There are a number of things we know about self-harm (National Collaborating Centre for Mental Health, 2004). It is a major public health problem accounting for up to 170 000 hospital attendances in the UK each year, its incidence seems to be increasing, and it confers a considerable risk of completed suicide. Self-harm is one of the most common reasons for admission to hospital. Effective intervention for self-harm probably represents one of the best opportunities for suicide prevention worldwide. What is still unclear, despite recent guidance, is how we might best manage individual patients when they present to health services.

GUIDELINES: OLD AND NEW

Guidelines on the management of deliberate self-harm were published by the Department of Health and Social Security (1984) and the Royal College of Psychiatrists (1994). These documents emphasised the role of psychosocial assessments, multidisciplinary approaches to working, adequate training and supervision, and the organisation of services. However, service provision for self-harm remained extremely variable.

Since then two sets of guidelines have been published (National Collaborating Centre for Mental Health, 2004; Royal College of Psychiatrists, 2004). Both have dropped the prefix ‘deliberate’ from ‘self-harm’ in response to the heterogeneous nature of the phenomenon and the concerns of service users. What needs to be emphasised (whatever the terminology) is that self-harm includes both self-poisoning and self-injury. There may be a belief among non-specialists that the term refers primarily to those who cut themselves, and even academic journals are sometimes guilty of misrepresentation (Reference Horrocks, Owens and HouseHorrocks et al, 2002).

The National Institute for Clinical Excellence (NICE) guideline (National Collaborating Centre for Mental Health, 2004) is certainly comprehensive and considers the short-term physical and psychosocial management of self-harm. The guideline was developed following extensive literature reviews, two focus groups with users and a lengthy consultation process. The main recommendations are uncontroversial and will be regarded by many as simply components of good practice. For example: treating patients who self-harm with care, respect and privacy; providing appropriate training to front-line staff; offering a preliminary psychosocial assessment to all patients; basing further treatment on a comprehensive assessment.

Other recommendations may be more challenging to implement across psychiatric services. The guidelines seem to suggest that all individuals who self-harm should receive an assessment by a mental health specialist. However, most specialists would agree that staff in acute medical settings are able to carry out adequate assessments if appropriately trained and supervised (Royal College of Psychiatrists, 1994, 2004). For patients at risk of repetition (and at least one in six will repeat within a year), the guidelines state that services should consider offering an intensive therapeutic intervention combined with outreach. This should last for at least 3 months and allow frequent access to a therapist, telephone contact and home treatment when necessary, and active follow-up when appointments have been missed. Unfortunately, given the current state of self-harm services this level of intervention seems unrealistic.

The Royal College of Psychiatrists report (2004) updates the 1994 document. It describes the clinical competencies that might be expected of both non-specialists non-specialists and specialists. General skills include assessment and treatment of the patient's physical condition, preliminary psychosocial assessment and a basic understanding of medico-legal issues. Specialist skills include providing a diagnostic formulation, assessing risk, and drawing up and implementing a treatment plan. The report also describes standards for the organisation of services, clinical procedures and facilities, and training and supervision in a variety of settings (the emergency department, the general hospital, the community setting and the psychiatric in-patient unit). The report seems particularly relevant to those planning services for self-harm. It acknowledges that some of the recommendations may be difficult to implement in smaller districts. For example, emergency department staff having access to self-harm specialists within 30 minutes in urban areas.

Both of the recent documents appear clinically sensible but ensuring their implementation may be extremely difficult. Guidelines are more likely to be adopted when there is strong professional support, no increased costs associated with their implementation, a system in place to monitor take up, and a strong evidence base (Reference Sheldon, Cullum and DawsonSheldon et al, 2004). Unfortunately, much of the evidence for the NICE guideline rated no higher than ‘GPP’ (good practice point based on the clinical experience of the guideline development group).

WHAT WORKS FOR SELF-HARM?

Various treatments have been evaluated but very few have led to clinically significant reductions in repetition. Systematic reviews have concluded that trials have been too small and have tended to recruit specific subgroups of individuals making the findings difficult to generalise (Reference Hawton, Townsend and ArensmanHawton et al, 1999; National Collaborating Centre for Mental Health, 2004).

A number of interventions probably warrant further investigation in large clinical trials. These include problem-solving therapy, interpersonal treatments and ‘emergency card’ type interventions. Other treatments may be helpful for subgroups of patients (e.g. dialectical–behavioural therapy for individuals who self-harm repeatedly, group therapy for adolescents). The largely negative results of some of the bigger trials (Reference Tyrer, Thompson and SchmidtTyrer et al, 2003) have led some investigators to argue that perhaps we should concentrate either on very large trials of low-intensity interventions (such as emergency cards or letter-writing interventions), or smaller trials of longer-term, more intensive psychological treatments (J. M. G. Williams, personal communication, 2005). Recent studies have provided some support for both approaches (Reference Brown, Ten Have and HenriquesBrown et al, 2005; Reference Carter, Clover and WhyteCarter et al, 2005). However, because one of the difficulties in this area of research is ensuring that patients actually receive the assigned treatment, an alternative might be a large-scale evaluation of a brief psychological intervention that specifically addresses issues related to engagement early on in the therapy. We should probably consider outcomes other than repeat presentation to hospital (such as self-reported self-harm, depression, hopelessness, loss of contact with services, quality of life and user satisfaction). Alternative methodological approaches may also be of benefit, such as qualitative or cohort study designs. A system for the multi-centre monitoring of self-harm is being implemented in England (Department of Health, 2002). This will provide valuable epidemiological data as well as allowing an investigation of the outcomes of treatments given in day-to-day practice. Although such approaches avoid the selection bias inherent in clinical trials, adjusting for relevant confounding variables can sometimes be problematic. It is therefore important that clinical databases are large, carefully constructed and measure all relevant outcomes in a standardised way (Reference Gilbody, House and SheldonGilbody et al, 2002).

HOW SHOULD MENTAL HEALTH SERVICES MANAGE SELF-HARM?

The clinical evaluation of patients following self-harm has been referred to as one of the most complex assessments in psychiatry (Reference Isacsson and RichIsacsson & Rich, 2001). In the context of current haphazard service provision (Reference Bennewith, Gunnell and PetersBennewith et al, 2004) and a lack of research evidence, how should mental health services manage self-harm?

The recent guidelines could help but attitudes among those responsible for providing services also need to change in order to ensure appropriate management of this patient group. There is still a body of opinion that views those who self-harm as immature individuals who divert resources from those with ‘serious’ physical or psychiatric illness (Reference JamesJames, 2004). One way of addressing these negative attitudes could be to involve users and carers in professional training, service delivery and service evaluation (Reference Simpson and HouseSimpson & House, 2003).

What other measures might improve our management of self-harm? One of the first tasks for services could be to broaden their priorities to include psychiatry in the general hospital as well as major mental illness in the community. Multidisciplinary self-harm teams are not a new idea and have several potential benefits – the range of available interventions is increased, a wide range of skills can be shared, administrative efficiency and speed of response may be improved, and the team approach helps to maintain morale in a service dealing with a complex patient group. Despite this, many services still consist of on-call junior psychiatrists carrying out rushed assessments between other commitments on a rota basis.

Reducing our preoccupation with risk assessment may help to improve the general hospital management of self-harm. Predicting the risk of future suicidal behaviour following self-harm is problematic because the outcomes we are interested in are rare and our assessment tools are relatively crude. There seems to be growing recognition of this fact, with a change in emphasis from ‘risk assessment’ to ‘needs assessment’ in recent guidance (National Collaborating Centre for Mental Health, 2004). A needs assessment aims to identify psychosocial factors that might explain an act of self-harm. This will lead to a formulation (describing short- and long-term vulnerability factors and precipitating factors) which will directly inform the management plan.

Follow-up after self-harm is important and perhaps the key thing about after-care is that it should be provided promptly – of those who repeat self-harm within a year of an episode, about a quarter do so within 3 weeks. However, these patients may be difficult to engage. Strategies that could be used to improve uptake of treatment include home visits, the use of written prompts, and after-care being provided by the health professional who carried out the initial assessment.

What form should this intervention take? Comorbid mental disorder and the suicidal risk associated with it should be managed in the usual way. The choice of any additional psychological therapy will be determined by the diagnostic formulation, or perhaps more pragmatically by local availability. Targeting intervention solely at individuals assessed to be at ‘high risk’ of further suicidal behaviour may represent an efficient use of resources but is less than ideal. This is because the large number of individuals assessed as at ‘low risk’ actually account for the majority of repeat episodes (Reference Kapur, Cooper and RodwayKapur et al, 2005). An alternative model of intervention would be to offer a basic intervention to all those who have harmed themselves, and use a combination of needs and risk assessment to identify individuals who might benefit from more intensive treatment.