In order to meet the government target of a reduction in the suicide rate of at least one-fifth by the year 2010 (Department of Health, 1998), a broad package of health and social measures is needed. The National Service Framework (NSF) for Mental Health (Department of Health, 1999) incorporates objectives aimed at specialist mental health services, primary care, social services and public health bodies. It offers an integrated approach to the enhancement of mental health and reduction in adverse outcomes, including suicide. In the NSF the prevention of suicide is one of seven key ‘standards’ and it makes clear that mental health services have a central role to play. Nevertheless, there is scepticism among mental health professionals regarding their role in suicide prevention and the common view persists within clinical practice that most suicides are not preventable. Suicide, it is said, is the product of adverse social factors such as unemployment, over which psychiatrists have no control.

Four papers in this issue touch on the prediction and prevention of suicide by mental health services. Slinn et al (Reference Slinn, King and Evans2000) highlight patients presenting with deliberate self-harm as a group at established high risk of eventual suicide. The authors examine the concordance between service provision for patients presenting with deliberate self-harm and the guidelines contained within a Royal College of Psychiatrists' consensus statement (1994). The findings demonstrate variation between trusts both in the targeting of high risk patients and in more general aspects of service provision and training. In 30% of trusts responding to the survey, not all patients admitted with deliberate self-harm were assessed by psychiatric services. Only 42% of trusts had a deliberate self-harm planning group, one of the key recommendations made by the College.

Raymont (Reference Raymont2000) reviews risk factors for suicide among patients with schizophrenia. She emphasises the need to incorporate research findings into staff training in risk assessment, the role of the multi-disciplinary review following a patient suicide and the need for large case-control studies to evaluate individual risk factors for suicide.

In a third paper, Hodelet and Hughson (Reference Hodelet and Hughson2000) provide a commentary on the role of the clinician when a patient commits suicide. They refer to the emotional impact that the suicide of a patient may have, a theme taken further by Courtenay and Stephens (Reference Courtenay and Stephens2000), who investigated the severity of emotional reactions among junior psychiatrists in London following a suicide. Their findings are in line with another recently published study that examined the effect of patients' suicide on consultant psychiatrists in Scotland (Reference Alexander, Klein and GrayAlexander et al, 2000). Although the rate of emotional distress among consultant psychiatrists was lower than that found by Courtenay and Stephens among junior psychiatrists, 33% of consultants reported being affected by psychological symptoms following a patient suicide. The same study also found a positive role for the multi-disciplinary review after suicide, highlighting opportunities for support, learning and improved management of suicide risk.

Recent studies have emphasised the preventive potential of mental health services. A Danish case register study (Reference Mortensen, Agerbo and EriksonMortensen et al, 2000) demonstrated that a history of mental illness necessitating hospital admission was the most significant risk factor for suicide in the general population among a range of mental health and socio-economic variables. Using a case-control study design, the authors noted that the effect of socio-economic variables decreased after adjustment for history of mental illness, with population attributable risk associated with mental illness of 45%, compared with 10% for being single and 3% for being unemployed. Previous studies, said the authors, had overplayed the importance of unemployment by not taking account of severe mental illness.

Further evidence for the link between aspects of mental health service provision and suicide is provided by a case-control study of suicides in Greater Manchester (Reference Appleby, Dennehy and ThomasAppleby et al, 1999a ). All patients had a history of inpatient psychiatric care in the 5 years before death. The study found that significantly more suicides (44%) than controls (26%) had a reduction in care at the last consultation before death. Such reductions in care included reduced drug dosage, reduced appointment frequency and discharge from services. In the majority of suicides (56%), death occurred within 3 months of the reduction in care.

The National Confidential Inquiry into Suicide and Homicide by People with Mental Illness (Reference Appleby, Shaw and AmosAppleby et al, 1999b ) estimated that 24% of general population suicides had been in contact with mental health services in the year before death. In 22% of these (30% in the case of suicides among in-patients), mental health teams judged that the suicide was preventable and in two-thirds of cases, expressed the view that more measures could have been taken to reduce risk. Forty per cent of suicides in the Inquiry sample were in-patients at the time of death or had left hospital in the previous 3 months. In the latter group, suicides clustered in the first post-discharge week; 40% occurred before the first follow-up appointment.

But the problem of predicting suicide remains. Because of this, the Inquiry recommended changes to the service as a whole, not just for patients who were identified to be at high risk. There are weak points in our services — immediate post-discharge care, ensuring compliance with treatment, maintaining contact with patients — and the Inquiry's recommendations were intended to strengthen these. The main benefit will be better general care for vulnerable patients, but another effect could be reduced suicide risk.

The inquiry also addresses some of the main points made in the papers published here. Our recommendations included better skills training for those in the clinical front line and multi-disciplinary case review. We have begun case-control studies of in-patient and post-discharge suicides. The results are intended to make prediction more precise and to lay the groundwork for the clinical trials on which future service development may be based.