Interview conducted by April Cashin-Garbutt, BA Hons (Cantab)
Please could you give a brief introduction to self-harming?
“Self-harm” is the term used in the UK, much of Europe and elsewhere, to describe intentional self-poisoning or self-injury. Thus it includes overdoses, attempted hanging, self-cutting etc.
“Self-harm” does not require a specific level of suicidal intent, rather it includes acts across the whole spectrum of suicidal intent, from no wish to die through to a definite wish to die. This reflects the well-recognised fact that people who self-harm have varying degrees of suicidal intent. They may also change their account of their intent over time.
Furthermore, a range of motives or reasons are commonly involved in self-harm, including, for example, to demonstrate distress to others, to escape temporarily from a situation, self-punishment, and tension reduction. Also, multiple motives are often involved in individual acts.
Thus the terminology of “self-harm” is in contrast to the binary classification that is becoming popular in the USA, in which self-harm is divided into “attempted suicide” and “non-suicidal self-injury”. We believe that this binary approach is misleading in terms of what we know about motives for the behaviour. It also introduces an artificial division. Furthermore, it omits people who, for example, take overdoses without suicidal intent.
How many people are thought to self-harm or have a history of self-harming?
Self-harm is extremely common. At the community level, around 7-10% of adolescents report having self-harmed. However, only a minority of these will have presented to hospital or other clinical services. Thus it may be a relatively secret behaviour - this particularly applies to self-cutting.
Fortunately, when such adolescents are followed-up into young adulthood self-harm has ceased in the majority of cases. In terms of people presenting to hospital, in the UK it is estimated that approximately 200,000 people per year present to emergency departments with self-poisoning or self-injury.
In contrast to the community, where the majority of cases involve self-injury (especially self-cutting), a large proportion of those seen in hospital have self-poisoned. Thus self-harm not only represents a considerable amount of distress, it all presents major challenges for health services.
How did your research into self-harming and premature death originate?
We were initially interested in the association between self-harm and suicide. Thus in the late 1980s we followed-up a sizeable number of people who had presented to our local general hospital with self-harm to look at the risk of suicide in both the short and longer term.
We were also interested in specific risk factors. At that time we also noted that the risk of dying from other causes, including natural causes, was considerably elevated relative to what one would have expected.
A few years ago we followed-up a much large cohort group of patients (approximately 12,000) and again looked at the risk of death from all external and natural causes. This built on our earlier findings, and further highlighted the risk of premature death from specific medical causes, in addition to those related to suicide and accidents.
In our recent study published in the Lancet we extended this approach. My colleague Dr Helen Bergen had the idea of also examining how many years of expected life were lost in those dying during follow-up after a self-harm episode. This emphasised the greatly elevated risk of death in this patient population.
What did your research show?
In the recent study we followed-up more than 30,000 people who had presented to hospital with self-harm in Oxford, Manchester or Derby. This project was part of the Multicentre Study of Self-harm (http://cebmh.warne.ox.ac.uk/csr/mcm/).
These individuals had presented to hospital between 2000 and 2007 and were followed-up until the end of 2009. The average (median) follow-up was six years. It is important to emphasise that a minority of people had died by the end of the follow-up period (6.1%). However, the risk of death was much higher than would have been expected in the general population, the increase risk being 3.6-fold overall, and greater in males (4.1) than females (3.2). For individuals who died of any cause the mean years of life lost (YLL) was 31.4 years for male patients and 30.7 years for female patients.
Did your research differentiate between deaths due to natural and external causes, such as suicides?
Yes, we not only assessed all causes of death but identified those which were due to natural causes and those due to external causes. Deaths due to natural causes were 2-7 times more frequent than was expected, depending on the specific cause. The YLL for natural-cause death was 25.9 years for male patients and 25.5 years for female patients.
We found that particular causes which contributed to the excess risk were diseases of the circulatory and digestive systems. Importantly, when we looked back at information that had been obtained when people were in hospital following self-harm, more of those with physical health problems at that time had died than those who had not had physical health problems. This was found specifically for those who died of circulatory disease.
Also, more people who were found to have alcohol problems at the time of self-harm subsequently died from digestive disease than those without such problems. Also, alcohol problems were more frequent in people who died from digestive disease than in those dying from other causes.
We confirmed the recognised high risk of death from external causes in this population. Thus suicides (including open verdicts) were 20 times more frequent than would have been expected in the general population of equivalent age, accidents (which might include some concealed suicides) were 9 times more frequent than expected, with a particular excess of accidental poisonings (25 times more than expected). Interestingly, the relative risk of suicide in female patients was somewhat higher than that of males.
Did your research take into account the socio-economic status of individuals?
We investigated socio-economic status of individuals, although this was done indirectly by using standard scores of socio-economic deprivation of the area in which they lived. When all causes of deaths were combined there was an increase in risk of death with increasing levels of socio-economic deprivation.
However, this primarily applied to deaths from natural causes. These are interesting findings. It is recognised that the risk of death from natural causes in general is related to socio-economic status. However, in contrast to this, we found that the risk of external causes of death following self-harm is largely unrelated to socio-economic status.
What impact do you think your research will have?
The findings of this study emphasise the need for a holistic approach to the management and treatment of patients with mental health problems, taking account of their physical health and also risk factors and behaviours which might contribute to increase risk of physical illnesses in future. Therefore we believe the findings will contribute to the growing call for integrated health care for patients with mental health issues.
The findings support the new National Suicide Prevention Strategy for England, published on September 10th 2012, which includes a need for attention to risk of suicidal behaviour in people with long-term physical health problems. We hope that the findings will be incorporated in future advice from the National Institute for Health and Clinical Excellence in its guideline on self-harm.
Overall, we hope that the research will contribute to closer working between clinicians from psychiatric and physical health services to provide better integrated care for patients.
Do you have any plans for further research into this field?
We will be extending our follow-up of these patients to look further at mortality over the longer term. We anticipate that the risk of those who self-harm dying from alcohol-related diseases may become stronger the longer the follow up.
We are also going to investigate the longer-term risks of suicide in this group of patients, together with a subsequent cohort, which will make this the largest ever study of its kind, certainly in the UK.
Would you like to make any further comments?
This type of research emphasises the need for maintaining large-scale and long-term clinical databases on health problems such as self-harm and linking these with death records.
Where can readers find more information?
About Prof Keith Hawton DSc
Keith Hawton is a Consultant Psychiatrist, Professor of Psychiatry and Director of the Centre for Suicide Research at Oxford University Department of Psychiatry. He is also a National Institute for Health Research Senior Investigator.
For more than thirty-five years he and his research group have been conducting investigations concerning the causes, treatment, prevention and outcome of suicidal behaviour.
He has received the Stengel Research Award from the International Association for Suicide Prevention (1995), the Dublin Career Research Award from the American Association of Suicidology (2000), the Research Award of the American Foundation for Suicide Prevention (2002), and a Lifetime Achievement Award for outstanding contribution in research, treatment and teaching in the field of suicidology and suicide prevention, presented at the 14th European Symposium on Suicide and Suicidal Behaviour, Tel Aviv, Israel (2012).
He has published more than 400 papers and chapters and 15 books. He is co-editor of The International Handbook of Suicide and Attempted Suicide (2000, Wiley), editor of Prevention and Treatment of Suicidal Behaviour: From Science to Practice (2005, Oxford University Press), and co-author of By Their Own Young Hand: Deliberate Self Harm and Suicidal Ideas in Adolescents (2006, Jessica Kingsley Publishers).