Does workers’ experience of knowledge on emotional demands protect against burnout?

Abstract

Background

Burnout is prevalent in human service work, such as healthcare, social work, and education. Within these professions, workers are often required to regulate and express their emotions according to their professional role and the organizational context, which can lead to high emotional demands and consequently burnout. Inspired by prevention strategies in professions with physical hazards, where instruction and information on risk factors are essential for safety procedures, this study aims to assess if workers’ experience of their own knowledge on the strain and possible reactions to emotional demands, moderates the association between emotional demands at work and symptoms of burnout.

Methods

The study utilized a longitudinal survey design, collecting data at baseline and six months follow-up. The sample consisted of workers from various professions with high emotional demands (n = 1336) and Generalized linear mixed models were used to analyze the data.

Results

The findings show that higher levels of emotional demands are positively associated with burnout symptoms across the six months (b = 0.2, t = 6.0, p < .001). Additionally, workers’ experience of knowledge on emotional demands acted as a moderator on this association (b = 0.02, t = 2.5, p = .014). Stratified analysis showed that workers with higher level of knowledge exhibited lower levels of burnout at both baseline (T1) and six months follow-up (T2) Contrast estimates at T1 = High vs. Low, contrast estimate = -1.9, t=-5.5, p < .001; High vs. Medium, contrast estimate = -0.8, t = -2.7, p = .007; Medium vs. Low, contrast estimate = -1.1, t = -4.2, p < .001. Contrast estimates at T2 = High vs. Low, contrast estimate = -2.0, t = 5.2, p < .001; High vs. Medium, contrast estimate = -0.7, t=-2.0, p = .045; Medium vs. Low, contrast estimates= -1.3, t = -4.3, p < .001. These findings were consistent across various levels of emotional demands.

Conclusion

The findings suggest that enhancing workers’ experience of knowledge of emotional demands might be a simple preventive measure that can supplement existing workplace initiatives on burnout prevention.

Introduction

Burnout is prevalent in human service work [1], with emotional demands being a possible core factor contributing to its development. The demands of emotional labor in professions such as healthcare, social work or education, require workers to regulate and express their feelings in accordance with their professional role and the organizational context [2]. Providing service to patients or clients in distress constitutes a high emotional demand that can result in negative mental health consequences, such as burnout [3, 4]. Despite the recognition of burnout as a severe mental health problem in human service work, there is still need of research that investigates potential factors of resilience in relation to emotional demands to avoid development of burnout [4, 5].

Prevention in relation to physical work hazards often includes instruction and information on the risks of different exposures and how to conduct work safely [6]. Substitution and the use of protective gear are primary preventive initiatives when working with hazardous exposures. However, several physical exposures are less hazardous and cannot be prevented in daily work, e.g., handling heavy loads, working with vibrating tools, and screen terminal work etc. A pivotal prevention initiative regarding these types of physical exposure is providing information and instructions. Securing employees’ knowledge of the risks associated with the exposure, how to handle the exposure, as well as information on signs of unwanted health-related reactions and how to act following an unwanted exposure, are all part of the prevention initiatives expected of employers and organizations [6–8]. The rationale behind the instruction and information is to increase safe behavior and safe work processes.

However, this kind of information and instructions on the risk of exposure, what warning signs to look for, and how to react, are not as explicitly expected in prevention initiatives regarding psychosocial exposure. It is plausible that increased knowledge of exposure and its associated risks may also improve safety behaviors and work processes within the psychosocial work environment. This study aims to utilize workers’ self-report on their experience of having knowledge regarding emotional demands as an indicator of their level of knowledge. The primary objective is to investigate whether this perceived knowledge serves as a protective factor against the adverse effects of high emotional demands on burnout. Specifically, we will examine the association between emotional demands and burnout at baseline and again at a six-month follow-up, in order to determine whether the extent of workers’ experience of knowledge can mitigate this association.

Background

Burnout

Today, the concept of burnout is still discussed both clinically and in research. The WHO’s 11th edition of the International Classification of Diseases (ICD-11) specifically classifies burnout as a syndrome resulting from stress at work that should not be applied to experiences outside the occupational context [9]. However, the ICD-11 do not describe what constitutes burnout, and burnout is not acknowledged as a clinical diagnosis. Recently, a harmonized definition of burnout was presented by researchers and experts within the field, defining burnout as “In a worker, occupational physical and emotional exhaustion state or occupational burnout is an exhaustion due to exposure to problems at work” [10]. Although this definition builds on both classical and recent research, it narrows the perspective on the dimension of exhaustion as the primary component of the syndrome. Classical definitions in research proposed by Herbert Freudenberger and Christina Maslach, which are still prevailing in today’s research, define burnout as a syndrome caused by unsuccessful adaptation to prolonged psychosocial stressors, resulting in symptoms related to three dimensions and not only one [11–13]. Here, the core dimension of burnout is exhaustion, described as a depletion of basic emotional and physical resources within the individual, similar to the consensus definition described above [14]. However, the classical definitions include further dimensions, namely the second dimension of increased disengagement from work, often expressed as a negative and cynical attitude towards core work tasks or clients/customers [14, 15]. Finally, the development of the third dimension of burnout can follow, namely reduced personal accomplishment, expressed as reduced professional efficacy and a feeling of reduced accomplishment at work [13, 14, 16].

The understanding of how burnout is best defined and whether burnout can solely develop in relation to occupational demands is still debated [17]. But, across definitions, the dimension of exhaustion is described as a core component.

The development of burnout is argued to be a complex process predicted by a multifactorial pattern of work and personal factors. Personality factors, individual attribution style, as well as lifestyle preferences, have been described as important to the development of burnout [18]. In the organizational context, meta-studies have found that burnout is associated with high demands, job control, social support, role conflict, and justice at work [19, 20]. However, most studies on factors predicting burnout are cross-sectional, and the temporal dynamics between risk factors and burnout symptoms are uncertain in these studies. A systematic review and meta-analysis of prospective studies and case-control trials found moderate evidence that exhaustion was explained by low job control and low workplace support [21]. The study found that demands, workplace justice, and high workload were prospectively associated with the burnout dimension of exhaustion, although the quality of scientific evidence was deemed weak. The disengagement dimension of burnout was associated with workplace justice, demands, high workload, and low support from colleagues and leaders, while no association was found between these work factors and the third dimension of burnout in the form of disengagement [21]. The meta-study therefore underlines an existing understanding that each of the three burnout dimensions may be explained by different work factors, which may necessitate differentiating between factors investigated in relation to each of the three components of burnout.

Emotional demands

Emotional demands are a work factor of continued interest in relation to burnout. Emotional demands are often conceptualized as work tasks that necessitate and compel the worker to exert emotional effort to manage their feelings and behavior in interactions at work with clients/patients, colleagues, or the workplace environment [22–24]. While emotional demands are possible in all types of work, e.g. the social interaction between colleagues and leaders, they are most often described in relation to service and healthcare work, where the daily tasks explicitly entail interaction with customers and clients [24–26]. These occupations are also described as high-risk occupations in regard to burnout and depressive reactions [25–27]. One way to understand how emotional demands can lead to burnout is through the framework of emotional labor.

Emotional labor was first termed by Hochschild in 1983 as “.the management of feeling to create a publicly observable facial and bodily display.” [2]. Today, emotional labor is linked to the process where workers have to conform to a desired expression of emotion that is determined by the professional context and the norms and culture of the workplace [28–30]. Emotional labor can be considered a part of all interpersonal processes of work, where the employee interacts with their patients/clients, with a demand to regulate his or her own emotions and emotional expressions to display behavior that is conducive to achieving the professional goals [5, 28, 30]. The regulation and expression of feelings can sometimes be incongruent with what one actually feels and has been associated with emotional dissonance and emotional strain [31, 32]. It is this process of emotional strain that is thought to increase the risk of exhaustion and burnout in work tasks with high emotional demands. Working with clients or patients who are in a stressful situation or are themselves stressed or sick can be considered an exceptional type of emotional labor. In this case, the workers not only have to regulate their own emotional reactions but also have to ease and modulate the emotional reactions of their clients in order to establish a relational process that promotes the professional goal [33]. Based on the theoretical understandings of emotional demands, it seems to be a seminal factor in the development of burnout. Interestingly, the meta-study by Aronson only identifies five longitudinal studies that focus explicitly on the association between emotional demands and burnout. Furthermore, the evidence from these studies is ambiguous. Two studies are on samples of police officers (N = 179) [34] and ambulance personnel (N = 123) [35] and find small but statistically significant positive associations between emotional demands at work and emotional exhaustion one year later. Another study on a sample of teachers (N = 274) [36] found a positive significant association between emotional demands and burnout 8 months later, but the effect was not significant when adjusted for burnout at baseline. In 2013 in a study on technology workers (N = 711), Van de Ven and colleagues found that emotional demands were associated with emotional exhaustion one year later, but the level of significance was weak (p =.01) [37]. Finally, a study on oncology care providers (N = 611) found no statistically significant association between change in emotional demands and change in exhaustion symptoms in a longitudinal case-control study [38]. The existing longitudinal evidence indicates an association between emotional demands and the level of burnout. But, studies are not in consensus, and the evidence is based on very few studies. Overall, the scientific evidence was assessed as weak [21], and existing research thus underlines the necessity of further longitudinal studies to corroborate the probable association between emotional demands and levels of burnout. While the understanding of factors leading to burnout is of outmost importance, the need to increase knowledge on how to reduce burnout in workers seems even more pressing.

Knowledge as a resilience factor

The enhancement of workers’ resilience, defined as their personal ability to adapt to and cope with adversity and demands in the workplace, has been identified as a general approach to reducing the risk of burnout [39, 40]. Resilience in individuals has been associated with various factors, including self-efficacy, self-insight, self-care, and emotional intelligence [39, 41–43]. Many intervention studies on burnout prevention in the workplace primarily focus on resilience and employ strategies such as cognitive behavioral interventions, mindfulness training, stress inoculation training, supervision, or job training [44, 45]. However, the effects of these interventions on burnout, have been found to be modest (ibid.). Moreover, some of these interventions are costly and challenging to implement in high-strain work environments with diverse emotional demands [45]. Therefore, there is a need for further research on generic preventive factors that can supplement existing prevention initiatives across different working populations.

In industrial organizations and manual workplaces, there is a history of providing workers with knowledge on workplace risks and the necessary behaviors and safety measures to prevent accidents or occupational diseases. This knowledge is imparted through instruction and education, enabling workers to increase their awareness and insight into potential risks and enhance their self-care practices, both of which are core components of resilience. In contrast, the preventive culture surrounding psychosocial risk factors, such as emotional demands, is less well defined, likely due to the intangible nature and complexity of these factors. However, providing basic instructions and education on how emotional demands can affect individuals, how to identify signs of psychological strain, how to respond to one’s own or colleagues’ reactions to emotional demands, and how to seek further information on emotional demands may help workers enhance their resilience by increasing self-insight and self-care. For example, increased knowledge on emotional demands may enable workers to recognize, understand, and accept negative reactions to such demands, prompting them to take early steps to address problems they experience. Workers may also seek social support at an early stage of emotional strain if they are better able to detect warning signs and engage in self-care activities or contact supervisors to address and resolve issues related to emotional demands. Recognizing the signs of possible emotional strain is particularly crucial, as studies have shown that both individuals experiencing burnout and their colleagues often overlook or ignore symptoms and behaviors associated with increasing burnout [46, 47]. It has been argued that equipping professionals with the skills to recognize personal stress and effectively respond to psychosocial work environment strains is essential for promoting professionalism and resilience [48]. Similarly, awareness of the impact of stressors and knowledge of when and where to seek assistance have been identified as crucial factors in promoting well-being and reducing burnout [49]. However, there is a lack of studies investigating how workers’ knowledge of emotional demands can protect against burnout and mental strain. Some stress inoculation interventions include psychoeducation and training in stress awareness, which may contribute to increased knowledge about reactions to work stressors. However, these studies do not separate the potential positive effects of psychoeducation itself and the results may be influenced by other components of the interventions, such as mindfulness training or cognitive behavioral treatment (ibid.). Only one study has been identified that specifically focuses on increasing knowledge of burnout. In 2020, Manning-Geist and colleagues examined the effects of a “knowledge of burnout course” on 135 medical students in a pre-post design [49]. They found that knowledge of burnout increased factors associated with resilience, such as perceived confidence, skills, and positive attitudes toward using protective coping strategies. The study did not measure the effect of knowledge on actual burnout, but it suggested that knowledge of stressors and coping strategies might enhance resilience factors. Investigations are needed to determine if knowledge on emotional demands alone can act as a protective factor against burnout.

Existing studies on interventions targeting general stress resilience indicate that the workers knowledge regarding the exposure to stressors may be important. At the same time, prevention on physical work hazards have a long-standing practice of instructing and educating workers on explicit workplace exposures to secure healthy and secure work. It is therefore suggested that the principles of increasing the workers knowledge of a psychosocial exposure, such as emotional demands, may serve as a protective factor.

Study aims

Currently, burnout is linked to various stressors within the work environment, with some stressors being specific to each profession. Both high quantitative demands and social support are considered important work factors in the development of burnout [21, 26]. Although the understanding of how emotional work contributes to burnout is theoretically crucial, the longitudinal evidence on this association remains scarce. Additionally, it is believed that workers can enhance their resilience against burnout through self-care and self-insight. However, there is limited research on more general factors that can promote these specific factors, and existing studies often involve complex and potentially costly resilience programs. Presently, there is a lack of studies on the protective effect of workers’ general knowledge of the potential strain caused by emotional demands. It is uncertain whether general knowledge on emotional demands at work can act as a preventive factor against burnout.

The present study investigates if work tasks entailing emotional demands at baseline T1 (e.g., working with clients in difficult life circumstances or working with clients who are angry or irritable) are associated with symptoms of burnout at T1 and also six month later at T2. We hypothesize:

• H1) That a higher level of work tasks entailing emotional demands at T1 is positively associated with the level of burnout symptoms at both T1 and T2, and that higher levels of emotional demands will be associated with a significant change in burnout symptoms from T1 to T2.

The study, investigates if the experience of having knowledge on emotional demands at baseline T1 can buffer the development of burnout symptoms at T1 and T2 following emotional demands at T1. We hypothesize:

• H2) That the level of experience of knowledge a T1 will moderate the association between emotional demands at T1 and burnout symptoms at T1 and T2.

• H3) That groups with a higher levels of experience of knowledge at T1 will have lower levels of burnout symptoms at both T1 and T2 compared to groups with lower levels of experience of knowledge, in a exposure response pattern.

• H4) That groups with higher levels of experience of knowledge will have a larger decrease in burnout symptoms from T1 to T2 compared with groups of lower levels of experience of knowledge.

Methods

Sample and procedure

The study was conducted with a longitudinal survey design as part of the project In Spite of All – Working With a Smile. The overall project is a mixed-method study using quantitative and qualitative data to understand how the different types of preventive measures affect workers’ burnout across a heterogeneous sample of workers from different professions with high emotional demands in Denmark. The data in the present study consists of survey data collected in 2022 in two waves: At baseline (T1) and at six months follow-up (T2).

The study used a convenience sample, inviting respondents from a total of 129 workplaces of different sizes, all relevant for emotional demands and burnout. Data was gathered across four areas of work: health professionals (including nurses, doctors, health assistants, physiotherapists, ergonomic therapists, and psychologists); social service workers (including social workers, municipality unemployment counselors, and social benefit authority workers); educational workers (including school teachers, kindergarten teachers, special educators, special teachers, teacher assistants, and teacher substitutes); and other workers (including administrative and front desk workers, unskilled helpers, and consultants). Invitations to the study were sent through work email in collaboration with the workplace leadership. The invitation described the study’s purpose and how anonymity was ensured.

Due to the convenience sampling, we could not estimate how many potential respondents received the invitation across the 129 organizations. In total, 2181 respondents chose to participate at baseline. At follow-up, 1336 responded to the burnout measure and were included in the study, resulting in an attrition rate of 39%. Only those who responded at both baseline and follow-up were included in the present study, yielding a sample of n = 1336 (61% responserate).

Human ethics and consent to participate

Participation in the study was voluntary and dependent on written consent. The invitation email included a link leading to an electronic form describing the purpose of the study, how data were handled, how anonymity was secured, and that participation was voluntary. Upon written acceptance, the respondents was directed to the survey. The data was hosted in a REDCap database and handled in accordance with GDPR and approved by the Danish Data Protection Agency (journal number 709898). Subjects were pseudo-anonymized to all researchers other than the data manager and the project leader. No information regarding participants or their responses was conveyed to any third parties, such as the workplace, unions, or authorities. All publicly shared results were fully anonymized.

The study was not presented to any scientific ethics committee because in Denmark, studies based on survey data alone, are not eligible for consideration at the Danish Research Ethics Committees [50]. The study was conducted in accordance with the Helsinki Declaration [51] and the Danish Code of Conduct for Research Integrity [52].

Measures

Outcome

Burnout symptoms was the main outcome. We used the validated scale of personal burnout from the Copenhagen Burnout Inventory (CBI) [12], a measure of mental and physical exhaustion considered to be a core factor of burnout. The scale consists of 6 items, measured on Likert scales from 0 = not at all to 4 = all of the time. The items were summed, resulting in outcome measures ranging from 0 to 30 points. The scale showed excellent internal consistency, with Cronbach’s Alpha = 0.892.

Explanatory variable

Emotional demands were the main explanatory variable in the study. We measured the level of emotional demands with the validated scale of emotional demands from the Danish Psychosocial Questionnaire (DPQ) [53]. The scale consists of 5 items that ask about emotionally demanding work tasks, such as being in contact with clients/patients/citizens who are in difficult situations (e.g., persons with a serious disease, experiencing a life crisis, or socially marginalized, etc.). All items are answered on a Likert scale from 1 = never/almost never to 5 = always, resulting in a sum scale from 5 to 25. The scale showed acceptable internal consistency, with Cronbach’s Alpha = 0.740.

In the post hoc analysis, we also divided respondents into 3 groups based on the quartiles of the sum scores of emotional demands. The Low group represents the lowest quartile (scores ranging from 5 to 13), the Medium group represents the two middle quartiles (scores ranging from 14 to 19), and the High group represents the highest quartile (scores ranging from 20 to 25). This grouping allowed the assessment of difference between groups with large differences e.g. high vs. low, while at the same time preserving statistical strength.

Moderator variable

Experience of knowledge on emotional demands served as the effect moderator. We aimed to assess the workers’ perception of their general knowledge regarding emotional demands as a work environment risk factor. Specifically, we wanted to assess their perceived knowledge of how emotional demands could affect themselves and their colleagues, as well as their perceived knowledge of how to address reactions to the strain of emotional demands and knowledge on how to obtain more information regarding the strain of emotional demands. To assess knowledge on emotional demands, we developed a 4-item scale:

1. I believe I have sufficient knowledge of how emotional demands may affect my colleagues and me.

2. I believe I have sufficient knowledge of emotional demands to know when I should be concerned about my own reactions to emotionally demanding work tasks.

3. I believe I have sufficient knowledge of emotional demands to know when I should be concerned about a colleague’s reaction to emotionally demanding work tasks.

4. I believe I have sufficient knowledge and information on how to act if I or a colleague experiences psychological strain related to emotionally demanding work tasks.

All items were answered using Likert scales from 1 = totally disagree to 5 = totally agree. The scale was presented together with the whole survey to a group of practitioners, including leaders and union representatives from the health, teaching, and social worker professions. The group judged the scale and accepted it in regards to face validity. Principal component analysis was conducted to assess the factor structure of the scale. Oblique rotation (promax) was used, with extraction based on eigenvalues > 1 with suppression of values < 0.30. The four items had good sampling adequacy: Kaisar-Meyer-Olkin = 0.84; Bartletts test of Sphericity, Approx Chisquare (6) 5340, p >.001. Scree-plot assessment showed a clear unidimensional factor explaining 76.9% of the variance and with no factor loadings below 0.84 indicating a measure that clearly measured a single factor.

The items were summed, creating a scale ranging from 4 to 20, with excellent internal consistency, Cronbach’s Alpha = 0.899. The scale was also divided into a categorical scale to use in the stratified analysis. Here, respondents were categorized into three categories based on the sum score quartiles. Low knowledge represents the lowest quartile (scores ranging from 4 to 10), Medium knowledge represents the middle two quartiles (scores ranging from 11 to 16), and High knowledge represents the highest quartile (scores ranging from 17 to 20). This grouping was chosen to ensure the assessment of groups with large differences, e.g. the high vs. the low group while securing data strength.

Confounders

Possible confounders were included in the study. We included age, measured in five age groups: 18–30 years; 31–40 years; 41–50 years; 51–60 years; and >60 years. Age was included because several studies have found that age is associated with burnout [1]. Gender was also included, measured as female, male, or other gender identity, based on studies showing that female gender is associated with an increased risk of burnout (ibid.). We included high quantitative demands because high demands at work are associated with burnout across several studies [21, 26]. Quantitative demands were measured with the validated scale of quantitative demands from the Copenhagen Psychosocial Questionnaire version two (COPSOQ II) [54]. The quantitative demand scale consists of 4 items and was measured on Likert scales from 1 = never/almost never to 5 = always, creating a sum scale from 4 to 20. The scale showed excellent internal consistency: Cronbach’s Alpha = 0.870. Support from leaders and coworkers was also included based on studies showing that support is associated with burnout [21, 26]. We used the two validated three-item scales of leader support and coworker support, respectively, from the COPSOQ II [54]. Both scales were answered on Likert scales from 1 = never/almost never to 5 = always, resulting in sum scores from 3 to 15 points. Both scales showed acceptable internal consistency for small scales: Cronbach’s Alpha for leader support = 0.76 and Cronbach’s Alpha for coworker support = 0.74. Finally, we included a variable on work area because each work area might entail context-specific characteristics of the overall type of emotional demands, just as each work area could entail psychosocial work environment factors related to burnout that we could not capture in the survey in a way that was relevant to all respondents. The 23 professional groups in the sample were recoded into 4 overall work areas: Health professionals (nurses, doctors, health assistants, physiotherapists, ergonomic therapists, psychologists); social workers (social workers, municipality unemployment consultants, social benefit authority workers); teaching/pedagogy (including school teachers, kindergarten teachers, special educators, special teachers, teacher assistants, and teacher substitutes); and other professions (administrative and front desk workers, unskilled helpers, consultants).

Data assesment

Missing data is of special concern when working with repeated panel data. We assessed attrition using logistic regression models, entering dropout/no dropout as the outcome, with burnout, emotional demands, knowledge on emotional demands, quantitative demands, support from co-workers and leaders, age, gender, and professional group as predictors at baseline. This analysis was followed by comparisons between the dropout group and the final sample group based on all the baseline variables. Finally, we investigated the pattern of missing data using the missing data analysis tool from SPSS version 28 (Armonk, US).

Data was assessed for critical outliers with standardized residual analysis. Multicollinearity was assessed based on basic bivariate correlation analysis and assessment of the variance inflation factors (VIF) calculated from a basic multiple regression model including all baseline measures (Fields, 2016). Assessment of basic assumptions for generalized mixed models was conducted. The distribution of residuals was assessed using residual plots, and the overall trends of burnout from T1 to T2 were assessed with spaghetti plots with subject ID as the group identifier [55, 56]. Descriptive analysis was conducted, assessing all variables at baseline in regards to the distribution of means and numbers of participants both within the whole sample and within the three groups of knowledge (low, medium, and high). The outcome measure of burnout symptoms was also assessed at both baseline and follow-up in the whole sample and within the three groups of knowledge on emotional demands (low, medium, and high).

Main analysis

All analyses were conducted in a repeated measure design based on generalized linear mixed models to secure inclusion of all available data points and to be able to account for the repeated measure of burnout symptoms. The models were conducted as multilevel models. Respondent ID was set as the subject identifier, and time as the repeated measure. We chose professional group and individual subject as level one and level two, respectively, since data within these levels could be presumed to be correlated [56]. Both levels (professional group and individual subject) were entered in the models as random effects with random intercepts, using the scaled identity covariance structure to secure convergence of the model. The models were conducted with the genlinmixed function, based on normal distribution and identity link and using the Satterthwaite approximation and robust estimation, due to the unbalanced data. Significance of pairwise comparisons was adjusted with the sequential Bonferroni adjustment in all relevant models. All analyses were done with IBM SPSS version 28 (Armonk, US).

The variance explained by the grouping structure (random effect) was assessed based on the calculation of the intraclass correlation (ICC). Here, ICC was calculated by dividing the random effect variance by the total variance, based on the intercept-only model, for both professional group and subject ID, respectively.

To investigate hypothesis 1 (H1), we used a model with burnout symptoms at T1 and T2 as the outcome, calculating the fixed effects of emotional demands at T1 as the primary explanatory variable, adjusting for age, gender, quantitative demands, time pressure, support from co-workers, support from leaders, and professional group. Adjustment variables were entered as fixed effects at T1. We then entered the variable time and included an interaction term time*emotional demands at T1 to assess if emotional demands at baseline were associated with change in burnout from T1 to T2. To show the relation between level of exposure to emotional demands at T1 and symptoms of burnout at both T1 and T2, we also ran the model with the categorical variable of emotional demands*time calculating the estimated means and contrast. This model was also adjusted for all confounding variables at T1.

To investigate hypothesis 2 (H2), we used the model from H1 without the interaction term of emotional demands*time. Instead, we entered an interaction term of knowledge*emotional demands at T1 with the outcome of burnout symptoms as a repeated measure at T1 and T2.

To investigate hypothesis 3 (H3), we used stratified analysis calculating the estimated means and pairwise contrasts of the different groups of knowledge at T1 by entering an interaction term of the category variable of knowledge*time. We calculated pairwise contrast of burnout symptoms at both T1 and T2 at the three levels of knowledge measured at T1. This model was also adjusted for age, gender, quantitative demands, time pressure, support from co-workers, support from leaders, professional group, and level of emotional demands, at T1.

To investigate hypothesis 4 (H4), we used the model from H3 but included an interaction term time*groups of knowledge at T1 to assess the effect of knowledge at T1 on burnout symptoms from T1 to T2, still including all the adjustment variables.

Finally, we chose to conduct post hoc analysis where we assessed whether the protective effects of knowledge were significant at different levels of emotional demands at work. The post hoc analysis were conducted to assess, if the possible interaction effects were primarily driven at certain levels of emotional demands or if knowledge could be considered a buffer across all levels of emotional demands. In the post hoc analysis, we also used stratified analysis. Here we used a model where we entered an interaction term of the category variable of levels of emotional demands at T1 with the category variable of the levels of knowledge at T1. We estimated the pairwise contrasts of levels of burnout symptoms as a repeated measure of both T1 and T2 between groups of knowledge within each of the three levels of emotional demands. The contrast estimates were also adjusted for age, gender, quantitative demands, time pressure, leader support, coworker support, and professional group at T1.

We used AI to proofread the manuscript before submission. Here we used the proofread template in the program KAHUBI with the average creativity and English (American) settings [57]. The AI proofreading was used to compensate for the fact that neither of the authors are native English speakers. All authors made the final approval of the manuscript after the AI proofreading.

Results

Attrition and missing data

Dropout analysis found that only age-group significantly predicted dropout (B = 0.25, p <.001), with younger respondents being at a higher risk of dropout. However, the overall classification of dropout for the model was less than satisfactory (75%), leaving some uncertainty in the results.

A comparison between the sample group and the dropout group showed significant differences in means on quantitative demands (Mean difference = 0.36, t(2158) = 2.6, p =.009), emotional demands (Mean difference = 0.47, t(2140) = 3.0, p =.002), and knowledge on emotional demands (Mean difference = 0.48, t(2142) = 2.9, p =.004). This indicates that those who felt higher demands at work and had higher knowledge on emotional demands were more likely to respond at follow-up. However, the effect sizes were small (Cohen’s d: Emotional demands = 0.13, Quantitative demands = 0.12, Knowledge on emotional demands = 0.15). Additionally, a comparison between the sample group and the dropout group showed statistically significant differences in professional categories (Chi2 (3) = 15.4, p =.002) and age groups (Chi2 (5) = 44.4, p <.001). Younger age groups and the professional group of teachers/pedagogy were less likely to respond at follow-up.

Missing data amounted to 14% across all data points, with most missing data placed within the items of burnout at T1 and T2 as well as support items at T1. Otherwise, no clear pattern of missing data emerged. Although dropout analysis did not predict dropout with high precision, the statistically significant differences between dropout and sample cases, as well as the majority of missing data being located on measures at T2, indicate that missing data may be predicted by baseline data. We believe that it was appropriate to assume Missing At Random (MAR) data, thus enabling analysis based on maximum likelihood estimations [58].

Table 1 presents descriptive data. The sample consisted primarily of women, and most respondents were between 31 and 60 years old, indicating a sample with workers who have working experience above 5 years since most of the job groups are typically educated at the age between 23 and 27 years. Comparing the sample on emotional demands across the groups of low, medium, and high levels of knowledge, a tendency was seen, where the low and middle knowledge groups seemed to have a larger percentage of the professional group of teachers/pedagogy. They also seemed to have a larger group of workers in the two youngest age groups and fewer in the oldest age groups. Additionally, the level of burnout at both baseline and follow-up showed a pattern where groups with less knowledge had higher reporting of burnout symptoms.

Table 1.

Descriptive statistics presented with mean (standard deviations) and percentages (number of respondents)

Factor		Mean (SD)/Percentage (n) Total sample	Mean (SD)/Percentage (n) Low knowledge	Mean (SD)/Percentage (n) Medium knowledge	Mean (SD)/Percentage (n) High knowledge
Age group	18–30 years 31–40 years 41–50 years 51–60 years > 60 years	11% (n = 244) 23.6% (n = 524) 30.4% (n = 675) 25.9% (n = 575) 9.2% (n = 205)	14.4% (n = 81) 33.5% (n = 189) 28.8% (n = 163) 18.6% (n = 105) 4.8% (n = 27)	12.3% (n = 132) 21.7% (n = 233) 32.1% (n = 345) 26.6% (n = 275) 8.5% (n = 91)	5.2% (n = 26) 15.6% (n = 78) 30.1% (n = 150) 33.9% (n = 169) 15.2% (n = 76)
Gender	Male Female	21.7% (n = 481) 78.3% (n = 1740)	23.8% (n = 134) 76.2% (n = 430)	22.3% (n = 240) 77.7% (n = 837)	17.9% (n = 89) 82.1% (n = 497)
Occupation	Health professionals Social workers Teaching and pedagogy Other	27% (n = 549) 18.7% (n = 380) 47% (n = 955) 7.4% (n = 150)	26.2% (n = 144) 16.8% (n = 92) 49.8% (n = 282) 5.6% (n = 31)	26.2% (n = 265) 18.5% (n = 187) 48.5% (n = 490) 6.8% (n = 69)	29.6% (n = 132) 21.1% (n = 94) 38.8% (n = 173) 10.5% (n = 47)
Quantitative demands		12.5 (3.1)	13.2 (3.1)	12.3 (3.1)	12.0 (3.0)
Leader support		10.5 (2.3)	9.8 (2.4)	10.6 (2.2)	11.2 (2.2)
Co-worker support		11.6 (2.0)	11.4 (2.0)	11.6 (1.9)	11.6 (2.0)
Emotional demands		16.3 (3.5)	16.7 (3.4)	16.2 (3.5)	16.1 (3.4)
Burnout baseline		16.5 (4.5)	18.1 (4.2)	16.4 (4.4)	15.0 (4.7)
Burnout follow-up		16.4 (4.6)	18.3 (4.4)	16.2 (4.3)	15.2 (4.6)
Knowledge on emotional demands		14.7 (3.8)	9.5 (2.3)	15.3 (1.2)	19.2 (0.8)

Intracorrelation coefficients for the professional group were low (ICC = 0.02), indicating that only 2% of the variance explained the difference in groups. In contrast, ICC for the individual subject was substantial (ICC = 0.30), indicating that individual variance explained about 30% of the variance in burnout. This underscores the importance of including the subject as a random effect.

Table 2 presents results from the model testing of hypothesis 1. Emotional demands at T1 was significantly and positively associated with burnout as a repeated measure at T1 and T2 (unstandardized coeff. = 0.2, t = 6.0, p <.001), even when adjusted for all confounding variables at T1. Stratified analysis on different levels of emotional demands at T1 showed that groups with higher levels of emotional demands had significantly higher levels of burnout symptoms at both T1 and T2, even when adjusted for all confounding variables (see Table 3).

Table 2.

Fixed effects of explanatory variables on burnout symptoms, presented with unadjusted coefficients (coeff), standard error (Std error), t statistic (t) and level of significance (p)

	Coeff.	Std. error	t	p
Intercept	8.3	1,1	7.4	< 0.001
Gender Female Male (ref)	1.1 -	0.3 -	4.4 -	< 0.001 -
Age 18–30 years 31–40 years 41–50 years 51–60 years < 60 years (ref)	1.4 0.9 0.9 0.4 -	0.50 0.40 0.30 0.30 -	3.2 2.5 2.5 1.2 -	0.001 0.013 0.014 0.219 -
Occupational group Health professionals Social work Teaching/pedagogy Other (ref)	0.2 0.8 1.6 -	0.40 0.50 0.42 -	0.4 1.9 3.7 -	0.674 0.064 < 0.001 -
Quantitative demands	0.5	0.04	13.5	< 0.001
Leader support	−0.2	0.05	−4.8	< 0.001
Co-worker support	−0.2	0.05	−3.0	0.003
Emotional demands	0.2	0.04	6.0	< 0.001
Time T1 Time T2 (ref)	−0.346 -	0.50 -	−0.6 -	0.528 -
Emotional demands*Time T1 Emotional demands*Time T2(ref)	0.02 -	0.03 -	0.6 -	0.738 -

T statistic is significant at level p ≤.050

Table 3.

Estimated means and mean contrasts of burnout symptoms at baseline (T1) and 6 months follow-up (T2) in groups of low medium and high levels of exposure to emotional demands at T1. Presented with estimated means, contrast estimates and standard error (std. err.) 95 % Confidence intervals (95% CI), t statistics (t) and level of significance (p)

	Burnout symptoms T1					Burnout symptoms T2
	Estimated mean	Std. err.	95 % CI			Estimated mean	Std. err.	95% CI
Low	15,9	0.2	[15.4- 16.5]	-	-	16,0	0.3	[15.4 – 16.6]	-	-
Medium	17,0	0.2	[16.4 -17.6]	-	-	17,0	0.3	[16.4 -17.6]	-	-
High	17,7	0.2	[17.1 -18.4]	-	-	17,7	0.2	[17.1 -18.4]	-	-
	Contrast estimate	Std. err.	95 % CI	t	p	Contrast estimate	Std. err.	95 % CI	t	p
High –Medium	0.7	0.2	[0.1 -1.4]	2.8	.014	0.8	0.3	[0.2 -1.3]	2.5	0.011
High – Low	1.8	0.3	[0.9 -2.7]	6.3	< .001	1.7	0.3	[1.0 -2.5]	5.2	< .001
Medium - Low	1.1	0.3	[0.3 -1.8]	3.7	.020	1.0	0.3	[0.2 -1.7]	3.0	0.005

Adjusted for age group, gender, quantitative demands, leader support, coworker support, and type of profession.

T-statistics are significant at level p<.05

However, there was no significant change in burnout over time, and the level of emotional demands was not associated with a change in burnout from baseline to follow-up. Our results only partly corroborated hypothesis H1, because the null hypothesis was not rejected in regards to the change in burnout symptoms from T1 to T2.

Table 4 show the results from the model testing hypothesis 2: Knowledge at T1 was significantly and negatively associated with burnout as a repeated measure at T1 and T2, and the interaction between emotional demands and knowledge at T1 was statistically significant (unstandardized coefficient = 0.02, t = 2.5, p =.014). Answering hypothesis H2, the null hypothesis was rejected.

Table 4.

Fixed effects of explanatory variables on burnout symptoms, presented with unadjusted regression coefficients (coeff), standard error (Std error), t statistic (t) and level of significance (p)

	Coeff.	Std. err	t	p
Gender Female Male (ref)	1.3 -	0.3 -	5.1 -	<.001 -
Age 18 – 30 years 31-40 years 41 – 50 years 51- 60 years < 60 years (ref)	0.9 0.3 0.6 0.3 -	0.5 0.4 0.4 0.4 -	1.8 0.8 1.5 0.8 -	.066 .405 .128 .449 -
Occupational group Health professionals Social work Teaching/pedagogy Other (ref)	0.3 0.9 1.5 -	0.4 0.4 0.4 -	0.8 1.9 3.5 -	.472 .050 <.001 -
Quantitavive demands	0.5	0.04	12.2	<.001
Leader support	-0.2	0.05	-3.6	<.001
Co-worker support	-0.2	0.05	-3.0	.003
Emotional demands	0.2	0.03	7.6	<.001
Knowledge	-0.2	0.02	-6.8	<.001
Emotional demands * Knowledge	0.02	0.00	2.5	0.014

t-statistic is significant at level p≤ .050

The stratified analysis testing hypothesis H3 showed a clear exposure-response pattern between the level of knowledge at T1 and the estimated means of burnout symptoms, at both T1 and T2 (see Table 5). The group with a high level of knowledge had lower levels of burnout symptoms compared to both the group of middle and low knowledge at both time points, just as the group of middle knowledge had lower levels of burnout symptoms compared to the low group. All differences between groups were statistically significant even when adjusted for all confounding variables. Answering hypothesis H3, the null hypothesis was rejected.

Table 5.

Estimated means and mean contrasts of burnout at baseline (T1) and 6 months follow-up (T2) in groups of low middle and high levels of knowledge. Presented with estimated means, contrast estimates and standard error (std.err.) 95 % Confidence intervals (95% CI), t statistics (t) and level of significance (p)

	Burnout symptoms T1					Burnout symptoms T2
	Estimated mean	Std. err.	95 % CI			Estimated mean	Std. err.	95% CI
Low	16.9	0.2	[16.4 -17.3]	-	-	17.0	0.3	[16.5 – 17.5]	-	-
Middle	15.8	0.2	[15.4 -16.2]	-	-	15.7	0.2	[15.3 – 16.1]	-	-
High	15.0	0.3	[14.4 -15.6]	-	-	15.1	0.3	[14.5 – 15.7]	-	-
	Contrast estimate	Std. err.	95 % CI	t	p	Contrast estimate	Std. err.	95 % CI	t	p
High –Middle	- 0.8	0.3	[-1.4 - -0.2]	-2.7	0.007	-0.7	0.3	[-1.3 - -0.0]	-2.0	.045
High – Low	-1.9	0.3	[-2.7 - -1.1]	-5.5	<.001	-2.0	0.4	[-2.9 - -1.0]	-5.2	<.001
Middle - Low	-1.1	0.3	[-1.7 - -0.5]	-4.2	<.001	-1.3	0.3	[-1.9 - -0.6]	-4.3	<.001

 Adjusted for age group, gender, quantitative demands, leader support, coworker support, type of profession and level of emotional work demands, t- statistics are significant at level p<.05

Testing hypothesis H4, we found no interaction effects of time on groups of knowledge, F(2,1296) = 0.6, p =.647, thus none of the groups had a significant change in burnout from baseline to follow-up. Answering hypothesis H4, the null hypothesis was not rejected.

Addressing the protective effect of the experience of knowledge in different levels of emotional demands in the post hoc analysis showed a statistically significant exposure-response pattern of different levels of knowledge at T1 and the level of burnout symptoms as a repeated measure at T1 and T2 across all levels of emotional demands at T1 (Table 6; Fig. 1), showing that the possible buffering effect of experience of knowledge was present across both low, medium and high levels of emotional demands.

Table 6.

Contrast estimates of burnout between groups of knowledge 1=low, 2=middle, 3=high at different levels of emotional demands at work, 1=low emotional demands, 2 medium emotional demands 3 = high emotional demands. Presented with contrast estimates, standard error (Std.err.), t statistics (t) and bonferroni adjusted level of significance (p)

	Groups of knowledge	Contrast estimate	Std. err.	t	p
Low level of emotional demands	Medium – Low	-2.2	.40	-5.6	<.001
	High – Low	-3.8	.50	-8.4	<.001
	High – Medium	-1.6	.40	-4.4	<.001
Medium level of emotional demands	Medium – Low	-1.9	.40	-5.3	<.001
	High – Low	-2.8	.40	-6.5	<.001
	High – Medium	-0.9	.40	-2.4	.017
High level of emotional demands	Medium – Low	-1.1	.40	-2.8	.008
	High – Low	-2.4	.50	-4.7	<.001
	High - Medium	-1.4	.50	-2.9	.008

Adjusted for age, gender, quantitative demands, leader support, coworker support and type of profession.

t-statistic is significant at p≤.050

Fig. 1.

presenting level of burnout symptoms, y axis, in three groups of level of knowledge x– axis (1 = low knowledge, 2 = middle knowledge, 3 = high knowledge) at three levels of emotional demands presented as lines (red line = 3, high demands green line = 2, medium demands, blue line = 1, low demands)

Discussion

The study showed a clear exposure-response pattern between the level of emotional demands at T1 and burnout symptoms at both T1 and T2. Furthermore, the study demonstrated that workers’ perception of their level of knowledge of emotional demands at T1 was inversely associated with the level of burnout symptom at both T1 and T2, even after adjusting for several confounders and across different levels of emotional demands.

The study hypothesis that emotional demands at T1 was positively associated with burnout at T1 and T2, was corroborated as the null hypothesis was rejected. Our results corroborate the findings from existing studies in other professions, such as police officers [34], ambulance personnel [35] and technology workers [37]. Interestingly, our study did not find a change in burnout symptoms over time, suggesting that a worsening of burnout symptoms from T1 to T2 could not be detected in association with emotional demands at T1. Studies on burnout have shown that the development of symptoms often occurs over long periods of time, and it has been argued that changes in burnout over shorter time periods are often small and difficult to detect [59]. Moreover, our sample of workers was already exposed to high emotional demands at baseline and may have already reached a plateau of symptoms, making it difficult to detect worsening or reduction of burnout symptoms in the short time span of six-month. The study results likely reflect the continued effects of ongoing exposure to emotional demands. This would also explain why the existing evidence, including the present study, all report small coefficients of association of emotional demands at baseline and level of burnout symptoms at follow-up, as the studies are conducted within a timeframe of 12 months or less and all on a working population already exposed at baseline. The small variation over time would also explain the findings by Le Blanc and colleagues, who found no significant associations between baseline level of emotional demands and burnout symptoms at follow-up, among oncology workers [38]. The results from the current study, based on a sample of different professions, corroborate the overall tendencies in extant longitudinal research, that emotional demands at work are a probable causal factor of burnout. But, further research is needed regarding emotional demands and burnout, with long follow-up periods, as well as studies based on samples of non-exposed individuals at baseline, such as newly educated workers. These studies would greatly expand our knowledge of how emotional demands at work can lead to burnout.

The study also revealed a significant interaction between exposure to emotional demands and worker`s experience of knowledge measured at baseline (T1), which was associated with burnout levels at both T1 and T2. Subsequent stratified analyses indicated an inverse relationship between experience of knowledge at T1 and burnout at both time points. The rejection of the null hypothesis in both analyses corroborates the idea that experience of knowledge serves as a protective factor, mitigating the adverse effects of elevated emotional demands on burnout. This was true even when adjusted for relevant confounders. Furthermore, the post hoc analysis showed that the effect was present at all levels of emotional demands. The post hoc analysis indicate that experience of knowledge is a possible protective factor at all levels of emotional demands, and not only at either high or low levels of exposure. It is worth noting that no change was seen in burnout from baseline to follow-up in either of the knowledge groups. So, the null hypothesis was not rejected when testing hypothesis H4. We believe that the main reason that our results do not corroborate this hypothesis is because of the probable plateau effect mentioned above. Burnout symptoms did not generally change from T1 to T2. We believe that studies with longer follow-up periods, as well as experimental studies on interventions that increase workers’ knowledge of emotional demands, are needed to ascertain the possible protective effects of knowledge.

Although no previous studies were identified that specifically investigated knowledge of emotional demands and burnout, our findings are consistent with the notion that general experience of knowledge of the possible impact of a stressor can increase resilience and induce better coping, as proposed by Dyrbye and colleagues [48] as well as Manning-Geist and colleagues [49]. Additionally, the findings by Manning-Geist, that education on knowledge of stress reactions increased students’ experience of being able to handle stressful situations, align with our results.

The study presents novel data on how workers’ perception of their knowledge of when and how to act on the strain from emotional demands, can act as a possible buffer against burnout. The study results should be seen as a first tentative step towards an awareness, that prevention initiatives against burnout caused by emotional demands might be seen in parallel to some of the existing preventive procedures seen in professions with physical hazards. Here, information and instruction regarding the potential harmfulness of an exposure, what signs of strain to look for and how to act, are part of the introduction and ongoing training of workers. A focus on increasing workers’ knowledge of emotional demands as a possible work hazard could be an important future initiative towards the prevention of burnout. However, our results should be understood as preliminary. Our study does not clarify the specific manner in which the experience of knowledge influences the strain associated with emotional demands. Additional studies are needed to corroborate our findings and to decipher how knowledge is gained and how it affects workers’ coping strategies. Qualitative studies, such as in-depth interviews or focus groups, could provide a deeper understanding of how workers perceive emotional demands, how they acquire and interpret knowledge about these demands, and how this knowledge affects their emotional well-being and burnout. Also, studies are needed, that assess the workers objective levels of knowledge on emotional demands and how it affects burnout. Here studies better suited to detect the effect of knowledge acquisition, such as controlled trials testing interventions that increases workers knowledge are preferable.

Limitations

The limitations of the study are important to consider when interpreting the results. Convenience sampling and the relatively high dropout rate raise questions about the generalizability of the findings. However, it is important to note that the study did not aim to report incident or prevalence rates within working populations. Instead, the focus was on investigating possible causal mechanisms. While the study sample may consist of workers who are specifically concerned with burnout or high emotional demands, the processes that are investigated are believed to be of a more general nature why the associations are also more generally applicable. Additionally, the use of a heterogeneous sample of workers from different key working areas improves generalizability compared to studies conducted on one specific working population. The reliance on a sample derived from a population characterized by high emotional demands limits the potential for contrasting exposure levels within the sample and precludes the possibility of comparing exposed individuals with non-exposed individuals. Nevertheless, the associations identified, along with the stratified results that demonstrate an exposure-response pattern, support the assertion that a relationship exists between emotional demands and burnout. One could contend that the variance in exposure observed within a uniformly exposed sample across similar occupational roles should be interpreted as measurement bias, arising from factors unrelated to actual differences in exposure. However, an understanding of the practices inherent to these occupations reveals that exposure to emotional demands is intrinsically linked to the patients and their relatives assigned to the individual worker`s caseload. Consequently, within the same occupational setting, organization, and even department, each colleague may experience varying degrees of emotional demands when evaluated over shorter time intervals, such as months. While these differences are likely to balance out over extended periods, it is plausible that variations in exposure exist within the timeframe of the current study. The instrument utilized to assess exposure to emotional demands has undergone validation and has demonstrated the requisite properties for evaluating emotional demands through self-report. Therefore, the differences in exposure reported in this study are unlikely to be predominantly attributable to reporting bias. Although sampling bias is a concern and generalization should be done with caution, the results are most likely not primarily the product of this type of bias and they are worthy of consideration in relation to emotional demands and burnout.

Another concern is the potential for common method bias, which is a well-established issue in survey studies [60]. The risk of inflated associations due to individual preferences, personality factors, or perceptions of response categories is acknowledged. However, Neglecting to include the subjective experience of the phenomena reported would reduce the accuracy in modeling the associations between psychosocial work environment factors and the individuals reaction since intra psychic processes are present [61]. Recent studies have discussed that the impact of common method bias on statistical results is unclear and may favor or not favor the null hypothesis [62]. It is still uncertain whether this bias will have a substantial impact (ibid.). Additionally, the use of random effects of the individual in the mixed models helps to accommodate for some of the systematic variance of the individual not explained by the response items. While common method bias cannot be ruled out entirely, it is unlikely that the study results are primarily a product of this type of bias. Furthermore, it is important to note that the interaction effects between emotional demands and knowledge are not affected by common method variance [63]. We believe that the study main results regarding this possible protective factor should not be seen as a product of common measures bias.

The use of self-assessment of knowledge as a proxy for actual knowledge is also a limitation in the study. While several arguments support the use of self - assessment as a methodological approach, it is essential to acknowledge its inherent limitations. For instance, self-assessments are often susceptible to biases stemming from social cognitive factors, which may lead individuals to either overestimate or underestimate their true levels of knowledge. Discrepancies in self-assessment can arise depending on contextual factors, resulting in variations in individuals’ perceptions of their competencies across different scenarios [64]. A primary rationale for utilizing self-assessment as a proxy for actual knowledge is its potential to reflect an individual’s confidence and self-efficacy. Such factors can significantly impact motivation and engagement in learning processes [65]. Thus, our proxy measure of knowledge may not accurately represent the actual level of understanding regarding emotional demands, but could be an indicator of heightened confidence in one’s capacity to learn about and manage high emotional demands [66, 67]. Also, burnout measured with the Personal Burnout Scale from the CBI can be considered a less definitive assessment of burnout compared to other scales. The personal burnout scale assesses the exhaustion component of burnout without assessing possible etiology or context dependence of these symptoms. As such, the symptoms of physical and mental exhaustion measured with this scale could result from factors other than the work environment.

Implications for research and practice

The implications of the study findings for research and practice are potentially significant. The results suggest that increasing workers’ knowledge regarding emotional demands, for instance, knowing when one should be concerned about one’s own or colleagues’ reactions to emotionally demanding work tasks, may enhance their resilience against the detrimental effects of such demands. This finding presents an opportunity for supplementary strategies in the prevention of burnout. Workplaces may benefit from implementing initiatives to increase staff learning about emotional demands and how it affect them. Simple courses, teaching sessions, or work groups could be used to enhance workers’ knowledge of this primary work exposure in relational tasks and its impact on psychological well-being. This approach, similar to instructions and teaching for other work environment risk factors, could be a simple and low-cost addition to existing prevention initiatives. It has the potential to heighten workers’ perception of and reaction to high emotional demands, thereby increasing individual resilience.

However, future studies should assess the long-term effects of this knowledge and determine if it can reduce symptom severity over extended periods of time. Moreover, further research is needed to understand the causal mechanisms through which knowledge acts as a resilience factor. It is important to investigate whether increased knowledge leads to changes in coping strategies at work, increased use of supportive resources, higher levels of self-care, or increased understanding of oneself, all aspects coined to increased resilience protection.

Additionally, exploring different sources of knowledge and their effects on overall workers’ knowledge is crucial. Our study asses the individual experience of having sufficient knowledge. However, the study does not assess what this knowledge consist of or how it is obtained. Understanding whether knowledge obtained through for example structured supervision has the same effect as knowledge obtained through for example teaching courses, instructions, or sustained focus and discussion in the work environment and organizational safety systems is needed. This knowledge would provide valuable insights for future dissemination of burnout prevention related to high emotional demands across working populations.

Conclusion

This study, conducted on a large sample of different work groups relevant for emotional demands at work, showed statistically significant associations between high emotional demands at baseline and the level of burnout symptoms, at both baseline and six months later. The study also found a clear tendency indicating that workers’ general knowledge of emotional demands might protect against burnout symptoms. While the study has limitations, the findings have potential implications for research and practice. Increasing workers’ knowledge of emotional demands may enhance their resilience against burnout. Future research should assess the long-term effects of this knowledge, investigate the causal mechanisms, and explore different sources of knowledge as well as the quality of knowledge to optimize prevention strategies. By addressing the limitations of this study future studies can further evidence about the protective effect of knowledge, and researchers and practitioners can potentially contribute to the development of effective interventions to mitigate the negative effects of emotional demands in the workplace. The current study presents initial evidence that enhancing workers’ knowledge of the possible detrimental effects of emotional demands and how to act on the warning signs of strain, can act as a protective factor against burnout symptoms.

Abbreviations

WHO

World health organisation

ICD

International classification of diseases

T1

Timepoint 1

T2

Timepoint 2

H1

Hypothesis 1

H2

Hypothesis 2

H3

Hypothesis 3

H4

Hypothesis 4

GDPR

General data protection regulation

CBI

Copenhagen burnout inventory

COPSOQ II

Copenhagen psychosocial questionnaire version two

DPQ

Danish psychosocial questionnaire

ICC

intraclass correlation

MAR

Missing at random

Authors’ contributions

JPT, DRA and LPA participated in the conception and design of the study. Data was collected by LPA and JPT. Analysis was conducted by JPT. Results were interpreted by LPA, DRA and JPT. JPT made first draft of the manuscript. Critical revision of the manuscript were done by LPA, DRA and JPT. LPA, DRA and JPT all approved the final version to be published and agreed to be accountable for all aspects of the work, to be named as authors and approved the full author list.

Funding

Open access funding provided by University of Southern Denmark. The study was funded by the Danish Working Environment Research Fund (Arbejdstilsynet) grant No. 20205100159.

Data availability

The data that support the findings of this study are not openly available due to reasons of sensitivity. Upon reasonable request a fully anonymized dataset can be made available from the third author.

Declarations

Ethics approval and consent to participate

In Denmark, studies based on survey data alone, are not eligible for consideration at the Danish Research Ethics Committees [50]. Therefore, the study was not presented to any scientific ethics committee. The study was conducted in accordance with the Helsinki Declaration [51] and the Danish Code of Conduct for Research Integrity [52]. Participation in the study was voluntary and dependent on written consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.