Is it time to abandon suicide risk assessment?

SUMMARY: Suicide risk assessment includes estimating the likelihood of suicide in words such as 'low, medium or high'. A 'high suicide risk' rating can trigger a powerful urge to eliminate risk immediately. But it is far from clear what 'high suicide risk' actually means. In the current state of knowledge, suicide reduction measures should apply to all psychiatric patients, irrespective of an individual patient's perceived risk. For patients presenting with suicidal thoughts, feelings and behaviour, assessment and management should focus on reducing or tolerating emotional pain. DECLARATION OF INTEREST: D.M. has received payments from Janssen-Cilag and Servier. COPYRIGHT AND USAGE: © The Royal College of Psychiatrists 2016. This is an open access article distributed under the terms of the Creative Commons Non-Commercial, No Derivatives (CC BY-NC-ND) licence.

Declan Murray is a consultant psychiatrist with the Health Service Executive Mental Health Services in North Dublin and adjunct senior clinical lecturer at the University of Limerick.

Psychiatrists in many settings routinely encounter patients who are at risk of suicide. Where the risk is apparent, assessment includes estimating the likelihood of suicide in words such as ‘low, medium or high’.[1] Clinical experience suggests that a ‘high’ suicide risk rating can trigger a cascade of events in which high anxiety creates a powerful urge to eliminate risk immediately. The result can be medicalisation and/or hospital admission for a psychosocial crisis. The immediate emotional relief for all involved and the lost opportunity to learn relevant coping skills[2] reinforce repetition of the cycle. All this occurs in spite of the fact that it is far from clear what ‘high suicide risk’ actually means, and many of those who die by suicide were not deemed to have been at high risk.[1] There are good reasons to question the practice of assessing an individual’s risk of suicide at all.

First, a ‘high suicide risk’ categorisation obscures the fact that suicide is uncommon even in psychiatric patients. The absolute risk of suicide in a total Danish national cohort followed up from the first psychiatric contact found that patients with psychiatric disorders have a lifetime suicide risk varying from 2 to 8% depending on diagnosis.3 The highest absolute risk of suicide was for men with bipolar disorder and self-harm – a rate of 17% over 36 years. This is equivalent to about 0.04% per month, a time frame more relevant to clinical situations.

Second, a ‘high suicide risk’ categorisation obscures the lack of predictive validity of suicide risk factors. For example, Pokorny[4] using the 20 best predictors correctly identified 35 of 67 patients who would eventually take their own life, but there were 1206 false positives. He found no item or combination of items helped identify to any useful degree who might die by suicide.

Third, such statistical predictions of human behaviour are actually superior to clinical assessments even when carried out by very experienced practitioners.[5] Clinical judgements can only be trusted in an environment that is sufficiently regular to be predictable with an opportunity to learn these regularities through prolonged practice. This is clearly not the case with an outcome as rare as suicide.

Finally, suicide determinations are not easy even after the event. Coroners vary considerably in the verdicts they give to individuals who probably died by suicide, and of 593 deaths classified as suicide by researchers only 385 (65.4%) received a suicide verdict.[6]

So why do we continue to do suicide risk assessments 5 years after the National Institute for Health and Care Excellence (NICE) recommended that ‘assessment tools and scales designed to give a crude indication of the level of risk (for example, high or low) of suicide’ should not be used.[7] Perhaps it is because cognitive illusions are particularly stubborn[5] and ‘…people can maintain an unshakeable faith in any proposition, however absurd, when they are sustained by a community of like minded believers’.[5] It is time to see through the illusion of validity surrounding suicide risk assessment and follow the NICE guideline.[7] We need measures aimed at making services safe for all patients irrespective of an individual’s perceived risk and at skillfully helping those presenting in emotional crisis.

Making services safe for all would include, among other things, addressing adverse events that precede hospital admission, careful, effective pre-discharge care planning and routine early follow-up following discharge from hospital.[8] Skillfully helping those presenting in emotional crisis might include reserving the language of suicide for the act of killing oneself intentionally, as most patients with suicidal thoughts, feelings or behaviour do not want to die, they just want an end to their psychological pain or ‘psychache’.[9] Management should address psychosocial problems and maladaptive coping strategies – all things that mental health professionals are usually trained to tackle.[9] Examples for research and development are safety plans[10] for emergencies, collaborative assessment and management of suicide[9] and brief cognitive–behavioural therapy (CBT)[7] for out-patient follow-up and dialectical behaviour therapy (DBT) for persistent self-harm.[2] These at least are learnable techniques, and there is evidence that CBT[7] and DBT[2] can reduce self-harm.

Making services safe for all might help reduce suicide. Although there is no evidence that the interventions for ‘psychache’ cited above do so, they are more likely to improve quality of care for psychiatric patients than ‘crude’ estimates of individual suicide risk.