27 Self-harm and suicide in women

This chapter will discuss self-harm and suicide separately: their epidemiology, associated risk factors, theoretical perspectives on causes and meaning, and prevention or treatment. We will also try to show the links between them. We will draw on general research into self-harm and highlight gender differences where they have been found.

Self-harm refers to actions that cause non-fatal, physical harm to the self, regardless of suicidal intent. Conventionally, the term excludes harm resulting from drug or alcohol use or from eating disorders. Self-harm involves either self-poisoning or self-injury. Self-poisoning is roughly synonymous with taking a drug overdose or ingesting substances never intended for human consumption, and self-injury refers to any form of intentional self-inflicted damage including cutting the skin, self-immolation, swallowing objects, hanging, or jumping off buildings. In neither case does the definition depend upon lasting harm being intended.

It is contentious whether ‘attempted suicide’ should be regarded as a separate phenomenon to ‘self-harm’ (Walsh and Rosen 1988). Some authors argue that there may be a ‘self-harm syndrome’ that describes people who cut themselves regularly as a way of coping with emotions rather than as an attempt at suicide (Pao 1969). However, attaching intent to an individual’s self-harm is difficult and unreliable. Severity of injury and suicidal intent do not always match (Plutchik et al. 1989). Intent can be transient and complex, making it difficult for people to report when asked. The fact that someone who has harmed themselves has a much higher risk of completed suicide, also suggests that the two are intricately linked and to separate them makes little practical sense except in extreme cases.

Our understanding of the epidemiology of self-harm relies heavily on data on attendance at Emergency Departments and from those cases referred for assessment by psychiatric services. Another limitation is that most of what we know comes from research in North America and Western Europe. The best current United Kingdom (UK) data came from a three-centre study. The overall proportion of females attending for self-harm was 57% (Hawton et al. 2007). Analysis of first recorded episodes showed markedly higher proportions of females presenting in the under 15, 20–24, 25–29, and 35–39 age groups compared to males. Annual rates of self-harm per 100 000 (age 15 and over) in the three centres of Oxford, Manchester, and Leeds were respectively 342, 587, and 374 for females compared to 285, 460, and 291 for males.

Past literature has often emphasized self-cutting as the main method of self-harm for women yet research has consistently shown that among hospital attenders, a significantly higher proportion of women compared to men self-poison (Horrocks et al. 2002; Hawton et al. 2007).

It is also worth noting that there is no evidence that women self-injure more than men, at least among hospital attendees. More self-injury in women involves cutting, but other methods are more common in males. There is some evidence that women use different drugs in self-poisoning (Hawton et al. 2007). Repetition of self-harm is fairly equal between the genders.

We can be fairly confident that rates of self-harm are much higher (probably at least double) in the general population than we estimate from hospital attendances.

Social factors associated with repetition of self-harm include current unemployment, lower social class, alcohol or drug misuse, criminality, and debt (Hatcher 1994; House et al. 1998).

Results from the WHO/EURO study suggested that in both genders, those with a lower level of education, the unemployed and the disabled were over-represented in people who had attempted suicide when compared with controls (Schmidtke et al. 1996).

Although it is accepted that there are strong ecological associations between socioeconomic deprivation, psychiatric morbidity, and attempted suicide (Gunnell et al. 1995) it is not clear what role social deprivation has as an independent determinant of self-harm.

One recent UK study found that 92% of 150 people who presented to a district general hospital in Oxford after self-harm had a diagnosable psychiatric disorder at that time (Haw et al. 2001). Affective disorder was the commonest (72%). Women were more likely to have an eating disorder and less likely to be dependent on alcohol than men. However, an earlier study found that although at the time of their self-harm episode 69% of people were measured as having a diagnosable mood disorder, this dropped to 39% a week later (Newson-Smith and Hirsch 1979).

This finding, and the observation of dramatic changes in rates of self-harm over time, suggests that rises in the number of psychiatric diagnoses found in people who self-harm could be due to changes that have occurred in methods of identifying psychiatric disorders and ways of thinking about mental health. Thus, self-harm is less a symptom of pervasive mood disorders and more a response to social circumstances, accompanied by transient states of distress.

The Oxford study also carried out a follow-up interview of their original sample. They diagnosed a staggering 46% as having a personality disorder and only two of these did not have a co-occurring psychiatric disorder (Haw et al. 2001).

It is not only such high rates that raise questions; the very status of personality disorders is a contentious one. For example, Livesley points out that the personality disorders identified by DSM are not clinically useful, since patients usually don’t map directly onto one disorder; they are neither mutually exclusive nor exhaustive in their identification of characteristics and traits; they have poor psychometric properties; they draw from a wide variety of theories and historical perspectives which means they lack reliability and validity and behaviours identified within the personality disorders merge with one another and with normality (Livesley 2003).

A comment regarding psychological attributions of self-harm and gender stereotyping: personality disorder as a diagnosis tends to serve as a catch-all for the types of behaviour that society finds unacceptable, so that individuals are pathologized by societal norms. Poor quality studies have led to spurious clusters of people diagnosed with personality disorder, which in turn leads to stereotyping of individuals.

The much higher rates of self-harm in women compared to men that were evident in the 1960s meant that there was a drive to understand ‘these women’ particularly those that self-harmed more than once (Grunebaum and Klerman 1967; Favazza and Conterio 1989). For example, Graff and Mallin (1967) described the ‘typical wrist cutter’ using a sample of only 21 patients:

In summary, the cutter is an attractive, intelligent, unmarried young woman, who is either promiscuous or overly afraid of sex, easily addicted, and unable to relate successfully to others.

The stereotype of the young woman who repeatedly cuts herself, who has suffered childhood abuse, and who is likely to be diagnosed with borderline personality disorder is a common one despite evidence to refute its accuracy. For example a number of studies have shown that relatively equal proportions of men and women cut themselves (Clendenin and Murphy 1971; Horrocks et al. 2003; Lilley et al. 2008).

Two aspects of self-harm are often neglected as a result of this preoccupation with psychiatric diagnosis. These are: the social context within which the self-harm occurs and self-harm as an adaptive and thus an understandable response to a range of individual situations and histories.

Although it is not always clear to the individual or to clinicians how an episode of self-harm has occurred, Fig. 27.1 attempts to describe the pathway to self-harm.

Early separation from either parent has been shown to lead to a fourfold increase in risk for further suicidal acts in men, but not for women (Oquendo et al. 2007). On the other hand women who perceive themselves as rejected or neglected by either of their parents have been found to be more likely to have made at least one suicide attempt, compared to females without this perception and all men (Ehnvall et al. 2008).

One of the more frequent stereotypes of women who self-harm, especially those who do so frequently, is that they have experienced abuse and the presumption is often that this is sexual abuse, though physical and psychological abuse may be as damaging. A community sample in Australia found that boys were more likely to attempt suicide after being sexually abused than girls. After adjustment for depression, hopelessness, and family functioning, the experience of sexual abuse no longer had a significant effect on suicidal behaviour in girls, whilst with boys the increased risk for self-harm and suicide remained (Martin et al. 2004).

In another study of antecedents of self-destructive behaviour in 18–34 year olds, histories of physical and sexual abuse, parental neglect, and separation were associated with a range of self-destructive behaviours in adulthood, including suicide attempts and self-cutting (van der Kolk et al. 1991). At follow-up, experience of neglect became the most powerful predictor of self-destructive behaviour, leading the authors to suggest that although trauma was a factor in initiating self-destructive behaviour, experience of neglect appeared to be a factor in maintaining it.

Many acts of self-harm have been in response to some immediate emotional or social difficulty that has occurred in the few days immediately before the episode. Examples of such factors include difficulties with partners, parents, children, friends, financial difficulties, anniversaries, bereavement, or other losses (Hawton et al. 2003).

These more immediate provocations are most likely to have their effect by interaction with longer-term vulnerabilities—perhaps especially if they mirror or reawaken feelings or dilemmas from those earlier experiences.

Self-harm is often seen as an abnormal behaviour, but it may have an adaptive function.

Early attempts to understand the function of self-harm have derived from accounts from individuals, typically women, who cut themselves and who do so on a frequent basis. As mentioned earlier, this has led to debate about the existence of a separate self-cutting syndrome (Graff and Mallin 1967; Pao 1969; Walsh and Rosen 1988).

Young women’s accounts of their self-injury describe it as an expression of emotional distress, such as sadness or anger (Abrams and Gordon 2003); the act of self-injury either communicates this to others or releases emotion and brings calm. Other young women have spoken about the need to turn emotional pain into physical pain, which is easier to deal with (Solomon and Farrand 1996).

The weakness of these studies on subjective experience of self-harm has been that they have been over-generalized to explain other forms of self-harm, which may or may not be repetitive. There is, for example, a strong link between self-cutting and other suicidal behaviour which shows that those who self-injure often also self-poison (Horrocks et al. 2003), and there are also people who repetitively self-poison, without specific suicidal intent.

In this chapter we have distinguished between self-reported functions and theoretical accounts of self-harm. Functions relate to the subjective experience and accounts of individuals who self-harm; theory is generated from interpretation, the research evidence and rhetoric.

Suyemoto (1988) has summarized many of the theoretical models associated with ‘self-mutilation’ and we will not mention all of them here. Although she focuses on explaining repetitive self-injury rather than all self-harm and we would disagree with the terminology, some of the models could also be used to explain other self-harming behaviour.

This model focuses on the two-way interaction between the individual who self-harms and the environment. People who self-harm experience secondary gains from their act, such as care and attention to their needs and this may reinforce or maintain the behaviour. However, the act of self-harm also serves the environment, so, for example, the self-harm act can maintain the status quo of family roles and relationships in a quasi-safe manner. This interaction between environment and individual acts is usually an unconscious one, and in most theories of family function it is seen as especially pertinent to women.

A psychodynamic approach sees self-harm as a secondary deviant behaviour that enables women to cope (Maris 1971). Early childhood experiences of neglect, abandonment or abuse lead to low self-esteem. Self-harm allows a woman to partially self-destruct without the finality of suicide, and without dealing with unresolved issues from her childhood. This partial self-destruction actually enables the woman to maintain a veneer of recovery and carry on living. This model sees self-harm as an active coping mechanism used to avoid suicide, rather than as a suicide attempt.

This model relates specifically to women, is based on classical Freudian theory, and focuses on sexual urges and conflicts that arise from repressing them. The theory proposes that women reject the female sexual organs as inferior to the penis because they are internal and therefore not as accessible. This leads to penis envy and a feeling of inferiority. Women become passive and reliant on male penetration for pleasure and internalize their aggression. This aggression may manifest itself in a number of ways, for example, through promiscuity or self-harm.

As with other psychodynamic theories this is difficult to prove or disprove. However, the idea that women are passive sexual beings seems outdated and misogynistic. Although the idea of internalized aggression makes sense, the reasons given for this do not fit well with 21st century sexual politics.

This model describes difficulties with tolerating the inner experience of emotions, expressing that emotion, and maintaining emotional equilibrium. Self-harm may be a way of communicating to others feelings of emotional distress. It can also be a way of communicating this to the self, since the act and its consequences can bring emotional distress to the attention of the conscious self. The self-harm either externalizes and releases the emotional pain (for example, through self-cutting) or deadens the emotion, providing escape through self-poisoning.

The dissociation model can be linked with the affect regulation model. Some people who self-harm report feelings of dissociation from their environment; a feeling of separateness or a lack of self. The function of self-harm is often to end that dissociation. Suyemoto (1988) cites other work that suggests that self-harm can also function as a means to become dissociated, so as to escape overwhelming emotion.

This model is based on object-relations theory and has been used to explain self-cutting. Rejection by or loss of someone close, perhaps through relationship breakdown, elicits overwhelming feelings of loss in the person who self-harms. The feelings are so intense that the sufferer feels they are losing themselves. Acts of self-injury reaffirm a sense of self, since the skin represents separateness from others and the scar that is left is a reminder of this distinction.

Cognitive theories have concentrated on styles of thinking that lead to an increased risk of self-harm. Cognitive studies have found that people who self-harm have more passive problem-solving styles than others with solutions being less versatile and less relevant to the problem (Orbach et al. 1990). Thus when confronted with a problem, that perhaps has high emotional content, the individual cannot think of a solution other than to harm themselves.

Poor problem-solving leads to hopelessness and/or helplessness, which increase the risk of self-harm (Schotte and Clum 1982). Hopelessness and poor problem-solving ability may, however, act independently of each other to increase risk (Mcleavey et al. 1987). There is no evidence that any of these cognitive styles is gender-specific.

Taylor (cited in Jack 1992) complains that theory is fragmented because research has focused on epidemiology, psychiatric diagnosis, and risk. Characteristics or behaviours are dealt with in isolation, rather than drawing on the wider field of suicidal behaviour and thus there is no cohesive theoretical drive to understand suicidal behaviour in all its forms.

One of the challenges faced by services wishing to stop people from self-harming is whether it is ethical to do so. If the function of self-harm is to avoid suicide, or to manage overwhelming emotions, then to try and stop it may have more dire consequences, and intervention is likely to be met with resistance and failure because the cause persists. It is therefore essential that treatment programmes attempt to unpick the function of self-harm for the individual or at the very least to provide alternative strategies for coping with emotions or situations that might otherwise lead to self-harm. Another challenge is that self-harm is not in itself a diagnosis; it is a behaviour engaged in by people with a range of diagnosable and non-diagnosable mental health problems.

A review of psychosocial and pharmacological treatments for self-harm (Hawton et al. 1998) identified 23 trials. Trials included the following forms of treatment or management: problem-solving therapy, intensive care/outreach, provision of emergency card, dialectical behaviour therapy, inpatient behaviour therapy, continuity of care, hospital admission, drug treatments, and long-term therapy. For most of these interventions, the majority of recipients were women and there is a consensus that women are more easily engaged in treatment after self-harm. The authors noted that both because of small size and the heterogeneity of the samples in the trials, results were unclear. No strong evidence was found for any of the interventions although the evidence was suggestive of benefit for brief problem-oriented psychological interventions. Dialectical behaviour therapy, which was used only with women who had self-harmed more than once, did show some reduction in self-harming behaviour during follow-up.

Goldney’s review (1998) notes that interventions do not have a sustained effect. Since suicidal behaviours are often the result of long-term problems, it is necessary to think of long-term solutions, rather than short-term interventions.

Epidemiological studies tell us that women self-harm more than men, though in the UK the ratio of women to men has become more equal over the last few decades. The evidence reported here suggests that there are some gender differences in the risk factors for self-harm. Physical and sexual abuse experienced by women in childhood may not lead to suicidal behaviours in the same way that they do for men, whereas neglectful parenting may be more of a risk factor for later self-harm or self-destructive behaviours. We know that anger is a feature of some self-harm in women, but can only hypothesize that the gender difference in rates of self-harm may be due to lack of opportunity and social norms that for women limit the expression of aggression externally, leading to internalized aggression and attacks on the self.

We do not know enough about the functions of self-harm in men or of a range of self-harm in both sexes, to understand where gender differences may lie. The following section on suicide may highlight further areas of difference or similarity that may help with explanations of the gender gap.

People who self-harm are at a much greater risk of suicide than any other group of people. However, while there are higher rates of self-harm in women, there are higher rates of suicide in men.

Age-standardized rates of suicide for the UK in 2006 show that the rate of suicide for females is 5.3 (per 100 000) and 17.4 for males (ages 15 and over).

In comparison with suicide data from Northern Europe, UK figures are the lowest overall and below the average for the whole of Europe.

There has been a decrease in suicides during both World War I and II; an increase in the 1920s and 1930s when there was severe economic recession, and a further decline in the mid 1960s and early 1970s (Gunnell 2005). The latter may be due to the change in domestic gas to become non-toxic, since the main method of suicide during the early 1960s was by domestic gas poisoning.

In the 1980s there appeared to be an increase in rates for widowed and divorced men and single men but by comparison there was a decrease in rates for widowed and divorced women and single women and the rates for married women remain relatively stable over time (Charlton et al. 1993).

A more recent study of suicide in young people between the ages of 15–34 years looked at rates between 1968–2005. By 2005, rates were at their lowest for almost 30 years, for women and men. Worryingly deaths by hanging have increased in young women, and death by overdose has reduced, which may indicate that young women are beginning to choose more lethal and violent methods of suicide (Biddle et al. 2008).

Younger suicides have been characterized by chronic interpersonal difficulties and recent interpersonal problems as well as acute and severe mental disorder (Appleby et al. 1999). After controlling for psychiatric history, a number of risk factors remain for men: being single, unemployed, retired, and having higher levels of sickness absence. For women there were no other significant risk factors although one study found a significantly reduced risk of suicide for women who had a child under two years old (Qin et al. 2000). This latter finding is consistent with findings that found pregnancy to be a time of reduced risk (Appleby 1991).

Although we briefly mentioned the link between self-harm and aggression earlier in the chapter, it is worth mentioning again. There is biological evidence on the relationship between serotonin function, aggression, and suicidal behaviour (Lidberg et al. 2000). In addition, research investigating personality traits has also found low socialization and high impulsive aggression scores in violent offenders (Gunilla Stalenheim 2001). This sheds an interesting light on the male predominance in suicide figures.

Research on individual risk factors has failed to explain why some people in high-risk groups take their own lives and others do not. Durkheim was interested in the interplay between the individual and the wider community; that this should be a mutually beneficial relationship, maintaining a health balance of individuality and communality. He identified four states that would upset this equilibrium between the individual and the community, leading to a higher probability of suicide within the population of that community. These four states were Altruism, Fatalism, Egoism, and Anomie.

Other sociological explanations also centre around themes of anomie, social integration and isolation, and industrialization and changes in the labour force. Stack’s paper (1978) cites the work of Gibbs and Martin (1964) who suggested that those who are divorced, retired, or have other changes in role, which either do not match with societal expectations or their own, experience role strain. They hypothesized that populations with large proportions of people experiencing this role conflict will have a relatively high rate of suicide. Other work has suggested that female labour force participation may influence suicide rates across countries but there is little evidence to support this (Platt and Hawton 2000).

Lester (1989) focuses on the presence of depression in those that complete suicide. His theory postulates that events in the lives of depressed individuals have taught them that they cannot influence the outcomes of their own lives (taken from the learned helplessness model of Seligman, 1975). Conversely these same depressed individuals believe they are responsible for their failures (drawn from Beck’s work, 1967). This ‘depression paradox theory’ leads to internal conflict between feelings of helplessness and responsibility. Eventually the individual may feel that the only way out of their situation is suicide (Lester 1989).

The World Health Organization strategy for suicide prevention (WHO 2007) sets out a number of objectives. These include:

Such steps seem to be logical, yet there is little evidence to support them (Goldney 1998). Some work has already been done on reducing the means to suicide. Changes to the toxicity of the domestic gas supply in the late 1950s and early 1960s coincided with a marked reduction in overall suicide rates. However, there was an increase in the rates of death by overdose from the late 1950s, though this levelled off by the mid 1960s. This could have been partly due to method substitution or the fact that prescribed drugs became both more toxic and more readily available (Gunnell et al. 2000). In England and Wales the introduction of catalytic converters in cars had some effect on the overall population suicide rate (all methods). However, there was no overall change in the suicide rates of young men and women, for whom the rates of hanging increased (Amos et al. 2001). Reduction in pack sizes of paracetamol has led to a reduction in deaths by paracetamol overdose (Hawton et al. 2004), though data on whether overall suicide rates have simultaneously decreased was not reported.

There is no evidence to guide a suicide prevention policy specifically at women.

The work on suicide has consistently shown that in most European countries, and certainly in the UK, suicide rates are higher for men than for women. The usual argument to explain this is that men have access to and choose more violent and therefore lethal means than women. It is unlikely that this is the sole cause for the difference in suicide rates, especially as there is a rise in women hanging themselves. Changes in the last century may shed some light on gender differences. For unmarried women rates of suicide have decreased and this may reflect the change in social norms, where marriage is no longer the overwhelmingly favourable option for women. In contrast the higher rates of suicide in widowed, divorced, and single men may point towards isolation and lack of social support as a key factor in suicide rates. Women are stereotypically more inclined to form supportive social networks than men, whose friendships may be less emotionally supportive. There is little evidence available to support or discount this theory.

It may be that there are more similarities between women and men than there are differences, but research on self-harm and suicide has evolved in such a way that we are unable to determine what gender differences exist. This is partially due to the focus on psychiatric diagnosis, which may be misleading, particularly in relation to personality disorders. We know that women are more likely to receive a diagnosis of borderline personality disorder, but given the lack of reliability of such diagnoses, this is of little help. The fact that subjective experiences, theory generation, and intervention research seem to be studied separately means that often unjustified assumptions are made and little headway is gained in the field of self-harm research as a whole. Given that different methods of self-harm are used by both men and women and that suicide methods may no longer be so gender-specific, research needs to be more encompassing and also specifically look for gender differences, in case gender-specific interventions are needed.