Of all of the following which is the most accurate statement regarding personality disorders?

The term ‘borderline personality’ was proposed in the United States by Adolph Stern in 1938 (most other personality disorders were first described in Europe). Stern described a group of patients who ‘fit frankly neither into the psychotic nor into the psychoneurotic group’ and introduced the term ‘borderline’ to describe what he observed because it ‘bordered’ on other conditions.

The term ‘borderline personality organisation’ was introduced by Otto Kernberg (1975) to refer to a consistent pattern of functioning and behaviour characterised by instability and reflecting a disturbed psychological self-organisation. Whatever the purported underlying psychological structures, the cluster of symptoms and behaviour associated with borderline personality were becoming more widely recognised, and included striking fluctuations from periods of confidence to times of absolute despair, markedly unstable self-image, rapid changes in mood, with fears of abandonment and rejection, and a strong tendency towards suicidal thinking and self-harm. Transient psychotic symptoms, including brief delusions and hallucinations, may also be present. The characteristics that now define borderline personality disorder were described by Gunderson and Kolb in 1978 and have since been incorporated into contemporary psychiatric classifications (see Section 2.2).

Either as a result of its position on the ‘border’ of other conditions, or as a result of conceptual confusion, borderline personality disorder is often diagnostically comorbid with depression and anxiety, eating disorders such as bulimia, post-traumatic stress disorder (PTSD), substance misuse disorders and bipolar disorder (with which it is also sometimes clinically confused). An overlap with psychotic disorders can also be considerable. In extreme cases people can experience both visual and auditory hallucinations and clear delusions, but these are usually brief and linked to times of extreme emotional instability, and thereby can be distinguished from the core symptoms of schizophrenia and other related disorders (Links et al., 1989).

The level of comorbidity is so great that it is uncommon to see an individual with ‘pure’ borderline personality disorder (Fyer et al., 1988a). Because of this considerable overlap with other disorders, many have suggested that borderline personality disorder should not be classified as a personality disorder; rather it should be classified with the mood disorders or with disorders of identity. Its association with past trauma and the manifest similarities with PTSD have led some to suggest that borderline personality disorder should be regarded as a form of delayed PTSD (Yen & Shea, 2001). Despite these concerns, borderline personality disorder is a more uniform category than other personality disorders and is probably the most widely researched of the personality disorders. While some people with borderline personality disorder come from stable and caring families, deprivation and instability in relationships are likely to promote borderline personality development and should be the focus of preventive strategies.

It is important to note that borderline personality disorder should not be confused with so-called ‘borderline intelligence’ which is a wholly distinct and unrelated concept. Nevertheless, borderline personality characteristics (notably self-harm) are sometimes present in people with significant learning disabilities and can be prominent (Alexander & Cooray, 2003).

The course of borderline personality disorder is very variable. Most people show symptoms in late adolescence or early adult life, although some may not come to the attention of psychiatric services until much later. The outcome, at least in those who have received treatment or formal psychiatric assessment, is much better than was originally thought, with at least 50% of people improving sufficiently to not meet the criteria for borderline personality disorder 5 to 10 years after first diagnosis (Zanarini et al., 2003). It is not known to what extent this is a consequence of treatment – evidence suggests that a significant proportion of improvement is spontaneous and accompanied by greater maturity and self-reflection.

There is some controversy over the possible age of onset of borderline personality disorder. Many believe that it cannot, or perhaps should not, be diagnosed in people under 18 years of age while the personality is still forming (although diagnosis is possible in the Diagnostic and Statistical Manual of Mental Disorders, 4th edition [DSM-IV; APA, 1994] based on the same criteria as adults with additional caveats). Nevertheless, borderline symptoms and characteristics are often identifiable at a much earlier age, and sometimes early in adolescence (Bradley et al., 2005a). More attention is now being paid to its early manifestations in adolescent groups (see Section 2.7).

Borderline personality disorder is associated with significant impairment, especially in relation to the capacity to sustain stable relationships as a result of personal and emotional instability. For many the severity of symptoms and behaviours that characterise borderline personality disorder correlate with the severity of personal, social and occupational impairments. However, this is not always the case, and some people with what appears to be, in other ways, marked borderline personality disorder may be able to function at very high levels in their careers (Stone, 1993). Many, but not all, people with borderline personality disorder recurrently harm themselves, usually to provide relief from intolerable distress, which for many can lead to significant physical impairment and disability. Moreover, suicide is still common in people with borderline personality disorder and may occur several years after the first presentation of symptoms (Paris & Zweig-Frank, 2001).

Although the prognosis of borderline personality disorder is relatively good, with most people not meeting the criteria for diagnosis after 5 years, it is important to note that a minority of people have persistent symptoms until late in life. Recurrent self-harm may occasionally be a problem in the elderly and the possibility that this may be because of borderline personality disorder should be considered in such circumstances. However, the prevalence of the condition in the elderly is much lower than in the young and one of the encouraging features about remission from the condition is that it is much less often followed by relapse than is the case with most other psychiatric disorders.

Borderline personality disorder is a heterogeneous condition and its symptoms overlap considerably with depressive, schizophrenic, impulsive, dissociative and identity disorders. This overlap is also linked to comorbidity and in clinical practice it is sometimes difficult to determine if the presenting symptoms are those of borderline personality disorder or a related comorbid condition. The main differences between the core symptoms of borderline personality disorder and other conditions are that the symptoms of borderline personality disorder undergo greater fluctuation and variability: psychotic and paranoid symptoms are transient, depressive symptoms change dramatically over a short period, suicidal ideas may be intense and unbearable but only for a short time, doubts about identity may occur but are short-lived, and disturbances in the continuity of self-experiences are unstable. For each of the equivalent comorbid disorders there is much greater consistency of these symptoms.

Borderline personality disorder is one of the most contentious of all the personality disorder subtypes. The reliability and validity of the diagnostic criteria have been criticised, and the utility of the construct itself has been called into question (Tyrer, 1999). Moreover, it is unclear how satisfactorily clinical or research diagnoses actually capture the experiences of people identified as personality disordered (Ramon et al., 2001). There is a large literature showing that borderline personality disorder overlaps considerably with other categories of personality disorder, with ‘pure’ borderline personality disorder only occurring in 3 to 10% of cases (Pfohl et al., 1986). The extent of overlap in research studies is particularly great with other so-called cluster B personality disorders (histrionic, narcissistic and antisocial). In addition, there is considerable overlap between borderline personality disorder and mood and anxiety disorders (Tyrer et al., 1997; Zanarini et al., 1998).

This guideline uses the DSM-IV diagnostic criteria for borderline personality disorder (APA, 1994), which are listed in Table 1. According to DSM-IV, the key features of borderline personality disorder are instability of interpersonal relationships, self-image and affect, combined with marked impulsivity beginning in early adulthood.

A stand-alone category of borderline personality disorder does not exist within the International Classification of Diseases, 10th revision (ICD-10; World Health Organization, 1992), although there is an equivalent category of disorder termed ‘emotionally unstable personality disorder, borderline type’ (F 60.31), which is characterised by instability in emotions, self-image and relationships. The ICD-10 category does not include brief quasi-psychotic features (criterion 9 of the DSM-IV category). Comparisons of DSM and ICD criteria when applied to the same group of patients have shown that there is little agreement between the two systems. For example, in a study of 52 outpatients diagnosed using both systems, less than a third of participants received the same primary personality disorder diagnosis (Zimmerman, 1994). Further modifications in the ICD and DSM are required to promote convergence between the two classifications, although greater convergence is unlikely to resolve the problems inherent in the current concept of personality disorder.

The reliability of diagnostic assessment for personality disorder has been considerably improved by the introduction of standardised interview schedules. However, no single schedule has emerged as the ‘gold standard’ as each has its own set of advantages and disadvantages, with excessive length of interview time being a problem common to many of the schedules. (The main instruments available for assessing borderline personality disorder are listed in Table 2.) When used by a properly trained rater, all of the schedules allow for a reliable diagnosis of borderline personality disorder to be made. Nevertheless, the level of agreement between interview schedules remains at best moderate (Zimmerman, 1994). In addition, clinical and research methods for diagnosing personality disorders diverge. Westen (1997) has found that although current instruments primarily rely on direct questions derived from DSM-IV, clinicians tend to find direct questions only marginally useful when assessing for the presence of personality disorders. Instead, clinicians are inclined to arrive at the diagnosis of personality disorder by listening to patients describe interpersonal interactions and observing their behaviour (Westen, 1997).

Currently, outside specialist treatment settings, there is still a heavy reliance on the diagnosis of borderline personality disorder being made following an unstructured clinical assessment. However, there are potential pitfalls in this approach. First, agreement among clinicians’ diagnoses of personality disorder has been shown to be poor (Mellsop et al., 1982). Second, the presence of acute mental or physical illness can influence the assessment of personality. The presence of affective and anxiety disorders, psychosis, or substance use disorder, or the occurrence of an acute medical or surgical condition can all mimic symptoms of borderline personality disorder; a primary diagnosis of borderline personality disorder should only be made in the absence of mental or physical illness. It is also preferable for clinicians to obtain an informant account of the individual’s personality before definitively arriving at a diagnosis of borderline personality disorder.

All personality disorders have been defined by their stability over time. Indeed, ICD and DSM definitions of personality disorders describe them as having an enduring pattern of characteristics. However, until recently, there was a paucity of longitudinal research into personality disorders to support the notion of borderline personality disorder as a stable construct. Reviews of the subject published over the past 10 years hinted at considerable variation in stability estimates (Grilo et al., 2000). Recent prospective studies have shown that a significant number of individuals initially diagnosed with borderline personality disorder will not consistently remain at diagnostic threshold, even over comparatively short periods of time (Shea et al., 2002). It seems that while individual differences in personality disorder features appear to be relatively stable (Lenzenweger, 1999), the number of criteria present can fluctuate considerably over time. Given the many problems associated with the diagnosis of borderline personality disorder, it seems clear that reclassification is urgently needed and this is likely to happen with the publication of DSM-V (Tyrer, 1999).

Although borderline personality disorder is a condition that is thought to occur globally (Pinto et al., 2000), there has been little epidemiological research into the disorder outside the Western world. Only three methodologically rigorous surveys have examined the community prevalence of borderline personality disorder. Coid and colleagues (2006) reported that the weighted prevalence of borderline personality disorder in a random sample of 626 British householders was 0.7%. Samuels and colleagues (2002) found that in a random sample of 742 American householders the weighted prevalence of borderline personality disorder was 0.5%. Torgersen and colleagues (2001) reported a prevalence of 0.7% in a Norwegian survey of 2,053 community residents. Despite methodological differences between these studies, there is remarkable concordance in their prevalence estimates, the median prevalence of borderline personality disorder across the three studies being 0.7%. Only Torgersen and colleagues’ 2001 study provides detailed information about the sociodemographic correlates of borderline personality disorder. In this study, there was a significant link between borderline personality disorder and younger age, living in a city centre and not living with a partner. Interestingly, the assumption that borderline personality disorder is over-represented among women was not supported by the data.

In primary care, the prevalence of borderline personality disorder ranges from 4 to 6% of primary attenders (Moran et al., 2000; Gross et al., 2002). Compared with those without personality disorder, people with borderline personality disorder are more likely to visit their GP frequently and to report psychosocial impairment. In spite of this, borderline personality disorder appears to be under-recognised by GPs (Moran et al., 2001).

In mental healthcare settings, the prevalence of all personality disorder subtypes is high, with many studies reporting a figure in excess of 50% of the sampled population. Borderline personality disorder is generally the most prevalent category of personality disorder in non-forensic mental healthcare settings. In community samples the prevalence of the disorder is roughly equal male to female, whereas in services there is a clear preponderance of women, who are more likely to seek treatment. It follows that the majority of people diagnosed with personality disorder, most of whom will have borderline personality disorder, will be women.

Borderline personality disorder is particularly common among people who are drug and/or alcohol dependent, and within drug and alcohol services there will be more men with a diagnosis of borderline personality disorder than women. Borderline personality disorder is also more common in those with an eating disorder (Zanarini et al., 1998), and also among people presenting with chronic self-harming behaviour (Linehan et al., 1991).

Many people who have at one time been given the diagnosis of borderline personality disorder are able to move on to live a fulfilling life. However, during the course of the disorder people can have significant problems which mean that they require a large amount of support from services and from those around them. The functional impairment associated with borderline personality disorder appears to be a relatively enduring feature of the disorder (Skodol et al., 2005). Studies of clinical populations have shown that people with borderline personality disorder experience significantly greater impairment in their work, social relationships and leisure compared with those with depression (Skodol et al., 2002). However, studies of selected samples of people with borderline personality disorder have shown that symptomatic improvement can occur to the extent that a number of people will no longer meet the criteria for borderline personality disorder and that the prognosis may be better than has previously been recognised (Zanarini et al., 2003).

People with borderline personality disorder may engage in a variety of destructive and impulsive behaviours including self-harm, eating problems and excessive use of alcohol and illicit substances. Self-harming behaviour in borderline personality disorder is associated with a variety of different meanings for the person, including relief from acute distress and feelings, such as emptiness and anger, and to reconnect with feelings after a period of dissociation. As a result of the frequency with which they self-harm, people with borderline personality disorder are at increased risk of suicide (Cheng et al., 1997), with 60 to 70% attempting suicide at some point in their life (Oldham, 2006). The rate of completed suicide in people with borderline personality disorder has been estimated to be approximately 10% (Oldham, 2006). A well-documented association exists between borderline personality disorder and depression (Skodol et al., 1999; Zanarini et al., 1998), and the combination of the two conditions has been shown to increase the number and seriousness of suicide attempts (Soloff et al., 2000).

The causes of borderline personality disorder are complex and remain uncertain. No current model has been advanced that is able to integrate all of the available evidence. The following may all be contributing factors: genetics and constitutional vulnerabilities; neurophysiological and neurobiological dysfunctions of emotional regulation and stress; psychosocial histories of childhood maltreatment and abuse; and disorganisation of aspects of the affiliative behavioural system, most particularly the attachment system.

Twin studies suggest that the heritability factor for borderline personality disorder is 0.69 (Torgersen et al., 2000), but it is likely that traits related to impulsive aggression and mood dysregulation, rather than borderline personality disorder itself, are transmitted in families. Current evidence suggests that the genetic influence on personality disorder generally, not specifically borderline personality disorder, acts both individually and in combination with anomalous environmental factors (White et al., 2003; Caspi et al., 2002; Caspi et al., 2003). More recent studies of heritability suggest that the heritability factor for cluster C disorders lies within the range 27 to 35% (Reichborn-Kjennerud et al., 2007) suggesting that genetic factors play a less important role than previously thought.

Regulation of emotional states is a core problem in borderline personality disorder. Neurotransmitters have been implicated in the regulation of impulses, aggression and affect. Serotonin has been the most extensively studied of these, and it has been shown that there is an inverse relationship between serotonin levels and levels of aggression. Reduced serotonergic activity may inhibit a person’s ability to modulate or control destructive urges, although the causal pathway remains unclear. Reduced 5-HT 1A receptor-mediated responses in women with borderline personality disorder and a history of prolonged child abuse have been noted (Rinne et al., 2000), suggesting the possibility that environmental factors might mediate the link between 5-HT and aggression.

Limited evidence exists for the role of catecholamines (norepinephrine and dopamine neurotransmitters) in the dysregulation of affect. People with borderline personality disorder have lower plasma-free methoxyhydroxyphenylglycol (a metabolite of noradrenaline), compared with controls without borderline personality disorder, but the finding disappears when aggression scores are controlled (Coccaro et al., 2003). The effects produced on administering amphetamines to people with borderline personality disorder suggest that such people are uniquely sensitive and demonstrate greater behavioural sensitivity than control subjects (Schulz et al., 1985).

Other neurotransmitters and neuromodulators implicated in the phenomenology of borderline personality disorder include acetylcholine (Steinberg et al., 1997), vasopressin (Coccaro et al., 1998), cholesterol (Atmaca et al., 2002) and fatty acids (Zanarini & Frankenburg, 2003), along with the hypothalamic-pituitary adrenal axis (Rinne et al., 2002).

Evidence of structural and functional deficit in brain areas central to affect regulation, attention and self-control, and executive function have been described in borderline personality disorder. Areas include the amygdala (Rusch et al., 2003), hippocampus (Tebartz van Elst et al., 2003) and orbitofrontal regions (Stein et al., 1993; Kunert et al., 2003; De la Fuente et al., 1997). Most studies are performed without emotional stimulation, however recent studies under conditions of emotional challenge suggest similar findings. People with borderline personality disorder show increased activity in the dorsolateral prefrontal cortex and in the cuneus, and a reduction in activity in the right anterior cingulate (Schmahl et al., 2003). Greater activation of the amygdale while viewing emotionally aversive images (Herpertz et al., 2001) or emotional faces (Donegan et al., 2003) has also been described.

Family studies have identified a number of factors that may be important in the development of borderline personality disorder, for example a history of mood disorders and substance misuse in other family members. Recent evidence also suggests that neglect, including supervision neglect, and emotional under-involvement by caregivers are important. Prospective studies in children have shown that parental emotional under-involvement contributes to a child’s difficulties in socialising and perhaps to a risk for suicide attempts (Johnson et al., 2002). People with borderline personality disorder (at least while symptomatic), significantly more often than people without the disorder, see their mother as distant or overprotective, and their relationship with her conflictual, while the father is perceived as less involved and more distant. This suggests that problems with both parents are more likely to be the common pathogenic influence in this group rather than problems with either parent alone. While these findings should be replicated with those who have recovered from borderline personality disorder, the general point about biparental difficulties being important in the genesis of borderline personality disorder is given further support from studies of abuse.

Physical, sexual and emotional abuse can all occur in a family context and high rates are reported in people with borderline personality disorder (Johnson et al., 1999a). Zanarini reported that 84% of people with borderline personality disorder retrospectively described experience of biparental neglect and emotional abuse before the age of 18, with emotional denial of their experiences by their caregivers as a predictor of borderline personality disorder (Zanarini et al., 2000). This suggests that these parents were unable to take the experience of the child into account in the context of family interactions. Abuse alone is neither necessary nor sufficient for the development of borderline personality disorder and predisposing factors and contextual features of the parent-child relationship are likely to be mediating factors in its development. Caregiver response to the abuse may be more important than the abuse itself in long-term outcomes (Horwitz et al., 2001). A family environment that discourages coherent discourse about a child’s perspective on the world is unlikely to facilitate successful adjustment following trauma. Thus the critical factor is the family environment. Studies that have examined the family context of childhood trauma in borderline personality disorder tend to see the unstable, non-nurturing family environment as the key social mediator of abuse (Bradley et al., 2005b) and personality dysfunction (Zweig-Frank & Paris, 1991).

Few of the studies point to how the features of parenting and family environment create a vulnerability for borderline personality disorder, but they are likely to be part of a disrupted attachment or affiliative system that affects the development of social cognition, which is considered to be impaired in borderline personality disorder (Fonagy & Bateman, 2007).

The literature on the relationship between attachment processes and the emergence of borderline personality disorder is broad and varies. For example, some studies suggest that people are made more vulnerable to the highly stressful psychosocial experiences discussed above by early inadequate mirroring and disorganised attachment. This is likely to be associated with a more general failure in families such as neglect, rejection, excessive control, unsupportive relationships, incoherence and confusion. While the relationship of diagnosis of borderline personality disorder and specific attachment category is not obvious, borderline personality disorder is strongly associated with insecure attachment (6 to 8% of patients with borderline personality disorder are coded as secure) and there are indications of disorganisation (unresolved attachment and inability to classify category of attachment) in interviews, and fearful avoidant and preoccupied attachment in questionnaire studies (Levy, 2005). Early attachment insecurity is a relatively stable characteristic of any individual, particularly in conjunction with subsequent negative life events (94%) (Hamilton, 2000; Waters et al., 2000; Weinfield et al., 2000). Given evidence of the continuity of attachment from early childhood, at least in adverse environments, and the two longitudinal studies following children from infancy to early adulthood (which reported associations between insecure attachment in early adulthood and borderline personality disorder symptoms [Lyons-Ruth et al., 2005]), childhood attachment may indeed be an important factor in the development of borderline personality disorder. Fonagy and colleagues (2003) suggest that adverse effects arising from insecure and/or disorganised attachment relationships, which may have been disrupted for many reasons, are mediated via a failure in development of mentalising capacity – a social cognitive capacity relating to understanding and interpreting one’s own and others’ actions as meaningful on the basis of formulating what is going on in one’s own and the other person’s mind.

This formulation overlaps with the importance of the invalidating family environment suggested by Linehan (1993) as a factor in the genesis of borderline personality disorder and further developed by Fruzzetti and colleagues (2003; 2005). Fruzzetti and colleagues report that parental invalidation, in part defined as the undermining of self-perceptions of internal states and therefore anti-mentalising, is not only associated with the young person’s reports of family distress, and their own distress and psychological problems, but also with aspects of social cognition, namely the ability to identify and label emotion in themselves and others. Along with other aspects contributing to the complex interaction described as invalidating, there is a systematic undermining of a person’s experience of their own mind by that of another. There is a failure to encourage the person to discriminate between their feelings and experiences and those of the caregiver, thereby undermining the development of a robust mentalising capacity.

Individuals constitutionally vulnerable and/or exposed to influences that undermine the development of social cognitive capacities, such as neglect in early relationships, develop with an impaired ability both to represent and to modulate affect and effortfully control attentional capacity. These factors, with or without further trauma, exemplified by severe neglect, abuse and other forms of maltreatment, may cause changes in the neural mechanisms of arousal and lead to structural and functional changes in the developing brain. Unless adequate remedial measures are taken, borderline personality may develop.

General adult mental health services in England and Wales offer varying levels of service provision for people with personality disorder. England and Wales have a health service in which personality disorder services are considered to be an integral part. As the decision to expand services to include the treatment of personality disorder was only made in 2003 the development of these services remains patchy and, in some areas, rudimentary. Although these services are for personality disorder generally, most users seeking services are likely to have a diagnosis of borderline personality disorder and this is anticipated in the service provision.

The programme in England includes the development of innovative psychosocial approaches to treatment, national service pilot projects and a workforce and training programme. The long-term plan is to develop capacity for specific personality services in all parts of the country.

Comorbid mental illness, particularly depression, bipolar disorder, PTSD, substance misuse disorder and psychosis are more common in people with borderline personality disorder than in the general population; lifetime prevalence of at least one comorbid mental illness approaches 100% for this group (Bender et al., 2001). In addition, many of the trait- and state-related symptoms of borderline personality disorder (including affective instability, transient stress-related psychotic symptoms, suicidal and self-harming behaviours, and impulsivity) are similar in quality to those of many types of mental illness and could intuitively be expected to respond to drug treatment.

The use of antidepressants, mood stabilisers and antipsychotics is common in clinical practice. One large study of prescribing practice in the US found that 10% of people with borderline personality disorder had been prescribed an antipsychotic at some point during their contact with services, 27% a mood stabiliser, 35% an anxiolytic and 61% an antidepressant (Bender et al., 2001); the lifetime prescribing rate for antidepressants was double that for patients with major depression. There are no published UK-based studies of prescribing practice, but given that people with borderline personality disorder tend to seek treatment, there is no reason to suspect that the prevalence of prescribing of psychotropic medication differs from that in the US. Such treatment is often initiated during periods of crisis and the placebo response rate in this context is high; the crisis is usually time limited and can be expected to resolve itself irrespective of drug treatment.

Often the prescribed drug is continued in an attempt to protect against further transient, stress-related symptoms and when these occur, another drug from a different class is likely to be added (Tyrer, 2002; Paris, 2002; Sanderson et al., 2002). A longitudinal study found that 75% of participants with borderline personality disorder were prescribed combinations of drugs at some point (Zanarini et al., 2003). Those who have repeated crisis admissions to hospital may be prescribed multiple psychotropic drugs in combination with a range of medicines for minor physical complaints. Adherence to medication in the medium term is often poor and the frequency with which prescriptions are altered makes it difficult to see which drug, if any, has helped and how.

The psychotropic drugs that are commonly prescribed are all associated with clinically significant side effects. For example, antipsychotic drugs may lead to considerable weight gain (Theisen et al., 2001), both compounding problems with self-esteem and increasing the risk of serious physical pathology such as diabetes and cardiovascular disease (Mackin et al., 2005). Lithium can cause hypothyroidism and is a very toxic drug in overdose; valproate can lead to weight gain and is a major human teratogen (Wyszynski et al., 2005); and selective serotonin re-uptake inhibitors (SSRIs) can cause unpleasant discontinuation symptoms if they are not taken consistently (Fava, 2006). The balance of risks and benefits of psychotropic drugs is generally even more unfavourable in adolescents and young adults: the risks associated with SSRIs, which have been associated with treatment-emergent suicidal ideation in young people (Hammad et al., 2006), may outweigh the benefits (Whittington et al., 2004), and valproate may increase the risk of young women developing polycystic ovaries (NICE, 2006a; NICE, 2007a).

No psychotropic drug is specifically licensed for the management of borderline personality disorder, although some have broad product licences that cover individual symptoms or symptom clusters. Where there is a diagnosis of comorbid depression, psychosis or bipolar disorder, the use of antidepressants, antipsychotics and mood stabilisers respectively would be within their licensed indications. Where there are depressive or psychotic symptoms, or affective instability, that fall short of diagnostic criteria for mental illness, the use of psychotropic drugs is largely unlicensed or ‘off-label’. Prescribing off-label places additional responsibilities on the prescriber and may increase liability if there are adverse effects (Baldwin, 2007). As a minimum, off-label prescribing should be consistent with a respected body of medical opinion (Bolam test) and be able to withstand logical analysis (House of Lords, 1997). The Royal College of Psychiatrists recommends that the patient be informed that the drug prescribed is not licensed for the indication it is being used for, and the reason for use and potential side effects fully explained (Baldwin, 2007).

The history of specific psychological interventions designed to help people with borderline personality disorder is intertwined with changing conceptions of the nature of the disorder itself. The emergent psychoanalytic concept of ‘borderline personality organisation’, intermediate between neurosis and psychosis (Stern, 1938; Kernberg, 1967), was influential in the introduction of borderline personality disorder into DSM-III in 1980, but was not an approach taken by ICD-10. The borderline personality disorder concept was therefore first adopted in the US and had no wide currency in the UK before the mid-1980s. At this time, although a range of psychodynamic, experiential, behavioural and cognitive behavioural therapies were available within NHS mental health services, they were very patchy and in short supply. Cognitive therapy (CT) for depression was only in the early stages of being adopted. Many people who would now be described in terms of having borderline personality disorder presented with depression, anxiety and interpersonal difficulties and were offered these therapies. This spurred innovation as practitioners began to modify these techniques in order to help people with more complex psychological difficulties, and during the 1980s and 1990s systematic methods were developed specifically for this client group.

Specific therapies for borderline personality disorder, therefore, developed through modification of existing techniques. In both the US and UK, psychoanalytic methods were adapted to provide more structure, containment (such as explicit contracts between therapist and client) and responsiveness; for example, the classical technique of the ‘blank screen’ of therapist neutrality and abstinence was modified so that the therapist became more active. Derived (but distinct) from classical analytic technique, an approach based on developmental attachment theory led to a specific therapy emphasising mentalisation. A behavioural approach to self-harm and suicidality that incorporated skills training in emotion regulation and validation of client experience developed into dialectical behaviour therapy (DBT), a specific intervention for borderline personality disorder per se. Cognitive analytic therapy (CAT), which had from its outset explicitly addressed interpersonal difficulties, gained greater application to borderline problems through theoretical and practical attention to partially dissociated states of mind and their functional analysis. CT for depression was also adapted to personality disorders. For example, one method paid greater attention to the early maladaptive schemas underpinning cognitive biases. Adaptations have also been made in cognitive behavioural therapy (CBT) and interpersonal therapy (IPT). Some of these adapted therapies are offered as psychological therapy programmes (for example, mentalisation-based partial hospitalisation and DBT); others are provided as more straightforward time-limited one-to-one or group treatments (for example, CBT or CAT).

Despite the developments of these specific psychological therapies (see Chapter 5), most ‘talking treatments’ offered to people with borderline personality disorder in the NHS are generic or eclectic and do not use a specific method. Clinical psychologists are trained to work flexibly around a range of assessment, treatment and rehabilitation needs, through psychological formulation, treatment planning, staff supervision and environmental change. The British Psychological Society requires chartered clinical and counselling psychologists to train in two evidence-based psychological therapies, with further post-qualification training required before they can register as practitioners. However, they may not use a specific approach during therapy sessions and, where a specific approach is used, it may not be available in the optimum format, that is, the one that was tested in clinical trials. A good example is DBT, which is a psychological therapy programme delivered by a team of therapists that includes one-to-one therapy sessions, psychoeducational groups and telephone support. Although NHS therapists may have trained in the method, it has proved organisationally difficult to ensure all elements of the DBT approach are available in practice.

Psychological and psychosocial interventions are delivered in a variety of ways and settings within the NHS by clinical psychologists and other staff trained in psychological therapies, such as psychiatrists, nurses, social workers and other mental health therapists. Individual and group therapies are available in psychology and psychotherapy departments, within day services and community mental health services. Day services have been established with specific expertise in programmes for this client group, some based on therapeutic community principles, but these are not universally available. In 2005, 11 pilot services were funded to demonstrate a range of service possibilities. All of these specified some element of psychological care, although few were based on provision of specific and formal psychological therapies (Crawford et al., 2007).

In practice, the limiting factor in providing access to psychological therapies is the very small proportion of NHS staff trained to deliver these to a competent standard. A further challenge is how to embed psychological treatment into the overall care programme in health and social care, which may involve liaison among staff from many agencies who do not share a psychological understanding of the nature of the disorder. To address this, a psychological therapies framework can be applied to the care programme through multidisciplinary team-based training (Sampson et al., 2006; Kerr et al., 2007).

Together with greater understanding of the developmental origins and psychological mechanisms underpinning this disorder and epidemiological evidence on its natural history, the emergence of at least partially effective psychological treatments has challenged traditional views of borderline personality disorder as immutable. The therapeutic nihilism so characteristic of earlier decades is giving way to a belief that psychological therapies have an important role to play in the overall care, treatment and recovery of people with these disorders.

Arts therapies developed mainly in the US and Europe. They have often been delivered as part of treatment programmes for people with personality disorders including those with borderline personality disorder. Arts therapies include art therapy, dance movement therapy, dramatherapy and music therapy which use arts media as its primary mode of communication; these four therapies are currently provided in the UK. Arts therapies are normally undertaken weekly, and a session lasts 1.5 to 2 hours. Patients are assessed for group (typically four to six members) or individual therapy. The primary concern is to effect change and growth through the use of the art form in a safe and facilitating environment in the presence of a therapist. Arts therapies can help those who find it hard to express thoughts and feelings verbally. Traditionally, art therapy is thought of as working with primitive emotional material that is ‘pre-verbal’ in nature, and thus made available to exploration and rational thought. The nature of the therapist’s work can thus be similar to the interpretations of psychoanalysis, or less interpretative and more supportive, to enable patients to understand what they want to understand from the work. For people with more severe borderline personality disorder, it is generally accepted that ‘plunging interpretations’ without sufficient support are unlikely to be helpful (Meares & Hobson, 1977).

Arts therapies are more concerned with the process of creating something, and the emotional response to this and/or the group dynamics of this. This can be very active (involving the physical characteristics of the art work and movement), playful, symbolic, metaphorical or lead directly to emotions that need to be understood. Such understanding may be achieved through subsequent discussion, and the use of the art materials when helpful.

A therapeutic community is a consciously designed social environment and programme within a residential or day unit in which the social and group process is harnessed with therapeutic intent. In the therapeutic community the community itself is the primary therapeutic instrument (Kennard & Haigh, 2009).

In England therapeutic communities first emerged in a form that we would recognise today during the Second World War, at Northfield Military Hospital in Birmingham and Mill Hill in London. The leaders of the Northfield ‘experiments’ were psychoanalysts who were later involved in treatment programmes at the Tavistock Clinic and the Cassel Hospital, and had considerable international influence on psychoanalysis and group therapy. The Mill Hill programme, for battle-shocked soldiers, later led to the founding of Henderson Hospital and a worldwide ‘social psychiatry’ movement, which brought considerably more psychological and less custodial treatment of patients of mental hospitals throughout the Western world.

Different forms of therapeutic community have evolved from these origins, one clear strand of which is for specific treatment of people with personality disorders. The therapeutic communities for personality disorder range from full-time residential hospitals to units that operate for a few hours on one day each week. Although, as stated above, the community itself is the primary therapeutic agent, programmes include a range of different therapies, usually held in groups. These can include small analytic groups, median analytic groups, psychodrama, transactional analysis, arts therapies, CT, social problem solving, psychoeducation and gestalt. In addition to specific therapies, there are community meetings and activities.

Therapeutic communities generally use a complex admission procedure, rather than straightforward inclusion and exclusion criteria. This results in diagnostic heterogeneity, and none claims to treat borderline personality disorder exclusively; however recent work has demonstrated that the admission characteristics of members show high levels of personality morbidity, with most exhibiting sufficient features to diagnose more than three personality disorders, often in more than one cluster. The admission phase includes engagement, assessment, preparation and selection processes before the definitive therapy programme begins and is a model of stepped care, where the service users decide when and whether to proceed to the next stage of the programme. A voting procedure by the existing members of the community, at a specifically convened case conference or admissions panel, is normally used to admit new members. Programmes and their various stages are time limited, and none of the therapeutic communities specifically for personality disorder is open ended. Some have formal or informal, staff or service-user led post-therapy programmes.

Staff teams in therapeutic communities are always multidisciplinary, drawn mostly from the mental health core professions, including direct psychiatric input and specialist psychotherapists. They also frequently employ ‘social therapists’, who are untrained staff with suitable personal characteristics, and ex-service users. The role of staff is less obvious than in single therapies, and can often cover a wide range of activities as part of the sociotherapy. However, clear structures – such as job descriptions defining their different responsibilities, mutually agreed processes for dealing with a range of day-to-day problems and rigorous supervisory arrangements – always underpin the various staff roles.

There are several theoretical models on which the clinical practice is based, drawing on systemic, psychodynamic, group analytic, cognitive-behavioural and humanistic traditions. The original therapeutic community model at Henderson Hospital was extensively researched in the 1950s using anthropological methods and four predominant ‘themes’ were identified: democratisation, permissiveness, reality confrontation and communalism. More contemporary theory emphasises the following: the role of attachment; the ‘culture of enquiry’ within which all behaviours, thinking and emotions can be scrutinised; the network of supportive and challenging relationships between members; and the empowering potential of members being made responsible for themselves and each other. This has been synthesised into a simple developmental model of emotional development, where the task of the therapeutic community is to recreate a network of close relationships, much like a family, in which deeply ingrained behavioural patterns, negative cognitions and adverse emotions can be re-learned.

For personality disorders, the non-residential communities are mostly within the NHS mainstream mental health services, and the residential units are in both NHS and tier 3 organisations. Standards have been devised to ensure uniformity and quality of practice, and all NHS therapeutic communities for personality disorder participate in an annual audit cycle of self-review, peer review and action planning against these standards. The Department of Health in England has supported the recent development of ‘NHS commissioning standards’ upon which accreditation for therapeutic communities will be based.

This section includes various modalities that are not part of the general psychological treatments for borderline personality disorder. Group analytic psychotherapy, humanistic and integrative psychotherapy and systemic therapy can all be routinely employed in work with people with personality disorder, either as stand-alone therapies for less complex cases or as part of multidisciplinary packages of care – or long-term pathways – for those with more intractable or severe conditions.

This is also often known simply as ‘group therapy’. It is characterised by non-directive groups (without pre-determined agendas), in which the relationships between the members, and the members and the therapist (‘conductor’), comprise the main therapeutic tool. Such groups generally, and deliberately, build a strong esprit de corps and are both strongly supportive and deeply challenging. The membership of a group is fairly constant, with each member staying typically for 2 to 5 years. Suitably qualified group therapists (to United Kingdom Council for Psychotherapy [UKCP] standards) undergo at least 4 years’ training, have regular clinical supervision and undertake continuing professional development (CPD) activities.

The group process can help prevent hazardous therapeutic relationships developing with a therapist, as can happen in individual therapy with people with severe personality disorders. They can actively address relationship difficulties that are manifest ‘live’ in the group, and they can avoid difficult dependency by helping participants to take responsibility for themselves by first sharing responsibility for each other and later learning how to ask for help for themselves, in an adaptive way.

Disadvantages include difficulty in initiating participation because of the fear of personal exposure; problems of finding a regular suitable meeting space; and issues of confidentiality.

These are therapies based on a variety of theoretical models that evolved in the mid-20th century as alternatives to the dominant model of psychoanalysis. There is a significant overlap with the term ‘action therapies’, which has increasing currency. They include: psychodrama, which is group-based and aims to understand particularly difficult past emotional episodes and link them to current problems and difficulties; transactional analysis, which is based on parent, adult and child ‘ego states’ (a person’s beliefs, mannerisms and emotional responses), and can be undertaken either individually or in groups; gestalt therapy, which aims to facilitate awareness and help achieve self-regulation and self-actualisation (therapeutic techniques include empty-chair work, role reversal and enactments); and person-centred therapy developed from Carl Rogers’ humanistic approach.

This is most commonly used for work with families (or support networks), for example, where the index patient is a child. It aims to maximise family strengths and resilience to help people overcome problems experienced by individual family members or the family as a whole. It helps family members to understand how they function as a family and to develop more helpful ways of interacting with and supporting each other. It uses a format with long but widely-spaced sessions, for example 2 hours every 6 weeks. It requires a supervising team who watch the session live or who listen to it with audio equipment, and who discuss hypotheses of how the system is working and actions to bring about change. The individual and family or support network have access to the ideas and hypotheses discussed in the team, so that different experiences and points of view can be heard and acknowledged. The therapists help the family (or support network) to bring about the changes that they have identified as therapeutic goals. There are a number of models of systemic theory and interventions, such as Milan, social constructionist, narrative, solution focused, structural and strategic. The interventions are generally ‘structural’ or ‘strategic’, and include the use of such techniques as circular questioning (for example, ‘what would your brother think about your mother’s answer to that question?’), reframing and mapping the system with genograms (a pictorial representation of a patient’s family relationships).

In cases of personality disorder where the dynamics within a whole family may be important in maintaining or exacerbating the presenting range of problems, and the family members are willing to participate, systemic therapy can be effective at starting new ways of communicating within a family that may be self-sustaining.

Nidotherapy, from the Latin, nidus, meaning nest (Tyrer et al., 2003a), is distinct from psychotherapeutic approaches in that the emphasis is on making environmental changes to create a better fit between the person and their environment. In this sense it is not specifically a treatment, but it does have a therapeutic aim of improving quality of life, through acceptance of a level of handicap and its environmental accommodation.

The perceived enduring and chronic nature of personality disorder poses a challenge to a healthcare system that is historically, and to a large extent still is, strongly influenced by the biological (illness) paradigm of mental health. Essentially, mental health services within the NHS have been configured in such a way as to ‘treat’ people during the acute phases of their illness. As personality disorders by their definition do not have ‘acute’ phases some have argued that a personality disorder should not be the responsibility of the NHS (see Kendell [2002] for further discussion).

Given the confusion that surrounds the nature of personality disorder, it is not surprising that this has impacted on NHS care for people with this diagnosis. Until recently, personality disorder services in the NHS had been diverse, spasmodic and inconsistent (Department of Health, 2003).

In line with the NSF for Mental Health (Department of Health, 1999a) the National Institute for Mental Health in England (NIMHE) produced policy implementation guidance for the development of services for people with personality disorder (Department of Health, 2003). The main purpose of this document was:

• to assist people with personality disorder who experience significant distress or difficulty to access appropriate clinical care and management from specialist mental health services

• to ensure that offenders with a personality disorder receive appropriate care from forensic services and interventions designed both to provide treatment and to address their offending behaviour

• to establish the necessary education and training to equip mental health practitioners to provide effective assessment and management.’ (Department of Health, 2003).

The Personality Disorder Capabilities Framework (NIMHE, 2003) soon followed. This document set out a framework to support the development of the skills that would enable practitioners to work more effectively with people with personality disorders. It also aimed to provide a framework to support local and regional partners to deliver appropriate education and training (NIMHE, 2003). This document did not focus solely on the needs of NHS organisations; it had a wider remit to include all agencies that had contact with people who met the diagnosis. These two documents, along with investments in pilot personality disorder services and training initiatives, have signalled a significant change in the perspective of the NHS on personality disorder and have led to its commitment to enhance and improve its service.

The role of social services, in providing care and support to people with mental health problems, covers a wide range of people, from those with mild mental health problems to people with severe and enduring mental disorders (Department of Health, 1998). Historically, care provided by social services is determined by the person’s social need and is less influenced by diagnosis and the biological paradigm than the NHS. After the 1998 White Paper on modernising social services (Department of Health, 1998), which aimed to set new standards of performance and to allow the NHS and social services to have closer partnerships in meeting the standards set down in the NSF for mental health, local implementation teams were set up across the country. With respect to personality disorder, their role is to review the progress that local mental health and social care services are making towards implementing the NSF’s targets for personality disorder.

In law, personality disorder is generally seen as distinct from ‘serious mental illness’ because it is not considered to reduce the person’s capacity to make decisions (Hart, 2001). Instead, it is thought of as an aggravating condition (Hart, 2001). Nevertheless, new legislation in the Mental Health Act amendment (HMSO, 2007) and the Mental Capacity Act (HMSO, 2005) will change both the rights and protections for people with personality disorders and their access to services. However, the legal position that people with personality disorder have held throughout the history of psychiatry has undoubtedly influenced the perspective of the criminal justice system regarding personality disorder and goes some way to explain why most people with personality disorder would generally find themselves in the criminal justice system as opposed to forensic mental health services. It is not uncommon within forensic mental health services for regional secure units to actively exclude patients with a primary diagnosis of personality disorder, because they do not consider this to be their core business (Department of Health, 2003). In many parts of the country there are no specific services, and, when services are offered, they tend to be idiosyncratic.

In March 1999, a report commissioned by the Department of Health about the future organisation of prison healthcare (Department of Health, 1999b) proposed that people in prison should have access to the same quality and range of services (including mental health) as the general public (Department of Health, 1999b). In the same year the NSF called for closer partnerships between prisons and the NHS at local, regional and national levels (Department of Health, 1999a). The emphasis was on a move towards the NHS taking more responsibility for providing mental healthcare in prisons and establishing formal partnerships.

In July 1998, the Secretary of State announced a review of the 1983 Mental Health Act (Department of Health, 1983), triggered by concerns that current legislation did not support a modern mental health service. These concerns were reiterated in the NSF for mental health since ‘neither mental health nor criminal justice law currently provides a robust way of managing the small number of dangerous people with severe personality disorder’ (Department of Health, 1999a).

Diagnosing borderline personality disorder in young people under 18 has often caused controversy. Although borderline personality disorder is thought to affect between 0.9 and 3% of the community population of under 18 year olds (Lewinsohn et al., 1997; Bernstein et al., 1993), there is some uncertainty about the rate (see Chapter 9). There are also certain caveats in DSM-IV and ICD-10 when making the diagnosis in young people (see Chapter 9). However young people with borderline personality disorder often present to services in seek of help (Chanen et al., 2007a). Because interventions for young people with borderline personality disorder will usually be provided by specialist CAMHS, which has a different structure from adult mental health services, a full discussion of the issues relating to young people with borderline personality disorder can be found in Chapter 9.

There are particular issues for people with borderline personality disorder regarding the diagnosis, the label and associated stigma, which can have an impact on people accessing services and receiving the appropriate treatment. These issues are fully explored in Chapter 4, which comprises personal accounts from people with personality disorder and from a carer, and a review of the literature of service user and family/carer experience.

The families and carers of people may also feel unsupported in their role by healthcare professionals and excluded from the service user’s treatment and care. The issues surrounding this are also further explored in Chapter 4. Although there are debates around the usefulness and applicability of the word ‘carer’, this guideline uses the term ‘families/carers’ to apply to all people who have regular close contact with the person and are involved in their care.

Besides functional impairment and emotional distress, borderline personality disorder is also associated with significant financial costs to the healthcare system, social services and the wider society. The annual cost of personality disorders to the NHS was estimated at approximately £61.2 million in 1986 (Smith et al., 1995). Of this, 91% accounted for inpatient care. Another study conducted in the UK, estimated the costs of people with personality disorders in contact with primary care services (Rendu et al., 2002). The study reported that people with personality disorders incurred a cost of around £3,000 per person annually, consisting of healthcare costs and productivity losses; in contrast, the respective cost incurred by people without personality disorders in contact with primary care services was £1,600 (1998/99 prices). In both groups, productivity losses accounted for over 80% of total costs. Dolan and colleagues (1996) assessed the cost of people with personality disorders admitted to a UK hospital over 1 year prior to admission; this cost was reported to reach £14,000 per person (1992/93 prices), including inpatient and outpatient health-care costs, as well as prison-related costs (which amounted to approximately 10% of the total cost). Although the two UK studies (Rendu et al., 2002; Dolan et al., 1996) differed in methodology and costs considered, this difference in costs may be partly attributed to the different levels of severity of the disorders apparent in the two study populations (people engaged with general practice services versus people admitted to hospital).

The economic cost of personality disorders has been assessed in other European countries as well: in Germany, inpatient treatment of borderline personality disorder was estimated at €3.5 billion annually, covering about 25% of the total costs for psychiatric inpatient treatment in the country (Bohus, 2007). In the Netherlands, the average cost of a person with personality disorder referred for psychotherapeutic treatment was estimated at €11,000 (2005 prices) over 12 months prior to treatment (Soeteman et al., 2008). Of this, 66.5% was associated with healthcare expenditure, while the rest reflected productivity losses. According to another study (Van Asselt et al., 2007), the average cost per person with borderline personality disorder in the Netherlands was €17,000 in 2000. Of this, only 22% was health-related. The remaining cost was incurred by out-of-pocket expenses, informal care, criminal justice costs and productivity losses. Based on this average cost and a prevalence of borderline personality disorder of 1.1%, the study estimated that the total societal cost of borderline personality disorder in the Netherlands reached €2.2 billion in 2000. The authors noted that the direct medical costs represented only 0.63% of total Dutch healthcare expenditure in 2000, which meant that, given the 1.1% prevalence of the condition, people with borderline personality disorder seemed to use a less than proportionate share of the healthcare budget. However, the authors acknowledged that people in institutional care were not part of the study sample, and therefore medical costs associated with borderline personality disorder might have been underestimated.

Treatment-seeking people with personality disorders have been reported to place a high economic cost on society, compared with people with other mental disorders such as depression or generalised anxiety disorder (GAD) (Soeteman et al., 2008). People with borderline personality disorder make extensive use of more intensive treatments, such as emergency department visits and psychiatric hospital services (Bender et al., 2001 & 2006; Chiesa et al., 2002), resulting in higher related health-care costs compared with people with other personality disorders and major depression (Bender et al., 2001 & 2006). In addition, they are more likely to use almost every type of psychosocial treatment (except self-help groups) and to have used most classes of medication compared with people with depression (Bender et al., 2001). However, an American prospective study that followed people with borderline personality disorder over 6 years (Zanarini et al., 2004a) reported that, although hospitalisation rates and rates of day or residential treatment were high at initiation of the study, these significantly declined overtime; similar patterns were observed for rates of intensive psychotherapy, although engagement in less intensive psychosocial therapeutic programmes remained stable over the 6 years of the study. Polypharmacy was a characteristic of people with borderline personality disorder that was not affected by time, with 40% of people taking three or more concurrent standing medications, 20% taking four or more and 10% taking five or more, at any follow-up period examined. The authors concluded that the majority of people diagnosed with borderline personality disorder carry on outpatient treatment in the long term, but only a declining minority continue to use restrictive and more costly forms of treatment.

The level of severity of symptoms of borderline personality disorder determines the level of usage of healthcare resources: in a study conducted in a primary care setting in the US, the severity of symptoms experienced by women with borderline personality disorder was shown to predict increased use of primary healthcare resources (Sansone et al., 1996). This finding was consistent with the findings of another American study that examined male veterans with borderline personality disorder (Black et al., 2006); the study reported that as the number of symptoms associated with borderline personality disorder increased, so did the levels of psychiatric comorbidity (such as depression, PTSD and GAD), the levels of suicidal and self-harming behaviour, as well as the rates of utilisation of healthcare resources (that is, inpatient stays, outpatient visits and emergency department visits). Moreover, the number of symptoms observed was positively related to rates of incarceration and other contacts with military forensic services (which are expected to incur extra costs). Psychiatric comorbidity is common in people with borderline personality disorder (Bender et al., 2001; Black et al., 2006) and, when present, results in a significant increase in total healthcare costs (Bender et al., 2001; Rendu et al., 2002).

The reported resource use and cost estimates have been made by studying people with borderline personality disorder in contact with health services. However, it is known that a significant proportion of people with personality disorders fail to seek treatment and, when they do, future disengagement with services is quite common. Moreover, contacts with social services, problems with housing, levels of unemployment and involvement with the criminal justice system incur further substantial costs that have not been thoroughly examined, if at all. Therefore, the financial and psychological implications of borderline personality disorder to society are likely to be wider than those suggested in the literature. Efficient use of available healthcare resources is required to maximise the benefits for people with borderline personality disorder, their family and carers, and society in general.