The study by Okorie et al ^{Reference Okorie, McDonald and Dineen1} of an accident and emergency (A&E) unit in Galway, Ireland, was of some interest given our experiences in an A&E in Middlesbrough. ^{Reference Kripalani, Nag, Nag and Gash2} Interestingly, the authors did not mention the proportion of mental health service users presenting with self-harm, a well-known cause of frequent attendance in our neck of the woods.

Whether the authors include presence of self-harm as evidence of mental disorder is another matter. It is an interesting debate, one which needs to happen with regard to accepting the role of services in managing emotional distress and at what threshold it should be labelled as mental disorder, an issue which was highlighted before. ^{Reference Patil, Rasquinha, Badanapuram, Kripalani and Gash3} Nevertheless, we know that self-harm and attempted suicide are represented at a higher rate in individuals with personality disorder, schizophrenia, bipolar disorder and alcohol/substance misuse. Yet Okorie et al have decided to disregard that particular and significant piece of the puzzle in those presenting to A&E seeking psychiatric assessment and care.

I am not sure how services are structured in Ireland. To appreciate the possible impact of a crisis team on local A&E services, it would have been useful to first describe how psychiatric assessments are currently made available to the attendees, including screening those not known to have previous involvement with mental health services. However, whether ‘community-oriented teams’ such as a home treatment team would lead to cost reduction (by reducing A&E attendances as the implication seems to be) might be stretching existing evidence and is evidently unsupported by this survey. This effort should also have accounted for a wide variety of variables which have an impact on individuals seeking psychiatric care and assessment in A&E and hence at least a replication should have been attempted before the study was submitted for publication.

Finally, Okorie et al's conclusions, as stated in the clinical implications in the abstract, are quite surprising, despite the mentioned limitations. Their wording, in my opinion, is unfortunate and overestimates evidence, and is completely out of synch with the survey.