We applaud Edward Ackling for his consideration of the importance of considering risks and agree that undertaking this difficult assessment is a core part of psychiatric practice. We also agree that suicide is ‘one of the worst outcomes’ and requires not only our clinical skills but our research interest. Ackling identifies the core issue of the problem of the low base rate, and again we agree, but we struggle to see how this provides a basis for suggesting that a structured risk assessment tool would be of benefit. The evidence to date simply does not support that stance. ^{Reference Chan, Bhatti, Meader, Stockton, Evans and O'Connor1} Ackling suggests the imprecision of future prediction supports a structured professional judgement (SPJ) tool and asserts that this allows us to remain associated to the patient while ‘following an evidence base’. We would be interested to see the evidence base that such SPJ tools reduce rates of suicide in the general population, the population presenting to the emergency department for psychiatric review, or any day-to-day psychiatric population. To our knowledge, the only randomised controlled trial which compared SPJ with assessment as usual failed to show any superiority in preventing subsequent violence. ^{Reference Troquete, van den Brink, Beintema, Mulder, van Os and Schoevers2} Our concern is that SJP feels like science and is therefore more comfortable for the clinician, while not actually providing any clinical benefit.

The problem is that suicidal behaviour has multiple causes and occurs in many psychiatric conditions, as well as in the absence of any psychiatric diagnosis. There is nothing wrong with trying to create a structure for suicide risk assessment, but the emphasis on the predictive ability of SPJ (typically measured using the area under the curve (AUC) statistic) conceals the potentially endless list of risk factors that must be filtered through a clinical algorithm, and which ultimately only an experienced clinician can process. And for some diagnostic categories, reliance on SPJ to predict outcome is as accurate as tossing a coin. ^{Reference Coid, Yang, Tyrer, Roberts and Ullrich3}

Ultimately, clinical experience and skill, rather than a checklist of commonly associated variables, are needed to determine the appropriate intervention. It is the intervention after assessment, or the risk management, that is the key and typically omitted from SPJ.

So, no more suicide risk assessments without risk management. It is also important to be wary of paying for courses taught by those with little relevant clinical experience such as the emergency room, as is often the case with SPJ. Finally, do psychiatrists who routinely assess suicide risk really carry out ‘unstructured’ assessments?