Hard evidence: how dangerous is mental illness?

Read with caution. David Jones/PA

Once again, the issue of mental illness and public danger has hit the headlines, following the murder of Christina Edkins by Phillip Simelane. It was reported Simelane was released from prison with no ongoing support in the community despite a history of violence to others and concerns about Simelane’s risk to himself.

Reporting should be accurate and balanced, with the aim of fostering an open and honest debate about how we as a society keep people safe from individuals who are potentially at risk to themselves or others. Instead, some newspapers, such as The Sun, sensationalised Christina Edkins’ murder, possibly entrenching views about mental illness and danger to the public, and strengthening the divide between two camps: those who feel that violent crime perpetrated by people with mental illnesses are common; and those who feel that talking about mental illness and violence is stigmatising, and that stories like Simelane’s receive disproportionate attention which does not reflect the rarity with which these incidents occur.

So, how do we move this debate forward in a useful way? Let’s look at the data.

The National Confidential Inquiry into Suicide and Homicide by People with Mental Illness (NCISH) is a UK-wide programme which has been based at the University of Manchester since 1996. We collect data on all suicides and homicides in the UK from national sources (for example the Office for National Statistics and the Home Office). We also collect detailed clinical data on those individuals who were in contact with mental health services in the year prior to their death or committing an offence directly from the clinicians who had been caring for them prior to the incident.

The data informs recommendations for clinical practice and policy to reduce homicide and suicide by individuals in contact with mental health services. Our figures relate primarily to perpetrators of homicide who were mental health patients although we also report figures on the victims of homicide.

Victims of homicide

Part of the negative reaction to the Sun article was in response to the misquoting of NCISH data. The Sun reported that there were 1,200 victims of homicide perpetrated by mental health patients in the UK over the previous ten years. What the data in fact shows is that there were 1,200 victims of homicide perpetrated by a person with a mental illness, or by someone who was a mental health patient.

The distinction between people with a mental illness and people who are mental health patients when they commit murder is important when discussing the role of mental health services in the prevention of homicides. Mental health services can only prevent homicides by people who are in their care.

Homicide perpetrators

Over a ten year period in England, there were on average 57 perpetrators per year who had been mentally ill at the time of the offence but had not been a patient. There were also on average 57 homicide perpetrators per year who were patients in England (74 in the UK). These figures have however been declining, since a peak between 2004 and 2006, with the lowest numbers of perpetrators reported in the most recent years.

Number of patient homicides, by gender of perpetrator.

What about perpetrators with a diagnosis of schizophrenia? How many people with schizophrenia commit murder?

Reports of homicide by people with a mental illness are often about people who have a diagnosis of schizophrenia. Between 2001 and 2010, an average of 17 patients per year who committed homicide had a diagnosis of schizophrenia although these figures have also decreased over this period.

Homicides by patients (that is, people in contact with mental health services) who have a diagnosis of schizophrenia are currently around ten per year compared to a peak of 27 in 2005.

Perpetrators with a primary diagnosis of schizophrenia.

What does this mean for the public?

In a paper published in the British Journal of Psychiatry in 2004, perpetrators of stranger homicides, where the victim was unknown to the perpetrator, were less likely to have a mental illness, or to have been under mental health care than perpetrators of homicide in general.

Stranger homicides were more likely to be men, killing other men during a physical attack, fuelled by alcohol or drugs. Our 2013 report shows that victims of homicide by patients were in fact more often acquaintances and family members (62%) than strangers (17%).

Patient homicide: relationship of victim to perpetrator.

Findings from Sweden indicate that, compared to the general public, people with a mental illness have a five-fold increased risk of being victims of homicide compared to people without mental illness.

The data tells us that people with mental illness and patients who are mentally ill are themselves vulnerable to violent acts. Also, in some cases homicide perpetrators are mentally ill or are mental health patients, and pose a risk to others. Only accurate and balanced reporting can help us move beyond stigmatising people with a mental illness, and move away from an “us and them” mentality. Equally, only by acknowledging that some people with mental illness are a risk to others, in some cases, can we address this public health issue.

Hard Evidence is a series of articles in which academics use research evidence to tackle the trickiest public policy questions.

We produce knowledge-based, ethical journalism. Please donate and help us thrive. Tax deductible.