Rapid Responses to:

EDITORIALS: John Geddes Suicide and homicide by people with mental illness BMJ 1999; 318: 1225-1226 [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Only 10 homicides?

Chris Manning   (9 May 1999)

Response to Geddes' editorial

Louis Appleby   (22 May 1999)

Look out for ethical costs of violence prediction

Tom Palmstierna   (4 August 1999)

Only 10 homicides? 9 May 1999
Chris Manning, GP Teddington Send response to journal: Re: Only 10 homicides?	Dear Sir, At last the evidence is emerging that the most dangerous people on our streets are people without or beyond a medical diagnosis. It is quite remarkable to realise that all the hype about BSE and severe and enduring mental illness has been allowed to rock the rational boat to the point where political outcome is all that matters. The truth is that we have more chance of being assassinated by a Bosnian hitman or being crushed by an asteroid. It is time we gained a sense of perspective and clinicians took the high ground rather than being bamboozled into submission by bureaucrats eager for votes and fuelled by massive public ignorance amongst the Murdoch majority with an average reading age of 8. In the face of the figures I would like to suggest however that GPs could uncouple ourselves from blame by not prescribing medicines that patients use to commit suicide and that, as with asthma, we do not become demoralised by outcome, but also remember the process. The asthma death rate is the same now as it was 20 years ago (admittedly with higher diagnostic and prevalence rates), but people do not seriously believe that asthma care has not improved beyond recognition. People are referred to hospital less, are admitted less and have far better quality and safer medicines than years ago. The same principles can apply to mental illness. There will be those whose outcome we cannot influence (and bashing carers and GPs for failure will lead to even lower morale), but Dr. John Geddes is right in emphasizing that "mental health services can be improved in many ways" and we must all work together to ensure that services are not skewed towards rare, if tragic events, to the detriment of improving the process of care for the majority. Yours Faithfully Dr Chris Manning
Response to Geddes' editorial 22 May 1999
Louis Appleby, Director of the National Confidential Inquiry into Suicide and Homicide by People with Mental Illnes Department of Psychiatry, University of Manchester Send response to journal: Re: Response to Geddes' editorial	Editor - In commenting on the National Confidential Inquiry report on suicide and homicide (1) that accompanied our recent papers (2,3), Geddes omits key points that are in the report (4). He devotes a detailed section of his article to the view that specific prevention measures would be wasteful of resources because suicide and homicide are rare outcomes. However, this argument loses its force if the same measures have other, much wider, benefits. As our report says, "In the relative absence of specific suicide and homicide prevention measures, the activities that are required - closer supervision, maintenance of treatments, etc - are in fact aspects of high quality care. The primary aim of a properly focussed service may then be to reduce risk but the secondary benefit is that many of the most needy patients will receive a better service." This approach underpins most of our recommendations on strengthening clinical services and not to mention it is misleading. Geddes is also critical because we are prepared to make recommendations on the prevention of suicide and homicide without the evidence of clinical trials, but this is precisely the point. It is the absence of clinical trials that makes it necessary to use other forms of evidence, as confidential inquiries in other specialities have previously done with benefit. Not that the evidence we quoted was flimsy. For example, we found that in a quarter of in-patient suicides, there were problems observing patients because of ward design, and recommended that wards should remove any physical obstructions to observation. We found that a quarter of community patients who committed suicide or homicide were not taking their medication in the previous month, and recommended measures to improve compliance with treatment. We found, as have others, that the transfer of information when patients move between services is sometimes poor, and recommended a simple form of documentation ("patient passports") to accompany them. These are hardly giant leaps of supposition. Even so, we did not make any major claims for the preventive impact of what we are proposing. In introducing our recommendations, we simply said that they amount to good clinical practice. Geddes is right about the difficulties facing research in this field. What is also needed is an equivalent understanding of the problems facing clinical services. Our report spells these out and our recommendations, which cover in-patient and community care, administrative aspects of care, contact with patients' families and training, are an attempt to address them. Professor Louis Appleby, Director Dr Jenny Shaw, Assistant Director 1. Appleby L, Shaw J, Amos T, McDonnell R, Kiernan K, Davies S, et al. Safer Services. Report of the national confidential inquiry into suicide and homicide by people with mental illness. London: HMSO, 1999. 2. Appleby L, Shaw J, Amos T, McDonnell R, Harris C, McCann K, et al. Suicide within 12 months of contact with mental health services: national clinical survey. BMJ 1999;318:1235-9 3. Shaw J, Appleby L, Amos T, McDonnell R, Harris C, McCann K, et al. Mental disorder and clinical care in people convicted of homicide: national clinical survey. BMJ 1999;318:1240-4 4. Geddes J. Suicide and homicide by people with mental illness. Editorial. BMJ 1999; 318:1225-1226
Look out for ethical costs of violence prediction 4 August 1999
Tom Palmstierna, Senior Psychiatrist and researcher Karolinska institute, dept of dependency disorders, S:t Gorans hospital Send response to journal: Re: Look out for ethical costs of violence prediction	Editor - It is with greatest interest that I have followed the ongoing debate regarding dangerousness among discharged psychiatric patients (1,2). In public debate (and sometimes also among professionals) it is often claimed that discharged patients are responsible for a substantial part of violent assaults in society. This is sometimes held as a reason for more custodial institutionalised treatment. In 1987 we performed a study of the possible effect of using violence prediction at discharge from involuntary psychiatric care as a mean to reduce violence rates in Swedish society (3). Our calculations showed that, at most, 100 convictions for serious assaults per year in Sweden were committed by patients discharged during latest year from involuntary psychiatric treatment, i.e. less than 1% of all patients discharged. Trieman et al (2) estimate that 2% of their population of discharged patients committed serious violent acts within 5 years after discharge (i.e. 0.4% per year). When discussing the possibility of preventing these serious acts of violence, the possibility of predicting statistically rare events must be taken into account (4). Beck (5) illustrated in 1985 why the psychiatric profession fails in predicting violence after discharge from institution. The reason is not lack of knowledge of risk factors important for future violence, but rather a statistical impossibility. The ability to perform a correct prediction is not only a question of accuracy in correct classification of future offenders (sensitivity of prediction), but also a question of correct identification of all future non-offenders (specificity). Assuming that the psychiatric profession in general succeeded in a prediction with 90 % sensitivity and 90% specificity, indeed a very high accuracy, our Swedish sample illustrated that for every correct prediction of future violence after discharge from institution, another eleven patients would be falsely predicted to commit serious violence. Trieman’s (2) annual rate of 0.4% assaultative patients, given the same prediction accuracy, a ratio of correct vs. false positive assault prediction would be about 1 to 30! If the consequences for the positively predicted patients include some kind of extended incarceration or violation of freedom, the question of prediction arises to a moral issue since those interventions made would not be for the benefit of the patient, but rather a precaution for society. From an ethical point of view, prediction of violence and restraining psychiatric patients into more custodial care is not only useless for society, but bears extremely high cost for those many patients falsely predicted to become violent Or, as Beck (5) put it: "the right not to be a false positive". Tom Palmstierna, M.D. PhD References Geddes J. Suicide and homicide by people with mental illness. BMJ 1999;318:1225-1226 Trieman N, Leff J, Glover G: Outcome of long stay psychiatric patients resettled in thecommunity: prospective cohort study. Brit. Med. J. 1999;319:13-16. Palmstierna T, Wistedt B. Psykiatrin kan inte axla samhällets ansvar för skydd mot våldsbrottslingar. Läkartidningen 1987;84:2768-9. (Psychiatry can not take society’s responsability for protection against violent criminal acts. Swedish medical Journal) Rosen A. Detection of suicidal patients: An example of some limitations in the prediction of infrequent events. J Consult Psychol 1954;18:397-403. Beck, J.C., Psychiatric assessement of potential violence: A reanalysis of the problem. In: The potentially violent patient and the Tarasoff decision in psychiatric practice (Ed. Beck JC), pp. 83-92, American Psychiatric Press Inc, Washington DC, 1985. Palmstierna T. Prediction of violent crime – a paradox. Proceedings Xth Nordic Congress of Forensic medicine, Linköping 1988:202-4.