The recent paper by Gunnell et al ^{Reference Gunnell, Metcalfe, While, Hawton, Ho and Appleby1} and the accompanying editorial by Pitman & Caine ^{Reference Pitman and Caine2} clearly outline the practice and principles of a contemporary approach to suicide prevention in mental health settings. However, I do not think the policy initiative that every patient with a serious mental illness or a recent episode of self-harm should be followed up within a week of discharge is really a high-risk approach to suicide prevention. Patients who self-harm and those with serious mental illness must constitute the vast majority of people who are admitted to psychiatric hospitals and therefore this recommendation is more like a universal precaution against suicide than a targeted intervention based on a high-risk model.

In my view there are compelling reasons to doubt the usefulness of high-risk categorisation for future suicide at the point of discharge from psychiatric hospitals. It is known that discharged patients have about a 100-fold increased risk of suicide compared with the general community in their first few weeks at home. ^{Reference Qin and Nordentoft3} However, those categorised as at high risk of suicide after discharge are only about four times more likely to take their own life than discharged patients categorised as at low risk of suicide. ^{Reference Large, Sharma, Cannon, Ryan and Nielssen4} Hence, compared with the risk of just being a discharged patient, being at high risk or low risk is virtually meaningless.

If the English guideline for early follow-up of patients has been successful, this is almost certainly because it approximates a universal precaution against suicide and not because of the success of a high-risk approach. We need to acknowledge that all those admitted to psychiatric hospitals have a very high absolute risk of suicide and that we are unable to tell who will be safe.