Virtanen et al draw our attention to the important problem of overcrowding in psychiatric wards and its association with increased risk of violence directed at staff. ^{Reference Virtanen, Vahtera, Batty, Tuisku, Pentti and Oksanen1} There appears to be an error in the results section of the paper, in which it is reported that men are more likely than women to be working in high-occupancy wards. This is contrary to what is presented in Table 1, where women are more likely than men to be working in such wards.

Among the limitations of their study that the authors list is that data were drawn solely from the retrospective self-reports of staff, potentially resulting in errors arising from recall problems and under- or overreporting. In future prospective research, the use of structured instruments such as the Overt Aggression Scale ^{Reference Yudofsky, Silver, Jackson, Endicott and Williams2} or the Staff Observation Aggression Scale ^{Reference Palmstierna and Wistedt3} could minimise under- and overreporting.

The authors suggest a dose–response pattern after they found a strong linear trend between higher bed occupancy rates and a high probability of assault. The dose–response relationship is one of the Bradford Hill criteria of causation, ^{Reference Abruzzi4} which outline the minimal conditions needed to establish a causal relationship between two items. In addition to the reported dose–response relationship, other criteria need to be fulfilled before one can establish a causal relationship between higher occupancy rates and violence. This study was unable to collect data relating to staffing variables and acuity levels of the wards, which may be associated with the incidence of aggression. Complex relationships have been reported between staffing, patient mix and violence. Risk of violence has been reported to increase with higher numbers of nursing and non-nursing staff on planned leave, of patients known to instigate violence, of disoriented patients, of patients detained compulsorily, and with more use of seclusion. Risk of violence has been reported to decrease with higher numbers of young staff (under 30 years old), of nursing staff with unplanned absenteeism, of admissions and of patients with substance misuse or physical illness. ^{Reference Owen, Tarantello, Jones and Tennant5} It will be necessary for future studies to take into account other possible explanations (as mentioned above) and effectively rule out such alternate explanations in order to fulfil all of the Bradford Hill criteria, one of which is ‘consideration of alternate explanations’.