Psychological treatment for personality disorders

Abstract

There is encouraging evidence that some patients with personality disorder are treatable. Psychotherapeutic interventions show promise, although interpretation of the literature is problematic: the number of patients in most trials is small, outcome measures are questionable, follow-up is limited, and treatments are multifaceted, complex interventions in which the effective components are unclear. The evidence base can be assessed according to efficacy and generalisability, and when both are taken into account the best verification is for psychodynamic therapy. However, there is inadequate evidence to make specific recommendations for any particular therapy.

Personality disorder is common and rarely treated in its own right, although interest in the subject is growing and, with it, a demand for effective treatments. This demand has not been met adequately but there has been a significant shift from the view that personality disorder is untreatable; we do have treatments that have at least some efficacy and this article describes the evidence base for these. However, the treatment of personality disorder is not just the simple application of therapy: the planning and management of treatment is at least equally important. This is the subject of a further article.

What treatments are available and how do we determine their success?

Although it is common for every known treatment to be used at some time for conditions that are deemed untreatable, almost all of the psychotherapeutic interventions for personality disorder that carry some elements of respectability are listed in Table 1. Each of these will be discussed in turn. In discussing the efficacy of treatments for personality disorder, we have to be aware of the special problems associated with assessment of these conditions:

Table 1 Summary of treatments used for personality disorder and the level at which they have been investigated

Type of treatment	Main purpose of treatment	Evidence level 1
Psychodynamic therapy	To increase reflective capacity and emotional and interpersonal understanding	1
Cognitive–behavioural therapy	To alter dysfunctional core beliefs	1
Dialectical behaviour therapy	Initially, to reduce self-harm; eventually, to achieve transcendence	1
Therapeutic community	To effect attitudinal and behavioural change	2
Cognitive analytic therapy	To achieve greater self-understanding	3
Behaviour therapy	To improve maladaptive behaviour	3
Nidotherapy	To achieve better environmental adjustment, thus minimising impact of disorder (Tyrer, 2002a )	4

1. • the high level of comorbidity with other disorders of both personality and mental state (Tyrer et al, 1997);

2. • the fluctuating nature of personality status over time, mainly as a consequence of concomitant mood changes (Clark et al, 2003);

3. • the need to have a long period of observation, preferably at least a year, before a treatment can be said to be properly evaluated;

4. • the recognition that personality disorder is a multifaceted condition that can be influenced in many different ways and fully justifies the use of what are now described as ‘complex interventions’ (Campbell et al, 2000) to treat it. Complex interventions lead to complex evaluations and consequent greater difficulty in interpreting results.

The other important distinction in determining the value of an intervention is whether this has been demonstrated in standard practice or under strict experimental conditions. This distinction was first highlighted by Schwarz & Lellouch (1967) in discussing differences between randomised trials, and it is commonly described as the difference between an explanatory trial, i.e. a trial in which treatments are compared under ideal (experimentally manipulated) conditions, and a pragmatic trial, in which the study is carried out under the conditions normally appertaining to ordinary practice. In the latter trial, possible confounders to the intervention may be present, and although they could be removed, to do so would create an artificial environment that would not allow the results to be transferred to ordinary practice.

Schwarz & Lellouch (1967) showed that the results of explanatory and pragmatic trials could be very different even though the treatments under test were the same. As personality disorders commonly occur in conjunction with other disorders, there is a place for both in their evaluation. In evidence-based psychiatry these are sometimes described as trials of efficacy (explanatory) and effectiveness (pragmatic). Each has its advantages and disadvantages, but in general it is common to establish efficacy first under controlled conditions, before testing an intervention in conditions of ordinary practice.

Because of the formidable difficulties of meeting these requirements (Box 1), there has been a tendency for investigators to abandon them or at least to fulfil them only partially, and the literature on personality disorder, particularly with psychotherapeutic interventions, is often difficult to interpret (Bateman & Fonagy, 2000).

Box 1 Requirements for establishing an effective treatment

1. • Efficacy in randomised controlled trial over control treatments when used for a pure form of the personality disorder

2. • Similar outcome in pragmatic randomised controlled trial

3. • Consistency across settings when used with appropriate treatment fidelity

4. • Maintenance of outcome over time (preferably more than 1 year) because of the long duration of personality disorder

Choice of outcome measures

Outcome measures have always been a problem in psychiatric disorders, but these difficulties become even more pronounced in the case of personality disorders because there are no agreed common outcomes. The wide range of those that have been measured are summarised in Table 2.

Table 2 Outcome measures used in efficacy studies of interventions for personality disorder

Outcome	Forms of measurement	Advantages and disadvantages
Change in symptoms	Standard rating scales (e.g. for depression, impulsiveness)	Useful to compare with other (symptomatic) conditions, but do not cover personality directly
Change in (social) function	Rating scale measures of disability and social function	Social function may be a good proxy measure of personality disorder, but is also influenced by factors independent of personality
Change in self-harm behaviour	Measures to record episodes of self-harm	Simple to record, but a problem to define, and only measure one part of personality disturbance (one of the core borderline features)
Recidivism rates	Standard records of re-offending	An easy measure of societal impact of antisocial behaviour but is not a direct measure of any form of personality disturbance
Rates and duration of psychiatric hospitalisation	Obtained from standard records (and usually costed by nature of institution)	Relevant to costs of care (in-patient costs constitute by far the greatest proportion of total costs), but is not necessarily related to personality disorder
Global outcome scales	Standard scales of global improvement	Relevant in that personality affects all modes of function so these are covered by global improvement, but unsatisfactory in that many other factors influence global outcome
Quality of life	Standard quality of life scales	The outcome measure with the greatest face validity, but most scales are not very sensitive to personality changes
Changes in personality status	Structured interview schedules	These are the most obvious measures to record, but the short-term stability of personality is poor (see text)

Personality disorders affect both the individual and society, and a range of outcomes can be measured to cover all these possibilities. Forensic psychiatrists and the general public may consider the effect on society to be the most important measure and insist that outcome for mentally disordered offenders is best measured by the frequency of re-offending. This is an easily measured, reliable statistic, although subject to social and judicial variation over time and distorted by a range of other factors such as the fact that patients who spend a long time in hospital or prison are not likely to reoffend. However, it does not record either symptomatic or personality changes, which are equally important, although they too may be subject to distortion. Changes in symptoms may be a consequence of change in mental state disorders quite independent of personality. In general, the interaction between personality disorder and Axis I disorder could result potentially in either an exaggeration of or an obscuring of a genuine treatment effect. Treatment might improve the mental illness but not the personality dysfunction or vice versa, and improvement in one may follow improvement in the other. Conversely, it is well known that mental illness can create the impression of abnormal personality, but that these characteristics change as the mental illness improves. Thus, any measured change in personality should be regarded in the first instance as an artefact related to improvement or deterioration in mental state.

Because of these difficulties, there is a tendency to use global outcome measures such as the Global Assessment Scale (Endicott et al, 1976) to determine the degree of improvement in personality disorders in long-term follow-up studies, although a battery of measurements, covering different domains, is normally used in short-term treatment studies. Unfortunately, there is no standardised procedure for recording global outcome measures, but it is reasonable to take into account any, several, or all of the measures in Table 2, provided that there is a clear distinction between primary and secondary outcomes.

Treatment evaluation

In summarising each of the treatments, we have taken the nature of the evidence base, the choice of outcomes and the quality of studies into account and given a scores for efficacy and generalisability. These are shown in Table 3.

Table 3 Efficacy and generalisability scores for psychological treatments of personality disorder

Treatment	Efficacy score 1	Generalisability score 2
Psychodynamic therapy	2	2
Cognitive analytic therapy	1	2
Cognitive therapy	1	2
Dialectical behaviour therapy	2	1
Therapeutic community	2	1

Many people with personality disorders have conditions with which they themselves are comfortable (i.e. they are egosyntonic) and have no wish to change. There is considerable variation between personality disorders in this respect. Those unwilling to have treatment can be described as type R (treatment-resisting), as opposed to type S (treatment-seeking); most of those with paranoid, antisocial and schizoid personality disorders come into the type R category, whereas those with borderline personality disorder are more often type S (Tyrer et al, 2003a ). This classification oversimplifies. Overt treatment seekers may be frightened of change, having gained a moderate level of psychological equilibrium over time through their interaction with services, or be seeking the gratification of an interactive relationship with medical services rather than looking for personal development. Others may avoid treatment but be desperate for help.

However, it reminds clinicians to assess motivation for change and explains why most of those who have personality disorder (3 out of 4) do not attend for treatment, as they belong to the type R group.

Our interpretation of the literature should be considered first in the context of personal bias (P.T. is especially interested in the role of cognitive–behavioural therapy and A.W.B. in psychodynamic therapy for the treatment of personality disorder), and second in the light of a recent objective meta-analysis of the effectiveness of psychodynamic therapy and cognitive–behavioural therapy in the treatment of personality disorders. Leichsenring & Leibing (2003) found that psychodynamic therapy yielded a large overall effect size of 1.46, with effect sizes of 1.08 for self-reported measures and 1.79 for observer-rated measures. This contrasts with cognitive–behavioural therapy, in which the corresponding values were 1.00, 1.20 and 0.87 respectively. In addition, the psychodynamic studies had a mean follow-up period of 1.5 years compared with only 13 weeks for the cognitive–behavioural therapy. However, the research literature is not extensive enough to draw firm conclusions from meta-analysis, and Leichsenring & Leibing were able to include only 14 psychodynamic studies and 11 cognitive-behavioural therapy studies. In addition, the effect sizes cannot be compared directly because the studies differ, even within the same therapy group, in terms of therapy content, patient populations, outcome assessments and other variables.

Psychological treatments

Psychodynamic therapy

Psychodynamic therapy has long been recommended, and has been modified to suit the particular problems associated with personality disorder. Most of the therapeutic interest in psychotherapy has related to borderline personality disorder (Higgitt & Fonagy, 1992), which differs from almost all others in frequent help-seeking behaviour and evident wish to change.

There have been few published randomised trials of dynamic psychotherapy (Table 4), and only two have been widely noted. Winston et al(1991) showed no difference between short-term dynamic psychotherapy and brief adaptational psychotherapy, but both interventions were somewhat superior to a waiting-list control. This study specifically excluded patients with borderline and narcissistic features, although a later study including some patients with Cluster B disorders produced similar results (Winston et al, 1994).

Table 4 Trials of psychodyamic therapy in personality disorder

Form of psychotherapy	Personality diagnosis	Reference	Type of study (n)	Main results
Psychodynamic with partial hospitalisation (day treatment)	Borderline	Bateman & Fonagy (1999, 2001, 2003)	Randomised controlled trial (n=38)	Psychotherapy superior to treatment as usual for self-harm, hospital admissions, costs and symptoms, but benefit delayed for 6 months
Brief psychodynamic therapy (two types)	Borderline	Winston et al(1991)	Randomised controlled trial (n=49, including 17 in waiting-list control group)	No significant difference between treatments
Psychotherapy and day hospital	All types (plus affective disorders)	Piper et al (1993)	Controlled trial (not randomised) (n=165)	No clear superiority of psychotherapy, but some indicators of preference
Psychotherapy preceded by short in-patient care	?Borderline	Chiesa & Fonagy (2003)	Controlled trial (not randomised) with community treatment and therapeutic community controls (n=135)	Psychotherapy preceded by in-patient treatment yielded best outcome, but results limited by fact that only those living close to unit could have this treatment

Mentalisation

The most recent support for a psychoanalytically based approach has come from a randomised study examining the effectiveness of a psychoanalytically oriented partial hospitalisation programme with standard psychiatric care for patients with borderline personality disorder (Bateman & Fonagy, 1999, 2001). This has now been fully manualised (Bateman & Fonagy, 2004) as a mentalisation-based treatment. Treatment interventions are organised to increase the reflective or mentalising capacity of the patient in the context of group and individual therapy. Mentalisation entails making sense of the actions of oneself and others on the basis of intentional mental states such as desires, feelings and beliefs. It involves the recognition that what is in the mind is indeed in the mind and reflects knowledge of one's own and others’ mental states as mental states. This capacity is enfeebled in borderline patients, and so group and individual therapy actively focuses on developing these patients’ understanding and recognition of the feelings they evoke in others and the feelings evoked in them by others.

On all outcome measures there was significantly greater improvement in those allocated to psychotherapy. The improvements in symptoms and function were delayed by several months but were greatest by the end of treatment, at 18 months. In a follow-up study, which was done on an intention-to-treat basis, gains were maintained after a further 18 months (Bateman & Fonagy, 2001), indicating that rehabilitative effects were stimulated during the treatment phase, although some treatment continued during the follow-up period. The treatment has also been found to be cost-effective (Bateman & Fonagy, 2003), but as yet the active components of therapy remain unclear, especially because it was not possible to show that mentalisation had increased in the patients who showed the most gains. An out-patient version of mentalisation-based treatment is currently being evaluated for borderline and antisocial personality disorder in a further randomised controlled trial.

Transference-focused psychotherapy

Another manualised dynamic therapy, known as transference-focused psychotherapy, that gives promising results. The outcome of a randomised controlled trial comparing transference-focused psychotherapy, dialectical behaviour therapy (see below) and supportive psychotherapy is not yet known, although its rationale has been described (Clarkin et al, 2004). In a cohort study (Clarkin et al, 2001), 23 women patients with borderline personality disorder were assessed at baseline and at the end of 12 months of treatment by means of diagnostic instruments, measures of suicidality, ratings of self-injurious behaviour and measures of medical and psychiatric service use. Compared with the year prior to treatment, the number of patients who made suicide attempts significantly decreased, as did the medical risk and severity of medical condition following self-injurious behaviour. In addition, patients during the treatment year had significantly fewer hospital admissions as well as number and days of psychiatric hospitalisation compared with the year before. The withdrawal rate was 19%. Conference reports of a comparison study between patients treated with transference focused psychotherapy and a matched untreated control group confirm the benefits of treatment (Clarkin, 2002). Costs have not yet been examined.

Community-based treatments

Chiesa and co-workers studied the relative effectiveness of three psychodynamically oriented treatment models for a mixed group of personality disorders – a long-term residential treatment using a therapeutic community approach; a briefer in-patient treatment followed by community-based dynamic therapy (step-down programme); and a general community psychiatric treatment. Initial results suggest that the brief in-patient therapeutic community treatment followed by out-patient dynamic therapy is more effective than both long-term residential therapeutic community treatment and general psychiatric treatment in the community on most measures, including self-harm, attempted suicide and readmission rates to general psychiatric wards, and is more cost-effective (Chiesa et al, 2002). Follow-up at 36 months has confirmed that patients in the step-down programme continued to show significantly greater improvement than the in-patient group on social adjustment and global assessment of mental health. In addition they were found to mutilate themselves, attempt suicide and be readmitted significantly less at 24-month and 36-month follow-up (Chiesa & Fonagy, 2003). However, the study was not a randomised trial and the groups were not strictly comparable.

In a related piece of work, in which personality status was not recorded, although disorder was likely to be present in a significant number of participants, Guthrie and her colleagues (Guthrie et al, 2001), in a randomised controlled trial, found that home-based psychodynamic/interpersonal psychotherapy was effective in reducing repetition of self-harm after 6 months.

Group psychotherapy

Non-controlled studies in which participants underwent day hospital stabilisation followed by out-patient dynamic group therapy indicate the value of using groups in borderline personality disorder (Wilberg et al, 1998). A randomised controlled trial by Marziali & Monroe-Blum (1995) concentrated on group therapy alone without the additional milieu and social components of therapy; however, their therapy was not formally psychoanalytic but focused on relationship management. They found equivalent results between group and individual therapy, concluding that, on cost-effectiveness grounds, group therapy is the treatment of choice. Further studies are needed to confirm their findings, especially since the treatment offered was less structured than most other treatments and withdrawal rates were high.

Cognitive analytic therapy

Cognitive analytic therapy has been manualised for treatment of borderline personality disorder and many are enthusiastic about its effectiveness. There are indications that this treatment method may be of help to some patients (Ryle & Golynkina, 2000), and in a small randomised trial (P. Richardson, personal communication, 2004) patients treated with cognitive analytic therapy showed significant improvement over time on a range of clinical measures. However, there was no difference between people receiving this therapy and those having other psychological treatments, so the effects may be non-specific. There was some indication that cognitive analytic therapy was judged more helpful than other psychological treatments by patients with borderline personality disorder, which might account for a lower but non-significant attrition rate in those undergoing this type of therapy. A formal randomised trial is now in progress comparing cognitive analytic therapy with ‘best available standard care’ for adolescent patients with borderline personality disorder in Victoria, Australia (Ryle, 2004).

Cognitive therapy

In cognitive therapy for personality disorders much greater emphasis is placed on changing core beliefs rather than dysfunctional thoughts and on maintaining a collaborative therapeutic alliance. Crisis intervention strategies are developed, training in self-help and self-monitoring skills is provided, and schema-focused conceptualisation is linked to behaviours that interfere with therapy, for example self-harm.

Davidson & Tyrer (1996), in an open study, used cognitive therapy for the treatment of two Cluster B personality disorders: antisocial and borderline personality disorders. They evaluated a brief (10-session) cognitive therapy approach using single-case methodology, which showed improvement in target problems (Table 5). Another small (n=34), randomised controlled trial used brief cognitive therapy, linked to a manual incorporating elements of dialectical behaviour therapy, in the treatment of recurrent self-harm in those with Cluster B personality difficulties and disorders. Eighteen patients with a history of repeated self-harm and a parasuicide attempt in the preceding 12 months were randomly allocated to manual-assisted cognitive–behavioural therapy; 16 similar patients received treatment as usual. The rate of suicide acts was lower with the intervention therapy (median 0.17 per month, compared with 0.37 per month with usual treatment; P=0.11), and self-rated depressive symptoms also improved (P=0.03). The treatment involved a mean of 2.7 sessions (each of 50 minutes duration), and the observed average cost of care was 46% less for the cognitive therapy group (P=0.22).

Table 5 Trials of cognitive–behavioural therapy in personality disorder

Form of cognitive therapy	Personality diagnosis	Reference	Type of study (n)	Main results
Beckian cognitive therapy adapted for personality disorder	Borderline and antisocial	Davidson & Tyrer (1996)	Single-case methodology study (n=10)	Slightly delayed improvement for target symptoms
Brief manual-assisted cognitive therapy	Cluster B personality disturbance	Evans et al (1999)	Randomised controlled trial (n=34)	Trend towards less frequent self-harm with MACT (P=0.11), with improvement in depression (P<0.05) compared with ordinary treatment
Manual-assisted cognitive therapy	Recurrent self-harm (42% with personality disorder)	Tyrer et al (2003b , c); Byford et al(2003)	Randomised controlled trial (n=480)	No difference between MACT and treatment as usual, but MACT more cost-effective

This study led to a much larger study (Tyrer et al, 2003c ), which probably comes closer to the model of effectiveness and generalisability than other studies. This was an extension of the manual-assisted cognitive–behavioural therapy study described earlier, with up to seven sessions of treatment offered to those with recurrent self-harm, 42% of whom had a personality disorder. It differed from other studies in being large (n=480), being multicentred (five centres in Scotland and England), using ordinary therapists (trained in the approach) in the course of their normal work, and offering no special service for those in the trial. In particular, those who did not attend appointments were not visited at home, as this was not part of normal practice. The results were, in general, negative in terms of efficacy compared with treatment as usual (which included psychotherapy and problem-solving treatment). Only 60% of patients attended for face-to-face sessions of the intervention therapy, and for the primary outcome, proportion of patients repeating self-harm, 39% of those allocated to manual-assisted cognitive–behavioural therapy repeated self-harm compared with 46% allocated to treatment as usual (P=0.20); there were seven suicides, five in the usual treatment group (Tyrer et al, 2003b ). Frequency of self-harming behaviour was also reduced by 50% in the intervention therapy group compared with the usual treatment group, but there was great variation in episodes of self-harm (Tyrer et al, 2004). There was no difference in any of the secondary outcomes. However, important differences between some of these outcomes have been shown between therapists judged as competent after assessment of taped interviews in the study compared with those who were less competent (Davidson et al, 2004), and manual-assisted cognitive–behavioural therapy led to a cost saving of £900 per patient compared with usual treatment at 6 months, although this did not remain significant at 12 months (Byford et al, 2003). Interestingly, in borderline personality disorder manual-assisted cognitive–behavioural therapy increased total costs, in contrast to its effect in other personality disorders, and had less satisfactory results in reducing self-harm (Tyrer et al, 2004).

The overall evidence in favour of cognitive–behavioural therapy in treatment of personality disorder is therefore relatively slim, with much of it coming from one research group, but it has involved more patients than any other form of treatment Further research is required into personality disorder itself, and manual-assisted cognitive–behavioural therapy is currently being evaluated for borderline personality disorder in a three-centre randomised controlled trial.

Dialectical behaviour therapy

Dialectical behaviour therapy is a special adaptation of behaviour therapy. It is a manualised therapy (Linehan, 1993) which includes techniques at the level of behaviour (functional analysis), cognition (e.g. skills training) and support (empathy, teaching management of trauma). The aim of the therapy is first to control self-harm but then, more importantly, to promote change in the emotional dysregulation that is judged to be the core of the disorder (Robins, 2003). Dialectical behaviour therapy was originally used for the treatment of a group of repeatedly parasuicidal female patients with DSM–III–R borderline personality disorder (Linehan et al, 1991). The 44 female patients admitted to the trial had each made at least two suicide attempts in the previous 5 years, with one in the preceding 8 weeks. Twenty-two of them were assigned to dialectical behaviour therapy and 22 to the control condition.

Treatment with dialectical behaviour therapy for 1 year compared with treatment as usual led to a reduction in the number and severity of suicide attempts and decreased the frequency and length of in-patient admission. However, there were no between-group differences on measures of depression, hopelessness or reasons for living.

Follow-up was naturalistic, based on the proposition that the morbidity of this group precluded termination of therapy at the end of the experimental period. At 6-month follow-up, dialectical behaviour therapy patients continued to show less parasuicidal behaviour than controls, although at 1 year there was no between-group difference (Linehan et al, 1993). Although at 1 year the dialectical behaviour therapy patients had had fewer days in hospital, at the 6-month assessment there was no between-group difference (Linehan et al, 1993).

Is its popularity justified?

The widespread adoption of dialectical behaviour therapy is a tribute both to the energy and charisma of its founder, Marsha Linehan, and to the attractiveness of the treatment, with its combination of acceptance and change, skills training, excellent manualisation, and a climate of opinion that is willing and able to embrace this multifaceted approach (Swenson, 2000). It is not, however, justified by the strength of the evidence (Tyrer, 2002b ) and conclusions about the long-term effectiveness of this therapy as a treatment for the personality itself are premature (Levendusky, 2000; Scheel, 2000; Turner, 2000). Since the original trial, handicapped by many methodological limitations, there has only been one randomised study that supports the findings unequivocally, that of Verheul et al(2003) (Table 6).

Table 6 Trials of dialectical behaviour therapy in personality disorder

Form of therapy	Personality diagnosis	Reference	Type of study (n)	Main results
Intensive (two group and one individual session per week)	Borderline (women only)	Linehan et al (1991)	Randomised controlled trial (n=44)	Significant reduction in self-harm and less hospitalisation in DBT group
Intensive group and individual therapy with specialist dialectical behaviour therapists	Borderline, drug misuse	Verheul et al (2003)	Randomised controlled trial (n=58)	Better retention than TAU group Reduction in self-harm No effect on drug use

In the study by Verheul and co-workers, 58 Dutch women who met DSM–IV criteria for borderline personality disorder were randomly assigned to either 1 year of dialectical behaviour therapy or treatment as usual, i.e. ongoing treatment in the community. Participants were clinical referrals from both addiction treatment and psychiatric services. Efficacy was measured in terms of treatment retention, and course of high-risk suicidal, self-mutilating and otherwise self-damaging behaviours. Dialectical behaviour therapy resulted in better retention rates (63% v. 23%) and significantly greater reductions of self-mutilating behaviours and self-damaging impulsive acts than treatment as usual, especially among those with histories of frequent self-mutilating behaviours. The study suggests dialectical behaviour therapy is superior to usual treatment in reducing self-mutilating and otherwise self-damaging impulsive behaviours in patients with borderline personality disorder; post hoc analyses suggested that those with more severe self-harming behaviour were helped most.

Other studies have been uncontrolled, too small to add any useful data (Koons et al, 2000) or have shown dialectical behaviour therapy to be no better than other active treatments such as the 12-step programme for alcohol dependence (Linehan et al, 2002). In addition, it is not clear which elements of dialectical behaviour therapy (psychotherapy, skills training, telephone consultation, therapist consultation team) make this treatment method effective. Two studies investigated the process of change in dialectical behaviour therapy by focusing on the possible influence of validation (Shearin & Linehan, 1992; Linehan & Heard, 1993; Linehan et al, 2002), but results are inconclusive. What we know thus far is that adding a dialectical behaviour therapy skills-training group to ongoing out-patient individual psychotherapy does not seem to enhance treatment outcomes. Given that dialectical behaviour therapy is described as primarily a skills-training approach (Koerner & Linehan, 1992), this finding indicates that the central skills-training component of dialectical behaviour therapy may not be of primary importance. The policy of not admitting patients to hospital, except for a bare minimum period, has also been challenged, as time spent in a structured in-patient setting has been used to apparently good effect (Bohus et al, 2004). There are also concerns that the rigidity of the programme leads to a greater proportion of people withdrawing from care, and about primary its preoccupation with the treatment of women.

For this reason we do not give dialectical behaviour therapy a greater efficacy rating than other treatments; its benefits are mainly confined to reduction of self-harm episodes.

Therapeutic community treatments

A therapeutic community may be defined as an intensive form of treatment in which the environmental setting becomes the core therapy, where behaviour can be challenged and modified, essentially through group interaction and interpersonal understanding. Although such communities have been in existence in the UK and Denmark for over 50 years, they have only recently been subjected to direct controlled evaluation. The treatments and patient populations treated are so varied that the results are difficult to interpret, but the consensus – mainly advanced by proponents of the treatment – is generally favourable. A systematic review of the literature by Lees et al(1999) concluded that therapeutic communities, particularly the so-called ‘concept’ communities in the USA, were effective, but the positive effects were found primarily in people who were substance misusers living in secure settings in which there was a considerable degree of coercion and no emphasis was placed on the treatment of personality disorder.

The more common version in Europe is the democratic therapeutic community, and no randomised trial has been carried out in this setting. However, Dolan and colleagues at the Henderson Hospital, in a creative attempt to find an appropriate control group, used a non-admitted comparison sample to assess the effectiveness of treatment on core symptoms (Dolan et al, 1997). A total of 137 patients were studied, of whom 70 were admitted and 67 not admitted (either for clinical or for financial reasons). However, this was not a strict comparison group, because less than 1 in 7 of those considered for the Henderson Hospital complete their treatment (Rutter & Tyrer, 2003). There was significantly greater reduction in core features of personality disorder on the Borderline Syndrome Index in the treated group than in the non-admitted group. However, using this index as a primary outcome measure is inadequate because significant state dependent fluctuations of core features of personality disorder may occur over time, and in a UK study this index has been found to lack validity and to be susceptible to distortion from current symptoms (Marlowe et al, 1996).

Further work has suggested that therapeutic community treatment may show cost savings over treatment in general psychiatric services primarily through reducing the need for hospital admission (Davies et al, 1999).

Extended hospital admission theoretically may engender pathological dependency and regression (Linehan, 1987), although there is little evidence for this. In a prospective study of 216 patients with severe personality disorder treated at the Menninger Clinic in the USA for variable lengths of time in two psychoanalytically oriented in-patient units, Gabbard et al(2000) found positive change at discharge and at 1-year follow-up, with no evidence of deleterious effects due to regression and dependency. Nevertheless, the classic paper by Main (1957) should act as a reminder that regression and countertransference may pose considerable difficulties for teams treating patients intensively.

As there are now many more treatments available for the treatment of personality disorder, therapeutic communities need to come into the frame of comparison studies. Further research should be undertaken (Haigh, 2002), but proponents will need to adopt acceptable experimental designs if therapeutic communities are to compete with other treatments.

Conclusions

Despite these somewhat cautionary views, it is important to note that, for the first time in the history of personality disorder, people are regarding the condition as potentially treatable. It is likely that effective treatments are going to be found in the course of the next two decades and, if this turns out to be the case, it will help greatly in removing the pejorative label that currently is attached to personality disorder.

MCQs

1. 1 Assessing efficacy of psychotherapeutic treatment for personality disorder is complicated by:

   1. a low comorbidity with other disorders

   2. b requirement for a long follow-up

   3. c lack of agreement on outcome measures

   4. d the fluctuating nature of this disorder over time

   5. e problems of finding patients.

2. 2 When considering efficacy and generalisability of treatments for personality disorder:

   1. a cognitive–behavioural therapy has the best evidence for both

   2. b dynamic therapy shows no evidence of generalisability

   3. c dialectical behaviour therapy is the treatment of choice

   4. d therapeutic community approaches are easily generalisable

   5. e cognitive analytic therapy shows excellent efficacy in borderline personality disorder.

3. 3 Psychodynamic therapy for personality disorder:

   1. a is too complex to manualise

   2. b may be effective in reducing self-harm

   3. c is commonly associated with a high drop-out rate

   4. d is better than standard psychiatric care

   5. e can only be given by highly trained practitioners.

4. 4 Cognitive therapy for personality disorder:

   1. a is more effective than any other psychotherapy

   2. b is no better than standard psychiatric care in patients who harm themselves

   3. c is simply a watered-down version of dialectical behaviour therapy

   4. d focuses on automatic thoughts

   5. e shows overlap with psychodynamic therapy in its use of the therapeutic relationship.

5. 5 Dialectical behaviour therapy for personality disorder:

   1. a has been shown to be effective in paranoid personality disorder

   2. b requires skills training if it is to be effective

   3. c has widespread acceptance because of frequent replication of good outcomes

   4. d may be used effectively on in-patients

   5. e has limited effect on drug misuse.

MCQ answers

1		2		3		4		5
a	F	a	F	a	F	a	F	a	F
b	T	b	F	b	T	b	T	b	F
c	T	c	F	c	F	c	F	c	F
d	T	d	F	d	T	d	F	d	T
e	F	e	F	e	F	e	T	e	T