Background Workplace bullying has been associated with mental health, but longitudinal studies confirming the association are lacking. This study examined the associations of workplace bullying with subsequent common mental disorders 5–7 years later, taking account of baseline common mental disorders and several covariates.
Methods Baseline questionnaire survey data were collected in 2000–2002 among municipal employees, aged 40–60 years (n=8960; 80% women; response rate 67%). Follow-up data were collected in 2007 (response rate 83%). The final data amounted to 6830 respondents. Workplace bullying was measured at baseline using an instructed question about being bullied currently, previously or never. Common mental disorders were measured at baseline and at follow-up using the 12-item version of the General Health Questionnaire. Those scoring 3–12 were classified as having common mental disorders. Covariates included bullying in childhood, occupational and employment position, work stress, obesity and limiting longstanding illness. Logistic regression analysis was used.
Results After adjusting for age, being currently bullied at baseline was associated with common mental disorders at follow-up among women (OR 2.34, CI 1.81 to 3.02) and men (OR 3.64, CI 2.13 to 6.24). The association for the previously bullied was weaker. Adjusting for baseline common mental disorders, the association attenuated but remained. Adjusting for further covariates did not substantially alter the studied association.
Conclusion The study confirms that workplace bullying is likely to contribute to subsequent common mental disorders. Measures against bullying are needed at workplaces to prevent mental disorders.
Statistics from Altmetric.com
Bullying in the workplace can be characterised as repeated insulting actions, practices and behaviours that interfere job performance of employees and lead the victims to feel upset, threatened and humiliated. The victims themselves perceive being bullied and often are in an underdog position compared with the bully.1 2 The general assumption is that bullying negatively affects the victims, their workmates and the workplace environment, creating ‘toxic’ atmospheres.2 3
While the general characterisation of workplace bullying is more or less shared, there is no common definition as yet. The terminology also varies, and terms such as mobbing, harassment, victimisation, emotional abuse and psychological terror at workplace have been used as synonyms for bullying.1 Building on earlier work, the definition by Einarsen et al2 refers to bullying as the result of systematic and enduring negative and intentional acts reflecting interpersonal conflicts in which the victim is in an inferior position and subject to harassment, social exclusion, offending or other negative acts affecting work tasks. Such adverse acts give rise to negative experiences which may undermine self-confidence as well as health and well-being among the victims.4–7
Bullying at workplace is a relatively common phenomenon, and its prevalence has ranged from 5% to 20% depending on the definition, measurement and the population under study.6 In the Nordic and other western European countries, the 6–12-month prevalence of workplace bullying varied around 10%.2 6 8
Bullying is closely related to conflicts in interpersonal relationships at workplace. Such conflicts enhance distress, and as a result mental health among the victims of bullying is likely to be affected.9 As a subarea of mental health, common mental disorders are widespread in the community and impair normal social functioning but do not meet diagnostic criteria.10 11 Nevertheless they are suitable for epidemiological studies of the mental consequences of workplace bullying. Common mental disorders typically include feelings of anxiety and depression,12 and are reported even by a quarter or a third among employees depending on the measurement.13 14
In previous studies, workplace bullying has been associated with common mental disorders, such as those indicated by the General Health Questionnaire (GHQ).15 Bullying has also been associated with depressive symptoms16 and symptoms of post-traumatic stress disorder.17 A study among a small group of bullied patients suggested that severe mental health repercussions are also possible.18 While mental health has been emphasised, bullying has even been associated with physical ill health15 and cardiovascular diseases.19
While there are suggestions that workplace bullying and common mental disorders are associated, the evidence almost exclusively hinges on cross-sectional studies. Follow-up studies are needed to confirm whether the hypothesis predicting that bullying contributes to subsequent common mental disorders is plausible. A Norwegian study followed up nurses' aides for 15 months but found no association between being exposed to bullying and subsequent symptoms of anxiety and depression.20 A Finnish study, in contrast, following up hospital employees over 2 years found that those who had been bullied reported subsequent depression.19 Nevertheless, in the latter study not only did bullying predict depression but also depression predicted bullying. Thus, bullying may contribute to common mental disorders, but stigmatising health problems or personality features may also contribute to falling subject to bullying.18 19 21 22
A further question is what factors might shape the mental repercussions of workplace bullying. Factors potentially affecting either common mental disorders or bullying, or both, include prior mental disorders,21 occupational and employment position,8 23 work stress24 as well as obesity and chronic illness.19 Childhood adversities in general and school bullying in particular are likely to have long-lasting consequences. Based on retrospective information, it has been suggested that those reporting school bullying are also at risk of adult workplace bullying.25
Relying on longitudinal data among middle-aged municipal employees, this study first aimed to examine the association of workplace bullying at baseline with subsequent common mental disorders at follow-up 5–7 years later. The second aim was to examine whether common mental disorders at baseline influence this association. The third aim was to examine whether further covariates, including bullying in childhood, occupational and employment position, work stress, obesity and limiting longstanding illness influence the studied association of workplace bullying with common mental disorders.
Materials and methods
This study is part of the Helsinki Health Study examining health and well-being among the employees of the City of Helsinki, the largest workplace in Finland, with about 40 000 employees. The majority of the employees (72%) are women. The main branches include public administration, social welfare, healthcare, education, cultural services, public transportation, environmental and technical maintenance. The jobs represent more than 200 occupational titles across the socio-economic ladder.
The data were derived from mail surveys among the staff of the City of Helsinki. The baseline data were collected in 2000, 2001 and 2002 among those who each year reached the age of 40, 45, 50, 55 and 60 years. The baseline data amounted to 8960 respondents (response rate 67%).26 The follow-up survey was conducted 5–7 years later in 2007 among the respondents to the baseline survey, and amounted to 7332 respondents (response rate 83%). The data with full information for analyses included 6830 respondents. For the baseline survey, an analysis of non-response has been conducted showing that the data are generally representative of the target population, with men, younger people and manual workers as well as those on long sick leaves being slightly under-represented among the respondents.27
Workplace bullying was measured at baseline using a similar procedure as in many prior studies.19 28 29 The instruction in the questionnaire read: ‘Mental violence or workplace bullying refers to isolation of a team member, underestimation of work performance, threatening, talking behind one’s back or other pressurizing.' It was then asked whether the respondent had been subject to such bullying. The response alternatives were: never; yes, currently; yes, previously in this workplace; yes, previously in another workplace; and cannot say. The two categories of previous bullying were collapsed.
Common mental disorders
Both at baseline and at follow-up, the 12-item version of the General Health Questionnaire (GHQ-12) was used to measure common mental disorders following Goldberg's original terminology for the instrument.10 11 GHQ-12 is a reliable and well-validated epidemiological and screening instrument based on self-reports. It primarily reflects general, non-psychotic and context-free affective ill health over the past 4 weeks. It predicts more severe mental disorders but does not provide diagnostic criteria.10–12 The items of the GHQ-12 cover mainly feelings of anxiety and depression but do not include, for example, substance abuse. The measure gives a quantitative estimate along a dimension ranging from scores 0 to 12 (Cronbach α 0.91). As recommended by validation assessments, the cut-off point of 3–12 GHQ-12 symptoms was used to indicate common mental disorders.10 30 31
The data were stratified by gender. Age was included as a covariate. Childhood bullying was elicited at baseline by a question asking whether the respondents before their 16th birthday had been subject to repeated bullying in school or among fellows. Occupational class included managers and professionals, semiprofessionals, routine non-manual employees and manual workers.26 Employment status included those employed and those non-employed at follow-up. Job strain indicating work-related stress was assessed at baseline by Karasek's job strain inventory.32 Job demands and job control were cross-tabulated to yield categories of low strain, active job, passive job and high strain. A body mass index of 30 or more was used to indicate obesity. Limiting longstanding illness was asked at baseline by a question ‘Do you have any longstanding illness, disability or infirmity?’ and a following question ‘Does your illness/disability restrict your work or does it limit your daily activities (gainful employment, housework, schooling, studying)?’ A positive response to both questions implied limiting longstanding illness.
Prevalence percentages for workplace bullying and common mental disorders were calculated. Table 1 presents prevalence percentages for bullying by gender, age and occupational class including p values from χ2 tests. Next, associations of bullying at baseline with common mental disorders at follow-up were examined using logistic regression analysis. Variables were entered into the models as categorical. First, ORs adjusted for age only were calculated. Second, baseline common mental disorders were additionally adjusted for to eliminate any prior mental disorders that might have contributed to workplace bullying at baseline. Third, after adjusting for age and baseline common mental disorders, further covariates were adjusted for each one at a time.
The prevalence of being currently bullied was 5% for both women and men (table 1). Nineteen per cent of women and 13% of men had been previously bullied in the current or in another workplace. Eleven per cent of women and 10% of men reported being unable to say whether they had been bullied or not. The rest had never been bullied. There were small variations in bullying by age and occupational class, but no clear patterns could be observed.
Among women, the prevalence of common mental disorders, that is, a GHQ-12 score of 3–12, was 25% at baseline and 24% at follow-up; for men, the corresponding figures were 22% and 20%.
Logistic regression analysis showed that compared with the never bullied, currently bullied (OR 2.34, 1.81 to 3.02) or previously bullied women (OR 1.64, CI 1.41 to 1.91) were more likely to report common mental disorders at follow-up after adjusting for age (table 2). Women who were unable to say whether they had been bullied or not were also somewhat more likely to report common mental disorders. After adjusting for baseline common mental disorders, the association between being currently bullied and common mental disorders at follow-up attenuated by about a third. The association of previous bullying with common mental disorders also attenuated. There was a strong association of common mental disorders at baseline with common mental disorders at follow-up (OR 3.53, CI 3.08 to 4.04) (data not shown). Next, the effects of further covariates were examined one at a time after adjusting for age and common mental disorders at baseline. Childhood bullying, occupational class, employment status, job strain, obesity and limiting longstanding illness had no effects worth mentioning on the association between bullying and common mental disorders among women.
For men as well, the currently bullied (OR 3.64, CI 2.13 to 6.24) or previously bullied (OR 1.78, CI 1.21 to 2.63) were more likely than the never bullied to report subsequent common mental disorders after adjusting for age (table 3). Also, men who were unable to say whether they had been bullied or not were more likely to report common mental disorders. Among men as well, after adjusting for baseline common mental disorders, the currently bullied at baseline were still more likely, although a third less so, to report common mental disorders at follow-up. The associations of both previous bullying and being unable to report bullying with common mental disorders attenuated but remained. There was a strong association of common mental disorders at baseline with common mental disorders at follow-up (OR 4.22, CI 3.11 to 5.74) as found for women as well (data not shown). Further covariates were adjusted for in accordance with the procedure followed among women. Only limiting longstanding illness slightly attenuated the studied association between bullying and subsequent common mental disorders.
This longitudinal study sought to examine whether workplace bullying at baseline is associated with common mental disorders at follow-up 5–7 years later among middle-aged employees of the City of Helsinki. The main findings can be summarised under the following points.
First, among both women and men, workplace bullying at baseline was associated with common mental disorders at follow-up. This held true in particular for those who reported being currently bullied and less so among those who had been previously bullied. Second, taking baseline common mental disorders into account weakened but did not abolish the association of bullying with common mental disorders. Third, taking further covariates into account did not substantially alter the studied association. Thus, the associations between bullying and subsequent common mental disorders remained among both women and men throughout the analysis.
The main findings are in accordance with assumptions based on prior work, as our study demonstrated that workplace bullying is associated with subsequent common mental disorders. The previous evidence is almost exclusively derived from cross-sectional studies,15 16 whereas our evidence is longitudinal and comes from a prospective follow-up. The previous longitudinal evidence is limited and has either lacked standardised instruments of common mental disorders19 or been focused on a specific occupational group.20 Compared with earlier research our study yields stronger evidence on the role of workplace bullying as contributing to common mental disorders than has been available so far.
The gender patterning of the mental health repercussions of bullying has been practically neglected. In a French study, the association of bullying and depressive symptoms was very similar among both women and men.16 In our study as well, the association of bullying with common mental disorders was observed among both women and men, with the association being particularly strong among men. While it is likely that both genders suffer from mental health repercussions of bullying, more work is needed on gender differences in the production of mental disorders as a consequence of workplace bullying.
Our analyses showed that taking baseline common mental disorders into account weakened the original association of workplace bullying with subsequent common mental disorders. Mental repercussions of bullying have been focused on, but an association the other way round is also possible.19 21 Thus, prior mental or other stigmatising health-related problems might provide a route to subsequent bullying. Bullying and mental disorders may even interact over time if mentally vulnerable people are stigmatised and become victims of bullying, with this further contributing to their mental disorders. For control purposes, cross-sectional analyses of bullying and common mental disorders both at baseline and at follow-up were made. The associations at both time points, among women as well as men, were very similar, the main difference from the longitudinal ones being that the longitudinal associations were somewhat stronger. Our study reconfirms the complex pathways between bullying and mental health, pinpointing the need for future clarification of the associations.
Over and above the influence of baseline common mental disorders, the studied association of workplace bullying with common mental disorders was practically immune to the influences of further covariates. Thus, the bullied victims were more likely to report subsequent common mental disorders than their never bullied counterparts, irrespective of bullying in childhood, socio-economic factors, psychosocial stress, obesity and chronic illness. Also, in previous studies, the association of bullying with mental disorders has remained after considering covariates, although the variety of covariates in our study was broader than in the previous studies.16 19
We measured common mental disorders with the General Health Questionnaire, which is a very frequently used, reliable and valid instrument reflecting primarily milder and widespread disorders but also predicting more severe disorders and the need for treatment.10–12 30 In addition to the conventional cut-off point (scores 3–12) for common mental disorders, control analyses were made with more severe cut points (5–12 and 7–12), but these yielded practically identical results (data not shown). Nevertheless, the 12-item instrument used by us is a generic one which cannot distinguish between subdomains of mental disorders. Furthermore, the General Health Questionnaire relies on self-reports. Therefore, future examinations would benefit from additional mental outcomes based on diagnoses, treatment and psychotropic medication in order to be able to confirm bullying as a factor contributing to mental disorders varying in nature and severity.
Examining employees from a large municipal workplace, a broad coverage of white-collar and blue-collar occupations were included. Nevertheless, the target was only one single, albeit large, workplace. Thus, caution is needed in extending the results even to the public sector, to say nothing about the labour market in general. The evidence from different occupational contexts might nevertheless be taken as a suggestion for the mental health repercussions of bullying not necessarily being strongly employment-sector- or occupation-specific,15 19 20 but further work is warranted.
This study included a longitudinal design, large sample, reliable measurement of common mental disorders and a broad variety of covariates. The follow-up, that, is 5–7 years, is relatively long, and the participants may have experienced other adverse life events influencing their mental health. The participation to the baseline survey was satisfactory, and according to our non-response analysis, the data were largely representative of the target population.27 The participation to the follow-up survey was high, leaving less room for bias.
The measurement of bullying followed a similar, although relatively simple, procedure to that in a number of previous studies.19 28 29 Bullying is a multidimensional construct, and while the instruction of our measure covers various dimensions, they cannot be examined separately. Neither can we measure the severity or the duration of bullying. There are more elaborate measures such as the Negative Acts Questionnaire4 and the Leymann Inventory of Psychological Terror,16 but no measure is yet generally approved and this complicates comparisons between studies. The measurement of bullying used by us has produced findings which are in accordance with studies following a similar as well more elaborate measurement.
Our measure of bullying yielded an additional potentially important finding, as about a 10th of the respondents were unable say whether they had been bullied or not. These respondents were also somewhat more likely to report common mental disorders than the never bullied. Bullying at workplace is a sensitive issue, and some victims may find it difficult to acknowledge for themselves or other people that they have been subject to such humiliating behaviour. If some victims choose to deny bullying, part of the repercussions may also remain hidden as previously noted as well.7 This source of bias should be considered in future work. More objective measures include asking about bullying from workmates or other observers.33
Bullying elicited in surveys is strongly based on people's own experience and so are common mental disorders. Self-reports are useful and often the only option in large-scale studies. Nevertheless, self-reports need to be critically assessed for potential bias. Personality factors such as negative affectivity, that is a disposition reflecting negative emotionality and self-concept, may lead to overtly negative responses to questionnaires.17 34 In contrast to denying bullying, negative affectivity might rather overestimate the problem.35
Using a longitudinal design and ruling out the influence of several covariates, our study confirms that prior workplace bullying is likely to contribute to subsequent common mental disorders. More objective measures reflecting more severe mental disorders would be beneficial for corroborating the current findings based on self-reports of common mental disorders.
Bullying at workplaces is relatively common. Organisational policies and occupational healthcare need to minimise the adverse health outcomes related to workplace bullying. Supervisory training should include the recognition of the signs of bullying, as well as guidelines to follow when dealing with the problem. Successful tackling of bullying behaviours at workplaces would help safeguard employee mental health and functioning.
What is already known on this subject
Previous studies suggest that there is an association between workplace bullying and mental disorders.
However, the evidence is mostly cross-sectional, and longitudinal studies are needed to establish the direction of the association found.
What this study adds
This longitudinal study confirms that workplace bullying is associated with subsequent common mental disorders.
The association remains, even after considering prior mental disorders and several other covariates.
Thanks to the City of Helsinki and all members of the Helsinki Health Study group.
Funding Academy of Finland 1121748, 1129225, 1135630, Finnish Work Environment Fund 106065, Helsinki, Finland.
Competing interests None.
Patient consent Obtained.
Ethics approval Ethics approval was provided by the ethics committees of the Department of Public Health, University of Helsinki and the health authorities of the City of Helsinki, Finland.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.