Why are psychosocial assessments following self-harm not completed?

Mullins et al ’s study of accident and emergency (A&E) presentations following self-harm added to the evidence for poor uptake of psychosocial assessments in the initial management of self-harm.[1][1] Of particular concern was the finding that single men under 45 represented 39% of those not


Why are psychosocial assessments following self-harm not completed?
Mullins et al's study of accident and emergency (A&E) presentations following self-harm added to the evidence for poor uptake of psychosocial assessments in the initial management of self-harm. 1 Of particular concern was the finding that single men under 45 represented 39% of those not assessed. Although suicide rates among men in the UK fell between 1992 and 2007, the 2008 figures show a rise to 17.7 per 100 000, with highest rates seen in men aged 15-44. 2 A young man's presentation to A&E following self-harm is a valuable opportunity to offer interventions which reduce his risk of repetition. The paradox is that with many of these opportunities being missed researchers cannot evaluate the effectiveness of interventions to reduce repetition in this group.
Those who discharge themselves from A&E before completed assessment are 3 times more likely to repeat selfharm in the following year than those who are assessed. 3 It is possible that impulsive personality traits are more heavily implicated than the lack of an assessment, but we need to know more about this group's behavioural characteristics so that we can learn how to engage them as soon as they present. From the Mullins et al study it is not clear whether patient factors or staff factors were more influential in determining completion of a psychosocial assessment. The National Institute for Health and Clinical Excellence (NICE) recommends that patients who self-harm are 'treated with the same care, respect and dignity as other patients', 4 and reforms to medical and nursing training in some areas of the UK have managed to achieve cultural change. 5 This is crucial because a humiliating or uncomfortable experience in A&E is likely to dissuade a patient from presenting should they self-harm again, and in cases of overdose this may increase mortality risk.
It is striking that of the 341 patients in Mullins et al's study who did not receive a psychosocial assessment, 141 (41%) subsequently presented within the year of data collection having self-harmed, of whom 74 (52%) slipped through the net a second time. We are unclear of the demographic characteristics of this subgroup, or whether there was a tendency for these individuals to leave A&E at the same stage in the referral process. However, if a study of this kind was repeated across a larger geographical area, it could be sufficiently powered to reveal valuable predictors which would help A&E staff decide which patients to fast-track.
Finally, NICE recommendations on the communication of findings after self-harm assessments require auditing in future similar studies. A patient's general practitioner (GP) or community mental health team may remain completely unaware of their presentation to A&E following self-harm unless a copy of the assessment is communicated to the relevant professionals. Even if the full psychosocial assessment was not performed, an outline of the presenting complaint would be of value. Armed with this information, a GP or key worker would be able to discern any patterns emerging in selfharm presentations, sometimes to many different hospitals, and would be in a unique position to manage apparent escalations in risk. Let's target screening more effectively I was very interested in the paper by Gumber et al, 1 which examined the monitoring of metabolic side-effects of antipsychotics in patients with schizophrenia. I commend them for their attempts to follow guidance for this monitoring and I agree that metabolic side-effects are important considerations for this group of patients. However, my critical review of the evidence of risk to patients with mental illness does not support the use of such widespread monitoring. I will use the example of lipid monitoring to illustrate this. A large general practice study in the UK 2 found that the relative risk of death from cardiovascular disease in people with mental illness when compared with controls was highest in younger people and reduced with age to a point that was not statistically significant in people over the age of 75. The authors of that study claim that the three-fold increase in deaths for people under the age of 50 is the most worrying. This may be so, but the finding is worthy of closer scrutiny, especially when the implications for screening are being considered. In fact, the absolute risk of death from coronary heart disease in people with mental illness aged 18-49 was 0.1% over a median follow-up period of 4.7 years.
European guidelines for prevention of heart disease 3 recommend monitoring of lipids only when the 10-year risk reaches 5% or more. It would seem difficult therefore to justify routine monitoring of mentally ill people aged 18-49.
Also of concern is the lack of evaluation of harm to patients caused by what is essentially a screening programme of high-risk individuals. Such programmes are known to be associated with harm in a variety of forms. These include overdiagnosis, overtreatment and anxiety concerning the illness being investigated. 4 Last, for a patient to give informed consent to participate in this kind of programme, they should be informed of the uncertainties inherent in it and the likelihood or otherwise of benefit to them of such a screening.
It is time to take stock and critically review which, if any, of these investigations are necessary for our patients.

Scarcity of evidence base on management of acutely disturbed patients
Brown et al give a useful insight into the practice at seven intensive care units all over the country. 1 Their results show that 22% of patients were given rapid tranquillisation using the intramuscular route and 68% were not given any rapid tranquillisation medication at all.
The results table is confusing and the numbers do not add up; 3% appear not to have been given any medication at all, which causes concerns about the referral process to psychiatric intensive care units (PICUs) and whether patients were appropriately placed.
The study does not clarify the legal status of the patients and does not throw any light on the level of aggression of the patients in PICUs.
The most common diagnosis was schizophrenia/schizoaffective disorder (54%), followed by mania (19%) and substance misuse (8%). The diagnosis for 19% of patients has not been provided in the study.

Authors' reply
We are keen to encourage a wider discussion of the issues around the treatment of patients admitted to psychiatric intensive care units (PICUs) and welcome the opportunity to address points raised by Acharya & Sadiq. In writing the paper 1 we made a series of judgements about how best to present a large volume of data in an easily assimilated form and we are sorry if some of these decisions led to a lack of clarity. One of the main findings of the study was that most PICU patients are safely managed without recourse to forced intramuscular (IM) medication, indeed that some patients are managed without any psychotropic medication at all. The study only collected data about treatment while the patients were in a PICU (this was a pragmatic decision as many patients came from and returned to distant units where data collection was not feasible). We suspect that some of the patients who did not receive any psychotropic medication in the PICU had received medication before transfer, possibly in the form of medium-or long-acting antipsychotic injection. Others will have received medication after transfer to the acute ward. The diagnoses of those patients who did not receive any medication were: schizophrenia (1), depression (2), druginduced psychosis (1), substance dependence (2), personality disorder (2), anxiety (1) and adjustment disorder (1).
The numbers in Table 1 do not always add up to 100% because some patients appear in several categories, for example: they were given IM rapid tranquillisation and IM zuclopenthixol acetate. All figures were rounded to the nearest 0.5%; with this caveat we are confident that the appropriate figures (from text and table) do add up to 100%.
The primary diagnoses of patients aggregated into the category 'other' were: learning (intellectual) disability, dementia, Asperger syndrome, obsessive-compulsive disorder, anxiety, adjustment disorder, and intoxication with drugs or alcohol.
With respect to measurement of behavioural disturbance and mental state, we used the Brief Psychiatric Rating Scale (BPRS) and those subscales (hostility score, three-item Factor V cluster and five-item hostility cluster) which focus on behavioural disturbance. The mean BPRS score fell from 58.2 on admission to 39.8 on transfer from PICU; the respective figures for the hostility score, Factor V and hostility clusters were: 4.2 to 1.8, 9.2 to 5.5 and 17.3 to 11.1.
We hope that these details clarify the points raised by Acharya & Sadiq.