When I first started in psychiatry 25 years ago, it was accepted that there was no increased risk of violence in people diagnosed with mental illness. Large population-based studies have since been published. The generally accepted academic view is now that mentally ill people do commit slightly more violent crimes. However, psychiatric diagnosis is not a good predictor of violence. Variables such as being male and young are in fact better predictors of violence than mental illness.

The issue of public safety has driven mental health policy over the last few years. The Zito Trust, formed after the tragic killing of Jonathan Zito by Christopher Clunis on the London underground, campaigned for changes in the Mental Health Act. The Trust closed in March 2009 after the implementation of supervised community treatment and other amendments to the Act.

Despite the reduction in the number of conventional psychiatric beds, numbers of involuntary admissions and secure beds have increased. Services have become defensive and bureaucratic, and staff fear being blamed when one of their patients behaves violently.

A useful review of studies of schizophrenia and violence was published in August 2009 in PLoS Medicine, an open access journal, freely available on-line. The authors focus on the wide variation in estimates of increase in violent offending in schizophrenia ranging from 7-fold to no association. The reason for this variation may be related to whether the person also has a problem with substance abuse. A diagnosis of substance use disorder as well as schizophrenia increases the risk of violent offending. In fact, schizophrenia and other psychoses do not appear to add any additional risk to that conferred by substance abuse alone.

In other words, the association of schizophrenia and violence may be an artefact. This is created because those diagnosed as schizophrenic may also abuse alcohol and drugs. Substance misuse may in fact be the driving factor for increased risk of violence, not schizophrenia.

The learning point from this study is the reminder that an association does not necessarily imply a cause. Violence is too often thought to be due to mental illness, which is seen as a sufficient factor without looking at other reasons. The stigma of mental illness is thereby increased, reinforcing the stereotype of the ‘dangerous lunatic’.