930 www.thelancet.com/psychiatry Vol 2 October 2015 Series Lancet Psychiatry 2015; 2: 930–41 See Comment page 858 See Series page 917 This is the second in a Series of two papers about bullying Interdisciplinary Centre (IDC), Herzliya, Israel (A B Klomek PhD); Research Centre for Child Psychiatry (Prof A Sourander MD, H Elonheimo LLD) and Faculty of Law (H Elonheimo), University of Turku, Turku, Finland; and Turku Unviersity Hospital, Turku, Finland (Prof A Sourander) Correspondence to Dr Anat Brunstein Klomek, Interdisciplinary Centre (IDC), Herzliya 46150, Israel bkanat@idc.ac.il Bullying 2 Bullying by peers in childhood and effects on psychopathology, suicidality, and criminality in adulthood Anat Brunstein Klomek, Andre Sourander, Henrik Elonheimo Bullying is shown to be associated with adverse outcomes in cross-sectional studies, but only a few studies have prospectively examined the effects of childhood bullying on adult outcomes. Our Series paper focuses on prospective longitudinal studies that used large, population-based, community samples analysed through quantitative methods and published between 1960 and 2015. We describe the results of childhood bullying in adulthood in three of the most burdensome areas: psychopathology, suicidality, and criminality. We note that the different groups involved (ie, victims, bullies, and bully–victims) are at risk of difficulties later in life, but their risk profiles differ and the contributions are probably not independent. Controlling for confounders reduces the risk and sometimes eliminates it. Victims are at a high risk of internalising disorders. Bullies seem to be at risk of later externalising disorders and criminality, mainly violent crime and illicit drug misuse. Bully–victims seem to be at risk of internalising disorders, externalising disorders, and criminality, but not all studies examined bully–victims as a separate group. Boys and girls differ in their long-term outcomes. A dose effect exists in which frequent bullying involvement in childhood is most strongly associated with adult adversities. Future studies need to control for additional factors (including genetic, psychosocial, and environmental) to account for the mechanisms behind the reported longitudinal associations. Introduction Cultural differences exist in the definition of bullying, but generally bullying is defined as a systematic misuse of power in which one person engages in repeated aggression against another that is intentional and involves an imbalance of power. 1,2 This aggression can be physical, verbal, relational (ie, social), or cyber-based. 1 People involved in bullying include the bullies, the victims, and the bully–victims who are both victims of bullying and bully others. The reported prevalence of bullying and victimisation varies in studies since it is associated with age, culture, and the definitions and informants used by investigators; 3 however, bullying is clearly a common form of childhood victimisation. 4 Evidence has shown that bullying has long-term outcomes that can persist into adulthood. 5 We will focus on three of the most burdensome outcomes that have been associated with bullying: psychopathology, suicidality (suicidal ideation and behaviour), and criminal behaviour. Bullying is part of conduct disorder and therefore the combination of psychopathology and criminality is reasonable. In these three domains we examine four issues. First, we assess the evidence that childhood bullying behaviour predicts adulthood adverse outcomes after controlling for some childhood comorbid psychopathology and family issues. Second, we assess the difference in these outcomes on the basis of the type of bullying behaviour (eg, bully only, victim only, or bully–victim). Third, we assess whether sex differences exist in long-term outcomes. Lastly, we assess whether a dose effect indicates that long-term outcomes are associated with the frequency of exposure to bullying behaviour. Long-term outcomes of childhood bullying behaviour The need for prospective longitudinal studies Data from cross-sectional studies have suggested a clear association between bullying behaviour and psychopathology, 6 suicide, 7–11 and delinquency 12 among children and adolescents. However, cross-sectional data cannot provide evidence about causality. Retrospective longitudinal studies have shown that bullying in childhood predicts various adverse effects when they are adults, 13–16 but these are limited by their selection biases and reliance on the memories of the participants. Longitudinal prospective studies have been the most useful in their ability to provide evidence about the causal nature of the association between bullying behaviour and long-term outcomes, beyond the effects of other risk factors including psychiatric disorders and adverse family or social relationships in childhood. Most available longitudinal studies are short term, 17–20 following children for a few months or years into adolescence, 21–30 with some studies following only boys. 31,32 Studies from the past 10 years have followed up girls and boys into young adulthood and have provided evidence for the detrimental effects of bullying. 33,34 A few population-based longitudinal studies worldwide are examining psychopathology and suicidality later in life (table table 1). The Finnish 1981 Birth Cohort study The Finnish study has a nationwide representative sample of children. 45 The baseline sample was about 10% of children born in 1981 in Finland. It includes a total of about 5400 boys and girls who were followed up