Objectives Workplace bullying may be a strong determinant of major depression, but only a few studies provide prospective data and none provide independent information on bullying. In a follow-up study we analysed newly-onset depression in relation to workplace bullying measured at the individual level (perceived bullying) and at the work-unit level (witnesses reporting bullying).

Methods Danish employees were recruited from two Danish cohorts of 3.743 and 2.617 workers, respectively. Cohort members received a questionnaire at baseline in 2006–07 with two-wave follow-ups in 2008–09 and 2011. Workplace bullying was measured by self-labelling and by the proportion of employees in a work unit who had witnessed workplace bullying “now and again” to “daily” over the past 6 months. For the latter purpose all participants were identified with their work-unit (471 work units, number of employees ranging between 1 and 161). The work-units were grouped according to the proportion of employees. who had witnessed workplace bullying within their work units. New cases of depression were diagnosed at the end of two-year follow-up periods using Schedules for Clinical Assessment in Neuropsychiatry (SCAN) interviews and the Major Depression Inventory questionnaire.

Results During the follow-up period, we identified 177 new cases of depression. The odds ratio for newly-onset depression among participants reporting bullying occasionally was 1.62 [95% CI 0.95–2.77] and among those reporting bullying often it was 5.73 [95% CI 2.37–13.90]. The risk of newly-onset depression by percentage of employees witnessing bullying in work-units was for 1–20%: 0.83 [95% CI 0.48–1.43], 21–30%: 0.87 [95% CI 0.49–1.55], and >30%: 1.08 [95% CI 0.61–1.90].

Conclusions Self-reported frequent bullying predicts development of depression but a work environment defined by witnesses of bullying does not. These findings have implications for the understanding of workplace bullying and options for preventive actions.