the training, and (4) the researchers to allocate responses at T1, T2, and T3 to an individual. However, the link did not include any information that could be used to identify participants. At T3 (12 weeks after training completion), participants were resent their personal link in order to fill in a follow-up questionnaire to evaluate the first mid-term effects of the intervention.
Any communication about the training initiative (eg, invitations), as well as personal links to training and questionnaires, was sent out via email by a human resources staff member of the participating organization, who was not involved in the study. Questionnaires were completed anonymously online, and responses were tracked and stored safely at the external training provider. The external training provider then replaced participants’ email addresses with a random, unique 3-digit identifier and posted the data back to the researchers at LMU Munich. To increase response rates, the external training provider informed the human resources staff member of the participating organization about any nonresponders so that he could send out reminders. The researchers were never told the names of individual respondents, and the human resources staff member in the participating organization never saw any completed questionnaires or individually identifiable data.
Ethical approval for the study was given by the Ethics Committee of LMU Munich, Germany. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional or national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.
Demographic questions included age, sex, level of education, marital status, whether they currently lived alone, and whether they knew someone with a mental health problem and had been diagnosed with or treated for a mental health problem themselves.
Other outcome measures matched the knowledge, attitudinal, and behavioral dimensions of stigma as defined above. We administered 4 validated instruments. To all of them, a 5-point Likert scale ranging from 1 (“strongly disagree”) to 5 (“strongly agree”) was applied. We calculated global scores on all instruments using sum scores, with higher scores indicating a better outcome, with the exception of stigmatizing attitudes. All measures were administrated at all 3 time points.
We assessed knowledge about mental health problems using the first 6 items, which are related to stigma, of the 12-item Mental Health Knowledge Schedule (MAKS) [
Additionally, we developed a set of 7 quiz questions to test participants’ knowledge on specific training content of LMHP, with 3 answer options, of which 1 was correct. An example item is “Which statement about business costs related to mental disorders is correct?” In this case, sum scores ranged from 0 to 7.
We assessed attitudes in the workplace toward coworkers who may have a mental illness using the 23-item Opening Minds Scale for Workplace Attitudes (OMS-WA), an adapted version of the Opening Minds Scale for Health Care Providers (OMS-HC) [
To assess behavioral change in leaders, we used proxy variables (eg, self-efficacy to deal with mental health situations at work and intentions to promote employee mental health), since in a 3-month period not very many mental health situations are likely to arise at work where leaders could possibly demonstrate actual support. However, prior research found that enhanced intentions and high self-efficacy increase the likelihood that a person will engage in newly learned behaviors [
In this study, we measured self-efficacy with regard to managing employee mental health by a previously adapted version of the 9-item New General Self-Efficacy Scale [
To assess participants’ intentions to promote employee mental health, we used a previously adapted 3-item version of a safety scale designed to assess managers’ safety promotion intentions [
We used descriptive statistics (mean, median, SD) to describe the study population. Multilevel growth models (with random intercept) were applied to investigate change over time in the dependent variables knowledge, attitudes, self-efficacy, and intentions to promote employee mental health [
Taking all formative research described above into consideration, we designed LMHP in a way to train managers in (1) understanding mental health and mental illness, (2) spotting warning signs, (3) taking early and appropriate action, and (4) monitoring and self-monitoring.
The training consisted of one single session, which took between 1.5 and 2 hours to complete, thereby meeting managers’ expectations of a particularly concise and time-efficient training format as expressed during interviews (see formative research described above). The setting was the office hub where, over a virtual time period of 7 weeks, the player was put into the position of a manager. During that time period, it was the manager’s task to supervise a virtual team and manage employee mental health effectively.
Outline of content and psychological constructs covered in the virtual scenarios of the Leadership Training in Mental Health Promotion program.
Scenario | Objective | Knowledge | Attitude | Skills | |
1. | Psychological well-being | Promotion of mental health | Create awareness of the importance of mental health at work and that stress or mental illness affects everyone | Develop more positive attitudes toward promoting mental health at work | Communication and behavioral strategies to ensure that healthy employees stay healthy |
2. | Acute stress | Prevention of mental illness | Create awareness that acute stress can result in psychological as well as physical symptoms | Develop more positive attitudes toward discussing the topic of stress more openly at work and to promote employee mental health | Communication, identification of warning signs, support strategies |
3. | Chronic stress | Prevention of mental illness | Create awareness that persistent stress has severe detrimental effects on the body and the mind and, if not dealt with, can lead to long-term sickness absence | Develop more positive attitudes toward employees with mental health problems with regard to avoidance, work competency, responsibility, and helping | Communication, identification of warning signs, and support and referral strategies |
4. | Mental Illness | Rehabilitation and return to work | Create awareness of common mental health problems and of return-to-work policies and procedures | Develop more positive attitudes toward employees with mental health problems with regard to perceived dangerousness, work competency, responsibility, avoidance, and helping | Communication, planning a successful return to work, workplace accommodations, monitoring, actively counteracting stigma and discrimination, facilitating open discussions |
The virtual team consisted of 4 employees showing diverse psychological profiles; thus, each represented a different mental health scenario likely to appear in real office life. Scenarios contained examples of the promotion of mental health, the prevention of mental illness, and the rehabilitation of employees with common mental health problems such as anxiety or depressive disorders (see
For example, to sensitize managers in the recognition and identification of warning signs, certain hints were placed into the virtual work environment (eg, medication, uneaten lunch, or work piling up on an employee’s desk) that may or may not signal a growing underlying mental imbalance. Once the manager had spotted something unusual or alarming, he or she could choose to engage in a conversation with the respective employee. Different dialogue options were provided to choose from, which were more or less appropriate given the sensitivity of a certain topic. Depending on how the manager behaved, the respective employee chose to either shut down and end the conversation or open up and share further information the manager needed to be able to offer appropriate and effective support.
To ensure continuous learning and improved self-efficacy to manage mental health situations at work, the player was provided with instant feedback regarding his or her actions after the end of each conversation. Furthermore, a video of an actual affected employee of the participating organization sharing his or her experience with burnout was shown automatically to every player. The personal testimonial was presented in a way to counter prominent stereotypes of people with mental health problems and with a strong focus on the road toward recovery and well-being, thus involving many features considered fundamental to reducing stigma [
Next to scenario-based learning, LMHP also offered a mental health toolbox that provided managers with practical information on topics found to be relevant to manage a given scenario successfully. The toolbox was presented in a way to improve managers’ knowledge of mental health and mental illness, improve their attitudes toward employees with mental health problems, and train them in skills to deal with mental health situations at work effectively. Topics of the mental health toolbox focused on 4 main areas: what mental health and mental illness mean, how to recognize signs of mental distress, how to start a conversation, and how to support affected employees effectively (see
Outline of content and psychological constructs covered in the Mental Health Toolbox of the Leadership Training in Mental Health Promotion program.
Focus areas of training | Module | ||
A | Understanding mental health and mental illness | A1 | Mental health affects us all |
A2 | Understanding mental health and mental illness | ||
A3 | Economic impact of mental illness | ||
A4 | Risk factors and treatment of mental disorders | ||
B | Recognizing signs of mental distress | B1 | What is stress? |
B2 | Work-related stressors and resources | ||
B3 | Warning signs | ||
B4 | Common mental disorders at work | ||
C | Starting the conversation | C1 | Stigma: a barrier to help-seeking |
C2 | Communication techniques | ||
C3 | Guidance for leaders | ||
C4 | In-house support services | ||
D | Supporting effectively | D1 | Key role of managers |
D2 | Providing support | ||
D3 | Return to work | ||
D4 | Self-care |
The idea behind the training—for example, the progression of employees’ mental state in scenarios—followed the principles of the mental health continuum model [
In LMHP, we used an adapted version of the mental health continuum model to suit our specific needs. Each phase of this continuum (health, acute stress, chronic stress, and illness) is assigned certain warning signs and recommended actions to take as an affected individual but also as a manager supporting affected employees. In this way, mental health becomes more concrete, which, in turn, facilitates managers’ understanding of mental health and warning signs.
On several occasions during the training, the manager was asked to assess each employee’s mental state along the phases of the mental health continuum model. Afterward, the player was given feedback on an employee’s actual mental state and on other parameters the manager influenced with his or her behavior, such as perceived managerial support or an employee’s willingness to seek professional help. This exercise was designed to improve managers’ self-efficacy in identifying warning signs and to strengthen their intentions to promote employee mental health.
Flow diagram showing progress through the phases of the trial.
Baseline demographic characteristics of the sample population (n=48).
Characteristics | Data | |
Age in years, mean (SD), median | 46.0 (7.2), 45.5 | |
<45.5 years | 24 (50) | |
≥45.5 years | 24 (50) | |
Male | 44 (92) | |
Female | 4 (8) | |
Graduate degree | 11 (23) | |
Bachelor’s degree | 12 (25) | |
Nonuniversity certificate | 13 (27) | |
High school | 10 (21) | |
Less than high school | 2 (4) | |
University degree | 23 (48) | |
Nonuniversity degree | 25 (52) | |
Married | 37 (77) | |
Divorced or separated | 6 (13) | |
Single | 3 (6) | |
Common-law relationship | 2 (4) | |
No | 42 (88) | |
Yes | 5 (10) | |
Prefer not to answer | 1 (2) | |
No | 13 (27) | |
Yes | 30 (63) | |
Prefer not to answer | 5 (10) | |
No | 41 (85) | |
Yes | 5 (10) | |
Prefer not to answer | 2 (4) | |
No | 30 (63) | |
Yes | 8 (17) | |
Missing values | 10 (21) |
aVariables included in multilevel analysis (model C).
Descriptive statistics for respondents who participated at all 3 time pointsa (n=37).
Measures | Wave 0 | Wave 1 | Wave 2 | |||
Mean | SD | Mean | SD | Mean | SD | |
Knowledge (MAKSb) | 22.1 | 2.6 | 24.2 | 2.5 | 24.0 | 2.8 |
Knowledge (quiz) | 4.4 | 1.4 | 5.6 | 1.4 | 4.9 | 1.2 |
Attitude total | 45.9 | 10.7 | 43.1 | 11.5 | 42.3 | 10.3 |
Attitude avoidance | 11.4 | 3.6 | 10.1 | 3.0 | 9.8 | 3.2 |
Attitude dangerousness | 10.5 | 3.0 | 9.3 | 3.3 | 9.1 | 2.7 |
Attitude work | 10.9 | 3.0 | 11.2 | 3.3 | 10.4 | 3.1 |
Attitude help | 8.0 | 1.6 | 8.0 | 2.2 | 8.6 | 2.7 |
Attitude responsibility | 5.0 | 2.0 | 4.5 | 1.6 | 4.4 | 1.7 |
Self-efficacy | 31.5 | 3.6 | 34.7 | 3.4 | 34.2 | 2.9 |
Promotion intentions | 12.2 | 1.3 | 12.4 | 1.2 | 12.3 | 1.2 |
aWave 0, baseline; wave 1, postintervention; wave 2, 3-month follow-up.
bMAKS: Mental Health Knowledge Schedule.
Mixed models (with random intercept) considering knowledge assessed by MAKSa, knowledge assessed by quiz, attitude (total), self-efficacy, and intentions to promote employee mental health as the dependent variable (n=48).
Dependent variable and predictors of change over time | Model A: unconditional means model | Model B: unconditional growth (with time) | Model C: time & age & education | ||||
Parameter estimate (SE) | Parameter estimate (SE) | Parameter estimate (SE) | |||||
Fixed effects | |||||||
Intercept (initial status) | 23.27 (0.324) | <.001 | 21.98 (0.372) | <.001 | 21.84 (0.572) | <.001 | |
Time (rate of change) | |||||||
Wave = 1 | 2.16 (0.335) | <.001 | 2.16 (0.335) | <.001 | |||
Wave = 2 | 1.88 (0.361) | <.001 | 1.87 (0.361) | <.001 | |||
Age | –0.09 (0.641) | ||||||
Education | 0.38 (0.642) | ||||||
Variance components | |||||||
Level 1: within-person (residual) | 4.13 (0.633) | <.001 | 2.65 (0.407) | <.001 | 2.65 (0.407) | <.001 | |
Level 2: in intercept | 3.51 (1.052) | .001 | 3.99 (1.024) | <.001 | 3.95 (1.017) | <.001 | |
Goodness of fit | |||||||
Deviance | 623.88 | 585.60 | 585.23 | ||||
AICb | 629.88 | 595.60 | 599.23 | ||||
BICc | 638.55 | 610.05 | 619.47 | ||||
Fixed effects | |||||||
Intercept (initial status) | 5.01 (0.138) | <.001 | 4.38 (0.191) | <.001 | 4.36 (0.259) | <.001 | |
Time (rate of change) | |||||||
Wave = 1 | 1.36 (0.239) | <.001 | 1.36 (0.239) | <.001 | |||
Wave = 2 | 0.55 (0.256) | .03 | 0.53 (0.256) | .04 | |||
Age | –0.34 (0.263) | ||||||
Education | 0.38 (0.642) | ||||||
Variance components | |||||||
Level 1: within-person (residual) | 1.86 (0.284) | <.001 | 1.36 (0.208) | <.001 | 1.36 (0.208) | <.001 | |
Level 2: in intercept | 0.24 (0.211) | 0.40 (0.197) | .04 | 0.33 (0.185) | |||
Goodness of fit | |||||||
Deviance | 474.48 | 446.59 | 443.09 | ||||
AIC | 480.48 | 456.59 | 457.09 | ||||
BIC | 489.15 | 471.04 | 477.32 | ||||
Fixed effects | |||||||
Intercept (initial status) | 43.77 (1.511) | <.001 | 46.13 (1.633) | <.001 | 47.93 (2.601) | <.001 | |
Time (rate of change) | |||||||
Wave = 1 | –3.49 (1.095) | .002 | –3.49 (1.095) | .002 | |||
Wave = 2 | –4.08 (1.185) | .001 | –4.06 (1.185) | .001 | |||
Age | –1.09 (3.002) | ||||||
Education | –2.64 (3.004) | ||||||
Variance components | |||||||
Level 1: within-person (residual) | 33.47 (5.147) | <.001 | 28.33 (4.356) | <.001 | 28.34 (4.361) | <.001 | |
Level 2: in intercept | 97.211 (22.562) | <.001 | 99.63 (22.644) | <.001 | 97.43 (22.218) | <.001 | |
Goodness of fit | |||||||
Deviance | 949.58 | 935.62 | 934.70 | ||||
AIC | 955.58 | 945.62 | 948.70 | ||||
BIC | 964.26 | 960.07 | 968.93 | ||||
Fixed effects | |||||||
Intercept (initial status) | 33.59 (0.396) | <.001 | 31.54 (0.507) | <.001 | 31.14 (0.742) | <.001 | |
Time (rate of change) | |||||||
Wave = 1 | 3.62 (0.551) | <.001 | 3.62 (0.551) | <.001 | |||
Wave = 2 | 2.78 (0.225) | <.001 | 2.77 (0.592) | <.001 | |||
Age | 0.47 (0.801) | ||||||
Education | 0.36 (0.801) | ||||||
Variance components | |||||||
Level 1: within-person (residual) | 11.28 (1.752) | <.001 | 7.18 (1.113) | <.001 | 7.20 (1.119) | <.001 | |
Level 2: in intercept | 3.41 (1.714) | .046 | 5.16 (1.685) | .002 | 5.03 (1.670) | .003 | |
Goodness of fit | |||||||
Deviance | 728.85 | 691.95 | 691.39 | ||||
AIC | 734.86 | 701.95 | 705.39 | ||||
BIC | 743.53 | 716.40 | 725.62 | ||||
Fixed effects | |||||||
Intercept (initial status) | 12.46 (0.151) | <.001 | 12.31 (0.185) | <.001 | 12.08 (0.269) | <.001 | |
Time (rate of change) | |||||||
Wave = 1 | 0.36 (0.192) | 0.36 (0.192) | |||||
Wave = 2 | 0.08 (0.207) | 0.07 (0.207) | |||||
Age | 0.00 (0.292) | ||||||
Education | 0.48 (0.292) | ||||||
Variance components | |||||||
Level 1: within-person (residual) | 0.91 (0.140) | <.001 | 0.87 (0.135) | <.001 | 0.88 (0.136) | <.001 | |
Level 2: in intercept | 0.76 (0.233) | .001 | 0.76 (0.231) | .001 | 0.70 (0.220) | .001 | |
Goodness of fit | |||||||
Deviance | 421.88 | 418.22 | 415.58 | ||||
AIC | 427.88 | 428.22 | 429.58 | ||||
BIC | 436.55 | 442.67 | 449.81 |
aMAKS: Mental Health Knowledge Schedule.
bAIC: Akaike information criterion.
cBIC: Bayesian information criterion.
Mixed models (with random intercept) considering attitudes regarding avoidance, dangerousness, workability, helping, and responsibility as the dependent variable (n=48).
Dependent variable and predictors of change over time | Model A: unconditional means model | Model B: unconditional growth (with time) | Model C: time & age & education | ||||
Parameter estimate (SE) | Parameter estimate (SE) | Parameter estimate (SE) | |||||
Fixed effects | |||||||
Intercept (initial status) | 10.50 (0.439) | <.001 | 11.44 (0.492) | <.001 | 11.69 (0.773) | <.001 | |
Time (rate of change) | |||||||
Wave = 1 | –1.37 (0.390) | .001 | –1.37 (0.390) | .001 | |||
Wave = 2 | –1.66 (0.422) | <.001 | –1.66 (0.422) | <.001 | |||
Age | –0.39 (0.880) | ||||||
Education | –0.12 (0.881) | ||||||
Variance components | |||||||
Level 1: within-person (residual) | 4.43 (0.681) | <.001 | 3.60 (0.554) | <.001 | 3.60 (0.555) | <.001 | |
Level 2: in intercept | 7.63 (1.926) | <.001 | 8.00 (1.932) | <.001 | 7.95 (1.924) | <.001 | |
Goodness of fit | |||||||
Deviance | 659.03 | 641.77 | 641.55 | ||||
AICa | 665.03 | 651.77 | 655.55 | ||||
BICb | 673.70 | 666.22 | 675.78 | ||||
Fixed effects | |||||||
Intercept (initial status) | 9.72 (0.404) | <.001 | 10.60 (0.440) | <.001 | 11.33 (0.688) | <.001 | |
Time (rate of change) | |||||||
Wave = 1 | –1.32 (0.308) | <.001 | –1.32 (0.308) | <.001 | |||
Wave = 2 | –1.52 (0.333) | <.001 | –1.51 (0.333) | <.001 | |||
Age | –0.40 (0.791) | ||||||
Education | –1.10 (0.792) | ||||||
Variance components | |||||||
Level 1: within-person (residual) | 2.96 (0.454) | <.001 | 2.24 (0.345) | <.001 | 2.25 (0.345) | <.001 | |
Level 2: in intercept | 6.76 (1.615) | <.001 | 7.03 (1.614) | <.001 | 6.67 (1.543) | <.001 | |
Goodness of fit | |||||||
Deviance | 616.80 | 593.42 | 591.23 | ||||
AIC | 622.80 | 603.42 | 605.23 | ||||
BIC | 631.47 | 617.87 | 625.46 | ||||
Fixed effects | 10.68 (0.409) | <.001 | 10.83 (0.472) | <.001 | 11.83 (0.707) | <.001 | |
Intercept (initial status) | |||||||
Time (rate of change) | |||||||
Wave = 1 | –0.08 (0.415) | –0.08 (0.415) | |||||
Wave = 2 | –0.47 (0.451) | –0.46 (0.452) | |||||
Age | –1.24 (0.791) | ||||||
Education | –0.78 (0.792) | ||||||
Variance components | |||||||
Level 1: within-person (residual) | 4.20 (0.642) | <.001 | 4.13 (0.632) | <.001 | 4.14 (0.635) | <.001 | |
Level 2: in intercept | 6.50 (1.666) | <.001 | 6.58 (1.676) | <.001 | 5.98 (1.565) | <.001 | |
Goodness of fit | |||||||
Deviance | 652.52 | 651.35 | 647.93 | ||||
AIC | 658.52 | 661.35 | 661.93 | ||||
BIC | 667.21 | 675.84 | 682.21 | ||||
Fixed effects | 8.07 (0.241) | <.001 | 8.17 (0.315) | <.001 | 8.00 (0.452) | <.001 | |
Intercept (initial status) | |||||||
Time (rate of change) | |||||||
Wave = 1 | 1.16 (0.587) | –0.51 (0.365) | |||||
Wave = 2 | 0.31 (0.484) | 0.31 (0.392) | |||||
Age | 0.38 (0.479) | ||||||
Education | –0.04 (0.479) | ||||||
Variance components | |||||||
Level 1: within-person (residual) | 3.32 (0.507) | <.001 | 3.17 (0.484) | <.001 | 3.16 (0.482) | <.001 | |
Level 2: in intercept | 1.58 (0.594) | .008 | 1.61 (0.587) | .006 | 1.59 (0.580) | .006 | |
Goodness of fit | |||||||
Deviance | 577.25 | 572.78 | 572.15 | ||||
AIC | 583.25 | 582.78 | 586.15 | ||||
BIC | 591.92 | 597.24 | 606.39 | ||||
Fixed effects | |||||||
Intercept (initial status) | 4.68 (0.248) | <.001 | 5.08 (0.274) | <.001 | 4.99 (0.428) | <.001 | |
Time (rate of change) | |||||||
Wave = 1 | –0.62 (0.208) | .004 | –0.61 (0.208) | .004 | |||
Wave = 2 | –0.69 (0.225) | .003 | –0.68 (0.225) | .003 | |||
Age | 0.54 (0.489) | ||||||
Education | –0.37 (0.490) | ||||||
Variance components | |||||||
Level 1: within-person (residual) | 1.18 (0.181) | <.001 | 1.02 (0.157) | <.001 | 1.02 (0.157) | <.001 | |
Level 2: in intercept | 2.52 (0.611) | <.001 | 2.58 (0.612) | <.001 | 2.49 (0.591) | <.001 | |
Goodness of fit | |||||||
Deviance | 491.42 | 479.80 | 478.11 | ||||
AIC | 497.42 | 489.80 | 492.11 | ||||
BIC | 506.09 | 504.25 | 512.34 |
aAIC: Akaike information criterion.
bBIC: Bayesian information criterion.
In this study we targeted the development and pilot evaluation of a digital game-based training program for managers to promote employee mental health and reduce mental illness stigma at work. Our study contributes to strengthen the evidence base that interventions targeting leaders may be effective in improving mental health literacy and reducing mental illness stigma in the workplace. In line with prior research and our hypotheses, we found statistically significant improvements in managers’ knowledge of mental health and mental illness, attitudes toward people with mental health problems, and self-efficacy to deal with mental health situations at work, with the exception of intentions to promote employee mental health [
Knowledge of mental health and mental illness is a key stigma component and a common target of antistigma interventions, as it enables recognition and is thus essential to the prevention of mental health problems [
Evidence of the potential impact of workplace antistigma interventions on managers’ attitudes toward people with mental health problems is generally mixed [
Behavioral change is key to creating an open and supportive work environment [
An open question is why LMHP did not lead to improvements in attitudes related to beliefs about workability and competency of people with mental health problems, and in managers’ intentions to promote employee mental health. One potential reason might be that managers in our sample already had quite positive attitudes at baseline regarding workability and competency of people with mental health problems, as well as intentions to promote employee mental health, which left little room for improvement postintervention. Moreover, even though people with mental health problems can function productively at work, the literature shows that employers’ beliefs about the workability and competency of people with mental health problems are often poor and may be particularly hard to change [
Due to a lack of sufficient follow-up in relevant prior studies, conclusions regarding the effectiveness of workplace antistigma interventions over the long term are limited [
While the use of digital game-based interventions in mental health promotion is scarce and especially so in the workplace, research in other settings such as schools shows promising effects, including significant improvements in students’ psychological well-being and increased engagement in a learning program [
This pilot study contributes to strengthen the evidence base of (digital) workplace antistigma interventions. Previous efforts in mental health promotion have largely neglected the role of leaders and instead have focused on employee-level interventions to address stress at work [
This pilot study has some limitations that must be mentioned. First, the study lacked a control group due to formal restrictions of the participating site. To what extent observed changes were due to the intervention is therefore questionable. To account for that, we recorded whether managers participated in further interventions during the study time, and the majority did not (30/48, 63%). Second, to measure knowledge, we developed our own quiz, which was not validated. Therefore, we used a second standardized instrument (MAKS, see Methods) and found similar change patterns in knowledge over time with both instruments. Third, while the OMS-WA as an adapted version of the OMS-HC [
Future analysis of data on employees and on EAP utilization, sickness absence rates, or the frequency and duration of disability claims before and after using the training program is essential in evaluating the full impact of LMHP. As the ultimate goal of the training was to create an inclusive and supportive working culture where employees feel comfortable to talk about mental health openly and seek help (early), it would be valuable to include employees’ perceptions on whether they feel supported by leaders, and whether and how that changed after the training. Investigating a change in objective data related to employee help-seeking would help establish the business case of investing in antistigma interventions in the workplace.
Even though we cannot be certain, it is very unlikely that a single intervention may be sufficient to end mental illness stigma and change the working culture in an organization. Hence, future research should explore whether training managers is an effective means of supporting employees with mental health problems or whether other interventions targeting employees instead or dual approaches (eg, campaign and training) may be more efficient to achieve cultural change in the long term. Finally, to increase the generalizability of our findings, workplace antistigma interventions targeting employees of different hierarchies in different types of workplaces are needed. Another appealing contribution of future research would be to compare different training formats (game-based vs standard Web-based vs face-to-face) and their effect on user engagement and learning attainment. In general, more digital workplace mental health interventions are needed that incorporate elements of positive psychology and focus on keeping employees healthy, motivated, and productive.
This pilot study provides first evidence on the effectiveness of LMHP, demonstrating its ability to positively affect managers’ knowledge, attitudes, and self-efficacy to deal with mental health situations at work. Further evaluation is needed to investigate potential beneficial effects on employees’ perceptions of management support, on their acceptance and use of existing mental health interventions (eg, EAP), and on the working culture in an organization. The benefits of digital game-based learning, such as increased participant engagement and reach, make it an effective strategy to facilitate widespread behavioral and cultural change in organizations.
employee assistance program
Leadership Training in Mental Health Promotion
Ludwig-Maximilians-University Munich
Mental Health Knowledge Schedule
Opening Minds Scale for Health Care Providers
Opening Minds Scale on Workplace Attitudes
The project received funding from the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme FP7/2007 – 2013 under REA grant agreement no. 316795. The authors alone are responsible for the content and writing of the paper.
We would like to thank Ovos Media GmbH, which supported us in the development of the game-based training program with regard to the technical solution and use of gamification elements. Moreover, we would like to thank Matthias Strack for his expert advice concerning the development of the intervention content, and Hans Bauer for his expert advice on data analysis.
None declared.