NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Mental Health (UK). Self-Harm: The Short-Term Physical and Psychological Management and Secondary Prevention of Self-Harm in Primary and Secondary Care. Leicester (UK): British Psychological Society; 2004. (NICE Clinical Guidelines, No. 16.)

Cover of Self-Harm

Self-Harm: The Short-Term Physical and Psychological Management and Secondary Prevention of Self-Harm in Primary and Secondary Care.

Show details

2Introduction to self-harm

2.1. What is self-harm and what does the guideline cover?

The guideline has adopted the definition that self-harm is ‘self-poisoning or self-injury, irrespective of the apparent purpose of the act’. The guideline has, therefore, used a shorter and broader definition than that adopted by the World Health Organization (‘an act with non-fatal outcome, in which an individual deliberately initiates a non-habitual behaviour that, without intervention from others, will cause self-harm, or deliberately ingests a substance in excess of the prescribed or generally recognised therapeutic dosage, and which is aimed at realising changes which the subject desired via the actual or expected physical consequences’) (Platt et al., 1992). The guideline focuses on those acts of self-harm that are an expression of personal distress and where the person directly intends to injure him/herself. It is important also to acknowledge that for some people, especially those who have been abused as children, acts of self-harm occur seemingly out of the person’s control or even awareness, during ‘trance-like’, or dissociative, states. It therefore uses the term ‘self-harm’ rather than ‘deliberate self-harm’.

The scope has been limited in this way because the term ‘self-harm’ is a broad one and could be applied to the actions of many people at some time in their lives. Many behaviours that are culturally acceptable can result in self-inflicted physical or psychological damage, such as smoking, recreational drug use, excessive alcohol consumption, over-eating or dieting. Also, self-harm can occur as part of religious practice, as a form of political or social protest or as an act of ‘body enhancement’ (Babiker & Arnold, 1997; Walsh & Rosen, 1988).

Even when these types of self-harm are excluded, the guideline must address the needs of people whose self-harm varies greatly in its nature and meaning.

Box 1 gives five vignettes that, while by no means encompassing all types of circumstance in which self-harm occurs, illustrate the extent of this diversity. Also, the outcome for the people described by the vignettes would be very different in the absence of an intervention by care services.

Box Icon

Box 1

Five vignettes to illustrate the diversity of self-harm that falls within the remit of the guideline A 55-year-old bank manager, married for 30 years and a mother of three children. She has had no recent major adverse life events. At age 30 she suffered (more...)

NICE guidelines are principally for those who work for or use NHS services. The emphasis is therefore on the care of those people whose act of self-harm brings them to the attention of statutory services. Although this is perhaps only a minority of people who self-harm, they are an important group both because they are statistically at much higher risk of suicide than the rest of the population and because they often report that services fail to meet their needs. The guideline is limited to how services should respond in the 48 hours after an episode of self-harm, and the effectiveness of follow-up treatments; it does not consider in detail the longer-term care of people who self-harm, including those who self-harm repeatedly. Finally, the guideline does not address the needs of people who have thoughts of self-harm or of suicide but do not act on these.

2.2. A note about terminology

The language of healthcare is always evolving. Also, healthcare workers from different professional backgrounds sometimes use different terms to describe the same concept. Furthermore, some terms are unacceptable to some service users. These differences reflect differing perspectives and can also sometimes exacerbate divisions. The Guideline Development Group had many discussions about terminology. It reached agreement on some issues; for example not to use the words ‘deliberate’ or ‘intentional’ to prefix self-harm and not to use the word ‘commit’ in relation to suicide. Many service users object to these terms, especially those who harm themselves during dissociative states, afterwards being unaware of any conscious intent to have harmed themselves. Also, it can be argued that prefixing the term ‘self-harm’ with ‘intentional’ would suggest that there may be accidental and non-intentional forms of self-harm. Clearly, the non-intentional forms, such as those carried out during dissociative states, are covered by the term ‘self-harm’ alone. And the suggestion that a person may accidentally self-harm would be misleading: we simply say that the person has had an accident.

Box 2 lists some of the terms that are commonly used to describe the behaviour that is the subject of this guideline.

Box Icon

Box 2

Common terms used to describe self-harm

2.3. Why do people self-harm?

As vignette 1 shows, an individual episode of self-harm might be an attempt to end life. However, many acts of self-harm are not directly connected to suicidal intent. They may be an attempt to communicate with others, to influence or to secure help or care from others or a way of obtaining relief from a difficult and otherwise overwhelming situation or emotional state (Hjelmeland et al., 2002). Paradoxically, the purpose of some acts of self-harm is to preserve life (as illustrated by vignettes 3 and 5). Professionals sometimes find this a difficult concept to understand.

One particular intention or motive might predominate or all might co-exist. This means that a person who self-harms repeatedly might not always do so for the same reason each time, or by the same method (Horrocks et al., 2003). Thus assumptions about intent should not be made on the basis of a previous pattern of self-harm; each act must be assessed separately to determine the motivation behind it. Failure to do this can result in the meaning of the act being misunderstood and in an interpretation that the service user finds judgemental or dismissive. This will inevitably lead to a breakdown in the therapeutic relationship, as well as making it less likely that appropriate help will be offered at times when a person is at high risk of suicide.

Consistent with these differences in intention and motive, people who self-harm might have very different expectations about how health services should respond and what constitutes a good outcome. In particular, people who harm themselves as a way of relieving distress (through cutting, for example) might be compelled to do this as a coping and suicide prevention strategy (as with the woman in vignette 5). They are likely to continue to need to do this until they receive appropriate and sufficient psychotherapeutic interventions and support.

Research into motivation and related risk factors comes from two main sources: asking service users at interview about their motivation to self-harm (usually recorded on checklists) and self-report. The interview and checklist approach is limited by the fact that the precise reason for a person to self-harm tends to be a very individual matter, and the same person may self-harm on different occasions for different reasons. The gain in statistical power using standard questionnaires may well be offset by the loss in specificity. The result is that scales to predict suicide have weak predictive power, because the absolute risk of suicide is so low (Dennehy et al., 1996) (see Chapter 8).

2.4. Methods of self-harm

The methods of self-harm can be divided into two broad groups: self-poisoning and self-injury. People who self-poison are more likely to seek help than those who self-injure (Hawton et al., 2002a; Meltzer et al., 2002a). For this reason, studies that focus on people who attend emergency departments paint a different picture about the respective prevalence of these two forms of self-harm from studies of the general population.

About 80% of people who present to emergency departments following self-harm will have taken an overdose of prescribed or over-the-counter medication (Horrocks et al., 2003). A small additional percentage will have intentionally taken a dangerously large amount of an illicit drug or have poisoned themselves with some other substance. The pattern of the type of drug taken in overdose has changed in recent years, largely with changes in their availability.

Recent studies of the method of suicide suggest that people who survive a medically serious suicide attempt may well have a poorer outcome in terms of life expectancy. For example, in a prospective study of 302 people who had made a serious attempt on their life, 1 in 11 had died within 5 years, with nearly 60% of deaths being by suicide, and a greater than expected number dying by vehicular accident (Beautrais, 2003). The method of self-harm may also depend on gender, with self-poisoning more frequent in adolescent females (92.4%) than in adolescent males (84.6%), and self-injury more frequent in adolescent males (15.4%) than in adolescent females (7.6%) (Hawton et al., 2003c). Table 1 lists those substances most commonly reported to the National Poisons Information Service (NPIS) London Centre (and therefore an approximate index of which drugs are taken in the more serious overdoses) as having been taken during acts of self-poisoning in 2002. Table 2 gives those drugs most frequently taken by people in self-poisoning incidents recorded by four surveys.

Table 1. Substances reported to the National Poisons Information Service (NPIS) London Centre in 2002 as implicated in self-poisoning.

Table 1

Substances reported to the National Poisons Information Service (NPIS) London Centre in 2002 as implicated in self-poisoning.

Table 2. Poisons implicated in self-poisoning episodes. The figures for each drug are percentages, rounded to the nearest whole number.

Table 2

Poisons implicated in self-poisoning episodes. The figures for each drug are percentages, rounded to the nearest whole number.

In contrast to those who attend emergency departments, self-injury is more common than self-poisoning in the population as a whole, perhaps by a ratio of 2:1 in teenagers (Hawton et al., 2002a). Cutting is by far the most common means (Hawton et al., 2002a; Horrocks et al., 2003). Less common methods include burning, hanging, stabbing, swallowing objects, insertion, shooting and jumping from heights or in front of vehicles. The specific medical and surgical treatments for these less common methods of self-injury are not covered in this guideline, although the general principles of care are likely to be similar.

2.5. How common is self-harm?

Since many acts of self-harm do not come to the attention of healthcare services, hospital attendance rates do not reflect the true scale of the problem (Hawton et al., 2002a; Meltzer et al., 2002b). A national interview survey suggested that in Great Britain between 4.6% and 6.6% of people have self-harmed (Meltzer et al., 2002a). However, even this might be an under-estimate. In a school survey, 13% of young people aged 15 or 16 reported having self-harmed at some time in their lives and 7% as having done so in the previous year (Hawton et al., 2002a).

Overall, women are more likely to self-harm than men. This is most pronounced in adolescence, where girls may be three times more likely to self-harm than boys (Hawton et al., 2002a).

Self-harm can occur at any age but is most common in adolescence and young adulthood (Meltzer et al., 2002a). For example, based on a survey of 12,529 children and young people aged 5 years to 15 years, 1.3% had tried to harm themselves (data collected from parents) (Meltzer, 2002a). In contrast, only about 5% of all episodes of self-harm occur in people over the age of 65 (Dennis et al., 1997; Draper, 1996; Owens et al., 1991).

2.6. Factors that are associated with self-harm

2.6.1. Socio-economic factors and life events

Self-harm occurs in all sections of the population but is more common among people who are disadvantaged in socio-economic terms and among those who are single or divorced, live alone, are single parents or have a severe lack of social support (Meltzer et al., 2002a).

Life events are strongly associated with self-harm in two ways. First, there is a strong relationship between the likelihood of self-harm and the number and type of adverse events that a person reports having experienced during the course of his/her life. These include having suffered victimisation and, in particular, sexual abuse (Hawton et al., 2002a; Meltzer et al., 2002a). Second, life events, particularly relationship problems, can precipitate an act of self-harm (Bancroft et al., 1977).

Many people who self-harm have a physical illness at the time and a substantial proportion of these report that this is the factor that precipitated the act (De Leo et al., 1999).

2.6.2. The association between self-harm and mental disorder

Most of those who attend an emergency department following an act of self-harm will meet criteria for one or more psychiatric diagnoses at the time they are assessed (Haw et al., 2001). More than two-thirds would be diagnosed as having depression although within 12–16 months two-thirds of these will no longer fulfil diagnostic criteria for depression.

People diagnosed as having certain types of mental disorder are much more likely to self-harm. For this group, the recognition and treatment of these disorders can be an important component of care. In one survey of a sample of the British population, people with current symptoms of a mental disorder were up to 20 times more likely to report having harmed themselves in the past (Meltzer et al., 2002a). The association was particularly strong for those diagnosed as having phobic and psychotic disorders. People diagnosed as having schizophrenia are most at risk and about one-half of this group will have harmed themselves at some time.

Certain psychological characteristics are more common among the group of people who self-harm, including impulsivity, poor problem-solving and hopelessness. Also, people who self-harm more often have interpersonal difficulties. It is possible to apply diagnostic criteria to these characteristics. This explains why nearly one-half of those who present to an emergency department meet criteria for having a personality disorder (Haw et al., 2001). However, there are problems with doing this because:

2.6.3. The association between self-harm and alcohol and drug use

About one-half of people who attend an emergency department following self-harm will have consumed alcohol immediately preceding or as part of the self-harm episode (Merrill et al., 1992; Horrocks et al., 2003). For many, this is a factor that complicates immediate management either by impairing judgement and capacity, or by adding to the toxic effects of ingested substances. About one-quarter of those who self-harm will have a diagnosis of harmful use of alcohol (Haw et al., 2001). Men are more likely to drink before an episode of self-harm than women (Hawton et al., 2003d), and are more likely to be misusing drugs or alcohol, as well as to have higher rates of several risk factors for suicide (Taylor et al., 1999).

2.6.4. The association between self-harm and child abuse and domestic violence

Child sexual abuse is known to be associated with self-harm (Hawton et al., 2002a; Meltzer et al., 2002a), especially among people who repeatedly self-harm, as well as a range of mental health problems in adolescence and adulthood for females in particular. In one review of 45 studies on the effects of childhood sexual abuse, immediate and long-term symptoms of severe mental distress were typical and included: anxiety, depression, substance misuse, self-destructive behaviour, suicide and a tendency towards self-harm and re-victimisation (Kendall-Tackett et al., 1993).

Prevalence rates for child sexual abuse are high: 20% to 30% for women, and 10% for men (Itzin, 2000), with 25% of women (and their children) experiencing domestic violence in their lifetime (Kershaw et al., 2000). Underlying this are high rates of physical abuse. In the NSPCC child maltreatment study more than a fifth of children experienced physical violence regularly, 12% reported injuries and 7% suffered serious physical abuse (Cawson et al., 2000). Comorbidity of domestic violence and child sexual and physical abuse is common (Hester, 2000).

It is important to note that socio-economic factors, such as unemployment and poverty, childhood experiences of abuse, and experiences of domestic violence are all associated with a wide range of mental disorders, as well as self-harm. How these experiences and factors interact needs to be explored and better understood. However, the NICE guideline on the treatment of PTSD in children and adults notes that there is a strong association between childhood sexual abuse and the development of childhood PTSD; the later emergence of self-harming behaviour may well be a later response to this (Posttraumatic Stress Disorder: The Management of PTSD in Primary and Secondary Care, forthcoming 2005; details available from NCCMH upon request).

2.7. Special groups

2.7.1. Diversity

There is no good evidence to suggest that the incidence of self-harm varies between different ethnic groups, with the exception of higher rates in young Asian women (Bhugra et al., 1999).

A recent survey of gay men, lesbians and bisexuals has suggested that this group of people may have an increased rate of self-harm, which some groups reported was linked to their sexual orientation (King & Mckeown, 2003). However, the report also identified higher rates of bullying and victimisation among these groups, factors also linked to self-harm. It may be that these groups of people are bullied more often and that this may lead some to self-harm.

2.7.2. Young people

The rate of self-harm is relatively low in early childhood, but increases rapidly with the onset of adolescence (Hawton et al., 2003c). Most acts of self-harm in young people never come to the attention of care services (Hawton et al., 2002a) and it is also likely that many parents are unaware of the problem (Meltzer et al., 2002a). Although clearly a manifestation of distress, self-harm in young people is often a ‘marker’ for the presence of other problems that might have an important bearing on outcome, such as substance misuse, poor school attendance, low academic achievement and unprotected sex (Kerfoot, 1998; King et al., 2001). Other issues relevant for young people include bullying, domestic violence, victimisation and child sexual and physical abuse.

2.7.3. Older people

Although it appears that older people are less likely to self-harm, the consequences are often more serious; it has been estimated that of every five older people who self-harm one will later die by suicide (Lawrence et al., 2000; McIntosh, 1992). Consistent with this, older people who have self-harmed score highly on scales that measure suicidal intent (Merrill & Owens, 1990; Nowers, 1993) and their profile resembles that of older people who die by suicide (Dennis & Lindesay, 1995). In particular, older people who self-harm have high rates of physical ill health, social isolation and depression (Draper, 1996; Merrill & Owens, 1990; Pierce, 1977). Those with persistent depression are at particular risk of repetition of self-harm or suicide (Hepple & Quinton, 1997).

2.7.4. People with a learning disability

For those working with people with a learning disability, the term self-harm usually refers to ‘self-injurious behaviour’ (SIB), which includes ‘head banging’ and ‘nail biting’. The prevalence of SIB varies between 17% and 24% and is more common in women and girls, those with very low IQ, with communication difficulties and with certain genetic disorders (Deb, 1998; Deb et al., 2001). The management of SIB in people with a learning disability is outside of the scope of this guideline.

There has been little research about the prevalence and management of self-harm, of a type that is the focus of this guideline, in people with a learning disability.

2.7.5. People within the criminal justice system

Self-harm is much more common among prisoners than among the general population. One-half of female remand prisoners have self-harmed at some time in their lives and more than one-quarter in the previous year. The corresponding figures for men are about half of these. Perhaps as many as 10% of prisoners will self-harm during their term with the likelihood increasing with the length of time in custody. Highest rates are among sentenced female prisoners who have spent two or more years in prison, 23% of whom will have self-harmed during the current term (Meltzer et al., 1999).

This high rate is largely explained by the fact that, among the prison population, there are much higher levels of the factors associated with self-harm. For example, between 12% and 21% of prisoners have at least four mental disorders simultaneously (including drug and alcohol dependence, personality disorder, neurotic disorder and psychosis); between 35% and 52% are dependent on opiates, stimulants or both; 20%–30% are severely dependent on alcohol; about one-half of female prisoners report having suffered violence in the home; and 10% of men and 33% of women report previous sexual abuse (Singleton et al., 1998).

Cutting or scratching is the most common method of self-harm in prison. In contrast 90% of suicides in prison are by hanging and self-strangulation, although these account for only one-fifth of self-harm incidents.

The Safer Custody Group of HM Prison Service is gathering increasing evidence about the link between prison conditions and the frequency of self-harm and of suicide (HM Prison Service, 2001).

The problem of self-harm is probably even greater among women in high security hospitals (Wilkins & Warner, 2001).

2.8. The consequences of self-harm

2.8.1. Repetition and suicide

Following an act of self-harm the rate of suicide increases to between 50 and 100 times the rate of suicide in the general population (Hawton et al., 2003b; Owens et al., 2002). Men who self-harm are more than twice as likely to die by suicide as women and the risk increases greatly with age for both genders (Hawton et al., 2003b). It has been estimated that one-quarter of all people who die by suicide would have attended a general hospital following an act of self-harm in the previous year (Owens & House, 1994).

About one in six people who attend an emergency department following self-harm will self-harm again in the following year (Owens et al., 2002); a small minority of people will do so repeatedly. The frequency with which some of the latter group self-harm means that they are over-represented among those who present at an emergency department or receive psychiatric care. There is no good evidence to support the widely voiced opinion that people who harm themselves repeatedly, particularly by cutting, are less likely to die by suicide.

2.8.2. Physical health

Regardless of the person’s intention, self-harm can result in long-lasting ill health or disability. Paracetamol poisoning is a major cause of acute liver failure requiring liver transplantation. Between 1998 and 2002, 111 liver transplants were carried out in England and Wales on people who had taken an overdose of paracetamol. This accounted for 4% of all liver transplants but 23% of all ‘super-urgent’ transplants –those in which the person is expected to die within 72 hours from fulminant liver failure (Chris Rudge, Medical Director UK Transplant, personal communication 2003). Self-cutting can result in permanent damage to tendons and nerves and scarring leading to disfigurement. More violent forms of self-injury often lead to permanent disability and/or hospitalisation.

2.8.3. The economic cost of self-harm

The assessment and treatment of people who self-harm uses a substantial amount of NHS resources. Most of this direct cost is accounted for by the estimated 150,000–170,000 attendances at an emergency department each year and the subsequent medical and psychiatric care (Yeo, 1993). Self-harm resulted in 68,716 hospital admissions in 2001/02. At the same time, over the past 10 years, the average daily number of NHS beds available for mental illness in England has almost halved with 63,000 beds available in the year 1988/89 and only 34,000 available in 2000/01 (Department of Health, 2003b). As one of the most common presentations to general hospitals and one which has a strong tendency for recurrence and increased severity, self-harm presents a considerable economic burden to the individual, family, health services, and society as a whole.

Despite the importance to healthcare authorities, there have been precious few economic evaluations of self-harm that meet rigorous criteria for health economic evaluation (Drummond & Jefferson, 1996). The extent of economic burden associated with self-harm is dependent upon both how it is defined and the method of economic evaluation (Drummond et al., 1997). Self-harm is associated with direct and indirect costs. Direct costs are incurred in the course of recognising, caring for and treating self-harm patients through primary care, secondary care and social care. Indirect costs include the effects of illness on work attendance and productivity, employer benefits (if any) extended to the individual, malpractice insurance and legal costs, cost of long-term disability and premature mortality, and intangible costs that may extend beyond the individual and her/his immediate family.

It is estimated that up to 20% of overdoses involve ingestion of antidepressants and that the rate of self-poisoning by this modality is increasing (Kapur et al., 1998; Hawton et al., 2003d). In terms of general hospital costs of antidepressant overdose, based on 240 episodes of self-poisoning over a five-month period in three teaching hospitals and three district hospitals in the UK, the total hospital cost for overdoses of SSRIs versus TCAs was estimated at £17,117 and £78,612 respectively (1999/2000 prices) (Kapur et al., 2001). Per self-poisoning episode, the additional hospital cost of TCA poisoning compared with SSRI poisoning was £461, largely due to the increased number of inpatient days in the intensive care unit, which accounted for a cumulative additional cost of £5.1 million per year in the UK (Kapur et al., 2002a).

The indirect costs of self-harm are unknown but, given its prevalence, are likely to be substantial, particularly in terms of days lost from work. However, what we do know is that after cardiovascular disease and cancer, suicide is the next most common cause of life years lost (Gunnell & Frankel, 1994), making self-harm and suicide important public health issues.

It is very difficult to evaluate the cost of treatments for self-harm because of the heterogeneous nature of self-harm and because there is little evidence about treatment effectiveness (see Chapter 9).

2.9. Contact with services

It is likely that many acts of self-harm do not come to the attention of healthcare workers and that only a minority result in assessment by specialist mental health services. A person who self-harms may seek advice or care from a variety of sources. These can be conceptualised as being at different levels of the care system. These levels also indicate the potential paths through care that an individual may follow. The picture of the number of people who present at each level and the extent to which they cross filters is incomplete.

Level 1: family, friends and acquaintances. The three-fold difference in prevalence of self-harm as reported by young people and by their parents (Meltzer et al., 2002a) suggests that many acts of self-harm in the young do not come to the attention of their families.

Level 2: contact with workers not employed by the health service. This might be either face-to-face (for example with a teacher, a counsellor, a police or prison officer, a social worker or a worker from a voluntary sector agency) or with a person staffing a help-line. The last are frequently used by people who have self-harmed or who believe themselves to be at risk of doing so. In 2001, the Samaritans had more than 3 million verbal contacts. They estimate that their volunteers explored suicidal feelings with the caller in more than one-quarter of these.

Level 3: primary care health workers and ambulance staff. This includes members of the primary care team, ambulance staff and NHS Direct. About one-half of people who attend an emergency department following self-harm will have visited their GP during the previous month and about the same proportion will do so in the two months afterwards (NHS Centre for Reviews and Dissemination, 1998).

People who have self-harmed, their friends and their relatives often turn to the emergency ambulance service for help in resolving incidents of self-harm. Ambulance staff are increasingly better trained and skilled to provide essential care and treatment at the scene and during transportation to hospital for patients with diverse medical conditions. They are placed in a privileged position to provide immediate treatment and psychological support for patients who have self-harmed prior to their arrival at a place of safety. They often have access to the person’s living environment and gain an insight from family and friends, who often are not present during hospital treatment, on events leading up to the incident of self-harm.

Level 4: accident and emergency departments. There are about 150,000 attendances at accident and emergency departments for self-harm each year. This is about 300 attendances per 100,000 of the population (Hawton et al., 2003d; Kapur et al., 1998), although the actual rates vary greatly in different parts of the country (Gunnell et al., 1996).

Level 5: secondary care health services:


hospital medical and surgical care. Self-harm is one of the top five causes of acute medical admission in the UK (Hawton & Fagg, 1992; Gunnell et al., 1996). The proportion of those who attend an emergency department following self-harm, who are then admitted to a medical or surgical ward, varies across different parts of the country, with a usual minimum in reported figures of 40% (Hawton et al., 1997; Horrocks et al., 2003). For children and young people, the consensus is that admission to a paediatric ward for time to ‘cool off’, to undertake assessment of the child and family, and to address child protection issues, should these arise, should be the normal course of events (Royal College of Psychiatrists, 1998).


mental health services. Perhaps one-half of those who present to an emergency department are assessed by a mental healthcare worker (Kapur et al., 1998; Horrocks et al., 2003) and between 5% and 10% will be admitted to a psychiatric ward (Horrocks et al., 2003). Some services achieve much higher rates of assessment (Hawton et al., 2003c). Once again there is wide variation between centres.

Level 6: tertiary services. The National Poisons Information Service provides a 24-hour telephone information service and maintains TOXBASE, which gives online information about the clinical management of poisoning. Both are free of charge to all NHS staff. Toxicology units are a further important component of services for people who have self-poisoned. In exceptional cases, people who have poisoned themselves might be transferred to a specialist centre.

Other health service facilities at this level include burns units, plastic surgery units and specialist tertiary psychotherapy and therapeutic community units, which sometimes offer help to people who repeatedly harm themselves.

At each of these levels there are opportunities for improving the rate of detection of acts of self-harm and for identifying those who would benefit from more specialist help and so should pass through the ‘filter’ to a higher level of care. This particularly relates to interfaces between the levels that involve services managed by the health service. For example, there is a danger that important information gathered by the ambulance service, about the circumstances of self-harm, is not always received at an emergency department. Also, as mentioned above, a substantial number of people who present to an emergency department are discharged or decide to leave before a psychosocial assessment has been carried out.

2.10. How people who self-harm experience services

The importance of the experience of assessment, treatment and care for people who have self-harmed should not be underestimated. Many people who self-harm do not come to the attention of health services, and when they do, many do not return or are lost to follow up. Service users describe contact with health services as often difficult, characterised by ignorance, negative attitudes and, sometimes, punitive behaviour by professionals towards people who self-harm. With the risk of death by suicide being considerably higher among people who have self-harmed, whatever the expressed intent, and with their high rates of mental health problems, and alcohol and substance misuse, it is no longer acceptable for healthcare professionals to ignore, or fail properly to address, the experience of care by service users and carers. Engaging service users in a therapeutic alliance and promoting joint clinical decision-making on the basis of understanding and compassion is essential, especially if further help and treatment are to be offered.

To examine the current experience of service users and carers, this guideline reviewed the service user literature, and arranged two focus groups and an individual interview with a service user; all participants were contacted through two national self-harm service user organisations. The service users were also asked to identify the changes they would want to make in the general approach to treatment as currently experienced. To gain confirmation of service user and carer views from a different perspective, an existing review of health professional attitudes to self-harm was also identified and reviewed. The findings and recommendations regarding the experience of care can be found in Chapter 5.

2.11. Assessment and treatment for people who self-harm

2.11.1. Aims and principles of treatment

As with any other treatment, the overarching aims are to reduce harm and improve survival while minimising the harm that may result from the treatment. In addition, the experience of treatment and care needs to be acceptable to service users and carers. This is especially so for people who self-harm and who may be suffering psychological, social or drug- and alcohol-related problems, which need further help after the immediate physical problems have been adequately addressed. It is essential that service users and carers, where appropriate, are engaged effectively by clinicians in an atmosphere of respect and trust. Without this, further psychosocial assessment and referral for treatment will be difficult if not impossible. Issues of consent and ethics are considered in Chapter 6.

The key aims and objectives in the treatment of self-harm should, therefore, include:

  • Rapid assessment of physical and psychological need (triage)
  • Effective engagement of service user (and carers where appropriate)
  • Effective measures to minimise pain and discomfort
  • Timely initiation of treatment, irrespective of the cause of self-harm
  • Harm reduction (from injury and treatment; short-term and longer-term)
  • Rapid and supportive psychosocial assessment (including risk assessment and comorbidity)
  • Prompt referral for further psychological, social and psychiatric assessment and treatment when necessary
  • Prompt and effective psychological and psychiatric treatment when necessary
  • An integrated and planned approach to the problems of people who self-harm, involving primary and secondary care, mental and physical healthcare personnel and services, and appropriate voluntary organisations
  • Ensuring that the special issues applying to children and young people who have self-harmed are properly addressed, such as child protection issues, confidentiality, consent and competence.

A flowchart detailing the ‘journey’ a service user may take within services is given in Appendix 2.

2.11.2. Primary care and pre-hospital environment

Primary care professionals, community-based mental health workers, ambulance staff and others come into contact with people who have self-harmed with varying frequency. In this context, assessment and referral to the emergency department is the most common action undertaken, in particular for self-poisoning. Sometimes, self-injury (but not self-poisoning) will be dealt with in primary care without referral for further physical treatment, usually by sympathetic GPs who have already had contact with the person. How many people who self-injure are treated in primary care is difficult to ascertain. Often, on-site counselling and psychological therapy services are available in primary care. In any event, psychosocial assessment should be undertaken by a professional trained to do so at the earliest opportunity. Treatment and care in these settings are considered in Chapter 7.

2.11.3. The emergency department and triage

On entry to the emergency department all patients undergo triage, a utilitarian system using predefined criteria in which the urgency with which a person needs treatment is evaluated using a formal and structured assessment. Triage allows patients to be categorised according to their needs and the urgency for treatment. It is widely regarded as essential in busy emergency departments, where overcrowding and shortage of resources are common (Brillman et al., 1997). The triage system most commonly used in the UK is the Manchester Emergency Triage system (Mackway-Jones, 1996), which gives priority to patients largely according to their physical state. The role and tools used for triage are reviewed in Chapter 7.

Following triage, patients who have self-harmed should receive the requisite treatment for their physical condition, undergo risk and full psychosocial needs assessment and mental state examination, and referral for further treatment and care as necessary.

The same principles and services should be available for children, who should be treated either in a children’s emergency department or in a separate area of the emergency department away from adults. Children also have a range of additional special needs, and for most children and young people admission overnight is indicated (unless they present to services earlier in the day) and psychosocial assessment can be undertaken the next day, including parents and significant others.

2.11.4. Medical and surgical treatment of self-harm

A range of medical and surgical interventions is available for the physical treatment of people who have self-harmed, including general and specific treatments for self-poisoning and self-injury. For this guideline, reviews of the evidence base of the following interventions have been undertaken, the findings and recommendations for which can be found in Chapter 7: triage, paracetamol screening, the general management of ingestion using gut decontamination, the specific treatment of overdose of paracetamol, benzodiazepine, opioid and other substances, and the treatment of superficial wounds. The closure of more complex wounds is beyond the scope of this guideline.

2.11.5. Assessment and psychological and pharmacological treatment of people who self-harm

The psychosocial and risk assessment of people attending emergency departments in the UK has been described as inadequate, characterised by low assessment rates and poor recording of mental health findings (Merrill et al., 1992). Nevertheless, much progress has been made in improving services for this group of people in some parts of the UK, and a number of emergency departments work closely with mental health services. However, given the potentially serious consequences of self-harm, with a significant number of people going on to kill themselves at a later date, effective and reliable means of assessment should be a priority for people who self-harm and present to services. In addition, following assessment, patients may be referred for further pharmacological or psychological treatment in inpatient, outpatient or other settings.

For this guideline we have undertaken a review of risk and needs assessment following self-harm (see Chapter 8). Also, psychological and pharmacological treatments specifically for people who self-harm have been reviewed for this guideline to help guide referral for additional treatment (see Chapter 9). Drug treatments reviewed include antidepressants and antipsychotics. Psychological treatments reviewed are: problem-oriented therapies, dialectical behaviour therapy, inpatient behaviour therapy and insight-oriented therapy, long- and short-term therapy, home-based family therapy and group therapy. Other psychosocial and service-level interventions are also reviewed in Chapter 9.

2.11.6. The relationship between emergency departments and mental health services for people who have self-harmed

Most people who self-harm and who present to health services will attend an accident and emergency department located on the site of a general hospital. As well as access to medical and surgical beds, many emergency departments have short-stay ward facilities where people who have taken an overdose can recover. Emergency departments must also be able to offer psychosocial assessments; this requires an interview room with adequate safety features.

With the exception of some minor injury units, which are staffed by nurses alone, all emergency departments are run by a team of doctors and nurses, including emergency nurse practitioners, with training in emergency medicine. Few employ staff with specialist mental health skills; it is therefore essential that there is a close working relationship between emergency departments and the local mental health service.

The structural relationship between emergency departments and mental health services varies greatly. In most parts of England and Wales, the two are managed by different NHS trusts and in some places there are no mental health services based on the same site as an emergency department. Also, in some settings there is no liaison psychiatrist dedicated to act as the bridge between mental health services and the general hospital. Regardless of the local service configuration, the guideline endorses the recommendations of the Royal College of Psychiatrists and the British Association for Accident and Emergency Medicine (Royal College of Psychiatrists, 2004), that:

  • There is a joint responsibility for commissioners, mental health service managers and acute service managers to ensure that the input of mental health services to emergency departments is not overlooked in negotiations
  • A consultant psychiatrist should be named as the senior member of staff in the local mental health services responsible for liaison with the emergency department
  • A liaison group, with representatives from the emergency department and from mental health services should review issues of joint working between the two services. This group might double up as a self-harm services planning group (Royal College of Psychiatrists, 1994).

In keeping with this organisational diversity, the nature and quantity of provision of specialist mental health input to emergency departments, and so to the assessment and care of people who have self-harmed, also varies greatly. In some sites, psychosocial assessments are made by emergency department staff while in others they are made by mental health nurses or social workers dedicated to working with people who have self-harmed. In yet others, these are undertaken by psychiatrists and/or psychiatric nurses as part of a more general mental health liaison service working closely with the emergency department team. In places where the relationship is less well developed, the only on-site, specialist mental health input to an emergency department is a trainee psychiatrist, whose main clinical responsibilities are elsewhere, working as part of an on-call duty rota. Often the nature and extent of mental health input to an emergency department will vary from shift to shift.

These differing arrangements are likely to affect the uptake of subsequent mental healthcare, as well as the quality and consistency of the psychosocial assessments. Where dedicated teams or workers make the assessments, they will often also provide short-term, follow-up mental healthcare for people discharged from an emergency department, or work assertively to ensure that such care is provided by local community teams. This is more difficult to achieve in settings where assessments are made by junior doctors working as part of an on-call rota.

Unfortunately, the training and supervision for junior psychiatrists in assessment, referral, treatment and follow-up varies considerably (Taylor, 1998). This is important because of the high proportion, estimated to be about 60% in one study (Van Heeringen, 1992), of those referred from an emergency department for specialist mental healthcare who fail to attend the subsequent appointment. This situation is even more serious when it is remembered that about 50% of people attending the emergency department following an act of self-harm are either not offered or do not wait for a psychosocial assessment in the emergency department. A number of studies have introduced different types of interventions to improve follow-up; these are reviewed in Chapter 9.

Whatever the arrangement, the principles underpinning the management of those who self-harm are the same (Royal College of Psychiatrists, 1994; also see Section 2.11.1). Whether the detailed psychosocial assessment is done by a member of the emergency department team or by a specialist mental health worker, the person must have had sufficient training and experience and have access to specialist supervision.

The requirement that 90% (rising to 98% by 2005) of patients must not wait longer than four hours from the time of entering an emergency department to the time of departure applies equally to people who have self-harmed (Department of Health, 2001b). The Royal College of Psychiatrists and British Association for Accident and Emergency Medicine (Royal College of Psychiatrists, 2004) have proposed standards for the response times of mental health staff requested to conduct psychosocial assessments (Table 3).

Table 3. Proposed response times (from being called) for mental health staff requested to conduct a psychosocial assessment in an emergency department.

Table 3

Proposed response times (from being called) for mental health staff requested to conduct a psychosocial assessment in an emergency department.

2.12. The prevention of self-harm

It is not within the scope of this guideline to make recommendations about the primary prevention of self-harm. However, it is important to acknowledge the potential for action, at the national level, that might reduce the number of people who self-harm or the seriousness of the consequences. These factors overlap with those that influence suicide rates (Department of Health, 2002). They include:

  • The socio-economic conditions, such as poverty, unemployment and housing problems, that influence the prevalence of self-harm and problems with which it is associated, such as alcohol and substance misuse and mental illness (Gunnell et al., 1995; Hawton et al., 2001a).
  • The relationship between self-harm and child sexual and physical abuse, and the association with domestic violence, suggests that if these could be addressed in the public services that have direct contact with children in terms of detection, protection and help, the need for many women to self-harm later in life could be reduced. This is also important with regard to secondary prevention: how healthcare workers respond to people who have been abused and go on to self-harm needs to be directly addressed through training. This is one of the objectives in the Department of Health’s National Suicide Prevention Strategy for England (2002), and the Women’s Mental Health Strategy implementation guidance (2003) makes recommendations on how this can be done.
  • The availability of the means of self-harm. There is some evidence that a reduction in pack size for over-the-counter drugs has reduced the severity of the adverse consequences of overdose of aspirin and paracetamol (Hawton et al., 2001b; Hawton, 2002).
  • The development and use of prescription drugs that are safer in overdose: the advent and rapid uptake of SSRIs was at least in part the result of their reduced toxicity in overdose compared with the older tricyclic antidepressants, although a parallel reduction in the death rate from tricyclic overdose has not been forthcoming (Wilkinson & Smeeton, 1987).
  • The overall rate of prescribing of psychotropic medication may causally influence the rate of self-poisoning (Forster & Frost, 1985), suggesting that prescription of psychotropic medication should be considered a public health issue.
  • Factors that might promote self-harm as a culturally acceptable behaviour, particularly in the young. This would include the connection between self-harm and its portrayal in the media (Blood & Pirkis, 2001; Pirkis & Blood, 2001a, b; Hawton & Williams, 2001).
  • Product information may have a place in suicide prevention strategies: the National Institute for Mental Health in England (NIMHE) is currently seeking to improve the safety warnings for over-the-counter medicines as a part of the UK Suicide Prevention Strategy (Department of Health, 2002), although it has to be recognised that information on the risks of overdose and poisoning could be used to aid suicide.

2.13. Research recommendations


Research, using appropriate survey and rigorous qualitative methods, should be conducted about the meaning of self-harm to people from different ethnic and cultural groups. This should include the exploration of issues of intentionality.


Epidemiological research should be conducted to determine the prevalence of self-harm in refugees and asylum seekers.


Appropriate multi-methodology research on self-harm as a response to child sexual and other abuse should be carried out, to include a review of the substantial service user literature, and to examine the range of interventions that service users believe to be supportive and helpful, including survivor organisations and networks and other voluntary organisations for women who self-harm.


An adequately powered epidemiological study, reporting all relevant outcomes, including quality of life, occupational status and potential, income, physical well-being and quality of relationship status, should be undertaken to establish the morbidity and mortality rates for specific drug ingestions used in self-harm.

Copyright © 2004, The British Psychological Society & The Royal College of Psychiatrists.

All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Enquiries in this regard should be directed to the British Psychological Society.

Bookshelf ID: NBK56398