Reporting the relationship between mental illness and homicide and suicide must be handled sensitively. When it is not, a social media firestorm can follow, explains Professor Louis Appleby.
It was about 18 months ago, late at night, when I checked my Twitter timeline for news. What I found took me by surprise. The mental health world – highly vocal on social media – was seething over a banner headline in the Sun: 1200 killed by mental patients.
Violence and mental illness is the most sensitive of subjects, a preoccupation of the media blamed by those of us in mental health for stoking-up stigma against the millions who suffer mental ill-health and present no risk except to themselves. It was a damaging story in the country’s biggest selling paper. But what alarmed me most was that the Sun was reporting my research.
In the 1990s a series of ‘community care killings’ had shocked the public, prompting the then government to set up a standing investigation into homicides – and, after arm-twisting by the Royal College of Psychiatrists, suicides – by mental health patients. The National Confidential Inquiry was established at the University of Manchester – it has been running now for 19 years. Our figures are quoted by all sides in the debate over mental health safety – ministers, campaigners, broadcasters and, yes, tabloids. Sometimes they are misquoted, as with the Sun – those figures were not about patients.
We have been aware from the start that just by publishing reports on patient homicide we are running the risk of fuelling popular prejudice and government alarm. We try to put the figures in context: 50 mental health patients are convicted of a homicide offence per year, but there are 1.5 million people in contact with mental health services. We stress that much of the risk of violence in patients is linked to alcohol or drug misuse rather than mental illness itself.
We have studied killings in which perpetrator and victim were unknown to each other – ‘stranger homicides’ are at the heart of public fear, but we found they were less likely to be committed by people with mental illness: it is young drunk men who kill strangers. Stranger homicides have not increased during the period when community care has come in: that media-maligned policy has not increased risk to the general public. And we have shown that, whilst successive governments have been concerned about mentally ill people as perpetrators, they are also at high risk of being victims.
Just as importantly, we have produced equivalent figures for suicide – currently 1300 deaths per year, 26 times the number of homicides. This is just over a quarter of all suicides – most people who take their lives are not under specialist mental health services, which cannot therefore be expected to prevent their deaths. This finding from our first studies made it clear that to reduce the national suicide rate, you need a broad-based approach with a role for numerous frontline agencies including the justice system and voluntary sector.
The government was persuaded and England’s first suicide prevention strategy appeared in 2002. By 2007 we had the lowest rate on record (records began in 1861) and only a churlish person would say the strategy didn’t play its part. Then 2008 brought the recession and as with every economic crisis in every country, the suicide rate rose again. It is deeply frustrating that the reckless selling of sub-prime mortgages in the American Mid-West has tracked through to the deaths by suicide of vulnerable people in the UK.
Walking the line between academic rigour and the needs of governments can be uncomfortable. Ministers are keen to show they are tackling a public concern and don’t have time for you to turn your findings into a series of clinical trials and a systematic review. They need you to draw the best conclusion you can from whatever findings you have: people talk about evidence-based policy, but the reality is that governments are run on the reasonable inference.
On that basis we advised the Government to require the removal of all weight-bearing curtain rails in mental health units: ward suicides by hanging fell by 70%. Several of our recommendations for safer community care became NHS policy. Services that adopted them had lower suicide rates subsequently. Patient homicides have fallen sharply.
When you examine the events leading up to a suicide or homicide, you can always see something that could have been done differently, an alternative sequence of events that doesn’t end with a death. Governments, the media, the public are quick to criticise when chances to prevent a tragedy weren’t taken. But retrospective enquiry can distort the picture, making judgements about patient risk look misguided when at the time they were reasonable – one of many difficult and ultimately safe decisions made every day. Risk can be examined backwards, but it has to be managed forwards.
Governments can become defensive when studies turn out findings they see as reflecting badly on them – especially when the issue is public safety. Once we mentioned in a press release that 1 in 1000 people dies by homicide – the point was that patients’ homicides are a fraction of this. Soon after, we had a call from the Number 10 press office, asking us to remove the 1 in 1000 figure from the university website, presumably in case the government was blamed for a collapse in law and order.
I would like to see government press offices become politically neutral – they put out statements on behalf of ministers, though they are publicly funded. And when government departments present misleading statistics, there should be a requirement that they publish a retraction on their website. And I would like to see policies and what they achieve subject to peer review, not self-assessment. The academic world is far from perfect, but it has standards others could learn from.