For most mental state disorders it is relatively easy to choose the period over which efficacy has to be demonstrated. In conditions that develop suddenly (e.g. panic), treatment trials could be for a very short time indeed. For others, particularly when maintenance treatment is being evaluated, at least 6 months may be necessary to establish continued efficacy. In the case of personality disorders, it has been thought that efficacy of treatment could not be judged adequately without at least a 2 to 3 year treatment phase. Personality disorder was regarded as being unlikely to change in the short-term. However, these ideas are changing in the light of evidence from longer-term follow-up studies of patients with personality disorder that show that these conditions do change over time in a consistent and predictable manner with substantial numbers of patients achieving full remission in the longer term.⁽¹⁾ If these longer-term follow-up studies are replicated, it would suggest that therapy should aim to accelerate the process of recovery. Determining what constitutes an adequate amount of therapy and over what length of time is an empirical question. More recent studies have offered treatment over 1 year with a 1 year follow-up
to examine maintenance of effect^(2,3) but other studies have chosen lengthier treatment phases of up to 3 years⁽⁴⁾ with some reporting continued therapy in the follow-up phase which does not allow the effect of maintenance of the original effect to be judged.⁽⁵⁾ More recent studies, examining the effect of psychological treatments, have included a 1 year follow-up. The purpose of this follow-up period is to determine if treatment effects are maintained following the termination of treatment. Such a requirement is not a purist position; if a treatment for personality disorder appears to be effective over a shorter period, this may be due to change in a concurrent (comorbid) condition. In addition, if a treatment is to be judged efficacious in personality disorder its effects should be lasting beyond the active treatment component.

Comorbidity has been defined as ‘the presence of any distinct clinical entity that has existed or that may occur during the clinical course of a patient who has the index disease under study’.⁽⁵⁾ The key word here is ‘distinct’. True comorbidity implies the presence of two completely separate disorders in the same person which are not causally related to each other in any way. Co-occurrence ranges from true comorbidity to the presence of the same disorder in two or more different forms.⁽⁶⁾

Comorbidity is the norm for most personality disorders both with other personality disorders or with mental state disorders. Borderline personality disorder is a major offender in this regard. Only about 1 in 20 of such disorders constitutes the pure condition,⁽⁷⁾ and multiple comorbidity with four or more disorders is common.

In deciding on the efficacy of any treatment for personality disorder it is impossible to be certain whether observed improvement is in the personality disorder or in a comorbid condition. This problem is made worse because personality assessment is allegedly confounded, or ‘contaminated’, by the effect of a concurrent mental state disorder. Thus personality status apparently changes during the presence of a mental state disorder such as depression, only to return to the baseline normal subsequently.^(8,9) Apparent improvement in a personality disorder following a treatment may be entirely due to improvement in a concurrent mental state disorder. However, this conclusion does not mean that personality function, as opposed to personality disorder, does not change. The underlying personality may remain stable, but if the setting and circumstances change, and this includes mental state changes, there can be marked changes in adjustment and so the manifestation of disorder will also change.⁽¹⁰⁾

In view of these problems, a treatment for a personality disorder should ideally be tested in those patients who have that personality disorder only. As these patients are uncommon and atypical, it is difficult to interpret the data from clinical trials.

People with personality disorders do not usually form good relationships with therapists. Although this is in keeping with their problems with relationships elsewhere, it can be a major problem in any form of therapy. The problem is particularly marked with psychotherapy, in which long-held views are challenged by the therapist. The consequence is that many patients dropout of care, and sometimes no amount of therapeutic skill can maintain them in care.⁽¹¹⁾ The failure to maintain prescribed treatment, in whatever form, is a constant handicap in accumulating an evidence base for interventions in personality disorder. Even within personality disorder there are differences between sub-groups. Most therapeutic trials have been inpatients with borderline personality disorder while those with schizoid, paranoid, histrionic, narcissistic, and antisocial personality disorders appear much less frequently. This is probably related to treatment attitudes. Borderline, anxious, and avoidant personality disorders contain a much higher proportion of treatment seeking (Type S) personality disorders as opposed to treatment resisting (Type R) ones, which are most prominent in those with antisocial, schizoid, and paranoid personalities.⁽¹²⁾

Any study of personality disorder is likely to have a large proportion of dropouts and this complicates the interpretation of the effects of treatment. The exception is when patients are treated in restricted settings such as prisons and other closed facilities,⁽¹³⁾ but as these circumstances are abnormal it is difficult to generalize from them.

The choice of outcome measures is a problem in the assessment of all psychiatric disorders, but difficulty is particularly great in studies of personality disorders. These disorders affect both the individual and society, and a range of outcomes can be measured to cover these possibilities. Forensic psychiatrists and the general public usually consider that the outcome of mentally disordered offenders is best measured by the frequency of reoffending. This is an easily measured and reliable statistic, but it does not necessarily record symptomatic or personality change, and may be distorted by a range of other factors (e.g. patients who spend a long time in hospital or prison are not likely to reoffend). Changes in symptoms also have limited use since they may be a consequence of changes in mental state disorders quite independent of personality. Repeated measures of personality status also have disadvantages since, as noted earlier, they may be affected by changes with concurrent mental state disorder. Personality also changes with ageing irrespective of treatment.⁽¹⁴⁾

Because of these difficulties, global outcome measures are often used to determine the degree of improvement in personality disorders in long-term follow-up studies, although a battery of measurements is normally used in short-term treatment studies. Unfortunately, there is no standardized set of measures of global outcome. It is reasonable to take into account symptomatic change, social functioning, quality of life, incidents of societal conflict (e.g. police contacts), and reports from informants. Even these may not correctly reflect change in personality status. Thus a person whose personality disorder does not change in any basic way may find an environmental niche in which the personality disturbance does not manifest itself as conflict. Such a person would show improvement on all the items listed above, but the improvement would be a consequence of environmental change, not of personality alteration.

The evidence base for effectiveness of treatment in personality disorders is also exacting:

1 The treatment should be effective when used in the pure form of the personality disorder (in an explanatory trial) and subsequently in other forms of the disorder in which comorbidity is more common (pragmatic trial).

2 Efficacy cannot be established satisfactorily unless the treatment is tested in a randomized controlled trial.

3 A suitable control treatment or management needs to be tested against the experimental treatment.

4 Efficacy should only be determined after a period of at least 1 year because of the long duration of personality disorder. More recent randomized controlled trials in borderline personality disorder examining psychological therapies have met these more demanding criteria.^(15,16)

Cognitive analytical therapy combines cognitive and analytical approaches and has been applied to the treatment of personality disorders, particularly the borderline group.⁽¹⁷⁾ The clinical manifestations of this condition are postulated to be a set of partially dissociated ‘self-states’ which account for the clinical features of this disorder. Such patients typically describe rapid switching from one state of mind to another, experiencing intense uncontrollable emotions or alternatively feeling muddled, or emotionally cut-off. Such ‘dissociative states’ (different from the conditions of similar name formerly linked to hysteria) are said to be activated by severe external threats and to be maintained by repetitions of threat and reinforcement by memories or situations which are similar to the original source of threat.

Cognitive analytical therapy is concerned with the identification of these different self-states and helping patients to identify ‘reciprocal role procedures’, or patterns of relationships which are learned in early childhood and are relatively resistant to change.⁽¹⁸⁾ The patient is taught to observe and try to change damaging patterns of thinking and behaviour which relate to these self-states and to become more self-aware. The therapist’s role is to gather information about the patient’s experience of relationships and the different states he or she experiences, including interpretation where necessary. Although the standard measure of evidence of effectiveness, the randomized controlled trial, has still not yet been reported for this treatment it has gathered an impressive group of adherents and has become widely used and now has a good theoretical and pragmatic base.⁽¹⁸⁾

In its original form, cognitive therapy for depression was used to help the patient to identify and modify dysfunctional thoughts and beliefs through the use of specific cognitive techniques such as Socratic questioning. The focus of therapy was here and now the aim was to return the patient to his or her usual functioning by relieving current symptoms. The cognitive model of personality disorder does emphasize cognitive, emotional, and behavioural factors but the origins of personality problems are regarded as originating in the temperament of the child, childhood development, and experiences. Early infant attachment patterns, the child’s internal working model of relationships, self-identity, self-worth, and the emotional availability of the infant’s caregivers are central to how the child develops and these shape the adult self-identity, interpersonal relationships in adulthood, and behavioural and emotional coping responses.⁽¹⁹⁾

One of the first tasks of cognitive therapy in personality disorders is to gain an historical account of the patient’s childhood development and background from which the therapist can derive a cognitive formulation linking past difficulties and presenting problems. Through the formulation, and understanding of the patient’s view of self and others, unique core beliefs are identified that are linked to affect and to overdeveloped behavioural patterns that prevent the individual from functioning in an adaptive manner, particularly in interpersonal contexts. Therapy focuses on beliefs that concern core concepts about the self and others that have developed from childhood onwards and associated behaviours that have developed as coping strategies. The content and meaning of the beliefs have had an impact on past and present relationships and are likely to impact the therapeutic relationship. These beliefs, formed through negative, possibly abusive and neglectful experiences with others, are likely to have resulted in low self-esteem, hypersensitivity to criticism, and poor relationships with peers, caregivers and others in adolescence. Once a clear understanding of the content of patient’s core beliefs and associated overdeveloped or compensatory behavioural patterns has been established, patients are encouraged to test out their beliefs and assumptions about others by learning new, more adaptive strategies for relating to others and to themselves. In borderline personality disorder, typically patients hold beliefs such as ‘I am a bad and inadequate person’ and ‘others will abandon or reject me’. Having formed these beliefs through experiences in childhood, borderline patients, for example, may have learnt to avoid close relationships, are highly sensitized to signs of disapproval in others and have developed a punitive, self-critical style of thinking and behaviour, including self-harm. The emphasis in cognitive therapy is in developing new ways of thinking about self and others and in testing out new ways of behaving that are less self-defeating and more likely to improve the patient’s interpersonal skills.⁽¹⁹⁾ In comparison with the treatment of Axis I disorders, cognitive therapy with personality-disordered individuals takes more sessions and spans a longer time because the underlying problems are more pervasive and ingrained. There are other important elements of cognitive and related therapies, of which schema therapy and dialectical behaviour therapy are the most prominent. One of these differences is on the emphasis and attention paid to the therapeutic relationship. In cognitive therapy for personality disorder, more emphasis is placed on establishing and maintaining a therapeutic alliance, as interpersonal difficulties which occur in the patient’s life outside therapy are also likely to arise within therapy. This is based on the hypothesis that the patient’s core beliefs are consistent across a wide range of settings and therefore are also likely to be manifest in therapeutic relationships. Patients are therefore likely to be highly sensitive to signs of criticism and approval in their therapists. The models of treatment for personality disorder proposed by Beck and Freeman,⁽²⁰⁾ Davidson,⁽¹⁹⁾ Young,⁽²¹⁾ and Linehan⁽²²⁾ have in common an attempt to integrate biological and psychosocial factors. All models of treatment recognize the importance of building a secure therapeutic relationship and transference and countertransference issues in therapy are increasingly recognized as important mediators of the therapeutic process. These therapies utilize cognitive techniques to repair breakdown in communication that can occur during therapy.
Cognitive analytical therapy⁽¹⁸⁾ also gives special attention to this aspect of therapy.

One of the goals of cognitive therapy with personality-disordered patients is to take advantage of these interpersonal difficulties in treatment by identifying and modifying the beliefs underlying them and, by extension, other relationships. Although people with personality disorders can recognize difficulties, they experience the problems as egosyntonic⁽²³⁾ (i.e. accepted as normal because they are an intrinsic part of usual functioning). As a result, alternative and potentially more adaptive beliefs about the self and others need to be identified and evaluated to see if they are indeed more adaptive and embraced as a consequence. These alternative more adaptive beliefs require to be systematically reviewed and reinforced, and new behaviours and ways of relating to others need to be practiced repeatedly if changes are to be consolidated. To achieve these changes, the therapist usually has to adopt a more directive approach than in cognitive therapy for depression and other Axis I disorders, and throughout will be more concerned with identifying and overcoming cognitive, emotional, and behavioural avoidance which maintains core beliefs.