Over the past 50 years findings from many experimental studies have established the effectiveness of pharmacological and psychosocial interventions for people with a range of mental health problems. In contrast, the evidence base for interventions for people with personality disorder remains poor. Most research into the impact of such interventions has focused on psychosocial treatment of borderline personality disorder. There is little evidence to guide the management of people with other forms of personality disorder. Even in relation to the borderline type, few high-quality trials have been conducted. A recent systematic review of psychological therapies for people with borderline personality disorder concluded that: ‘studies are too few and too small to inspire full confidence in their results’ (Reference Binks, Fenton and McCarthyBinks et al, 2006). Residential treatments for people with personality disorder have been evaluated (Reference Lees, Manning and RawlingsLees et al, 1999), but such services can inevitably only be offered to – and may only be suitable for – a minority of all those with this disorder.

CONSEQUENCES OF A WEAK EVIDENCE BASE

The relative absence of research into community-based interventions for people with personality disorder is matched by a paucity of services for such people (National Institute for Mental Health in England, 2003). Various reasons have been proposed to explain this state of affairs. It has been argued that interpersonal problems experienced by people with personality disorder make it unrewarding for healthcare professionals to work with this group (Reference HinshelwoodHinshelwood, 1999). Previous surveys have demonstrated professional ambivalence to working with people with personality disorders: in a recent British study examining the views of general practitioners and psychiatrists about which patients should be referred for treatment in secondary care, general practitioners were less likely to state that people with anxiety, depression and most other mental disorders needed to be referred to secondary care than were psychiatrists; in contrast, psychiatrists were less likely than general practitioners to state that people with personality disorders should be referred to secondary care (Reference Walker, Haeney and NaikWalker et al, 2005).

An alternative explanation for the reluctance of psychiatrists and other mental health professionals to work with people with personality disorders is that they feel that they are not equipped to provide these people with satisfactory treatment. With their focus on monitoring mental states, psychotropic medication and powers of compulsory treatment at times of crisis, general mental health services were certainly not designed to meet the needs of people with personality disorders. Although such people may be referred to psychotherapy services, the limited availability of this resource, together with the relative lack of evidence, means that healthcare workers may be reluctant to refer people with a primary diagnosis of personality disorder. Given this context, findings from a randomised trial of social problem-solving therapy for people with personality disorder by Huband et al (Reference Huband, Duggan and Evans2007; this issue) in this month's journal are to be welcomed.

SOCIAL PROBLEM-SOLVING THERAPY

Various attempts have been made to help people improve their social problem-solving skills. Initial studies investigating problem-solving therapy for people who self-harm showed little effect (Reference Gibbons, Butler and UrwinGibbons et al, 1978). Since then, more structured approaches to helping people manage social problems have been developed which synthesise cognitive–behavioural techniques and elements of social skills training. These focus on helping people identify goals and exploring how existing patterns of thinking and behaviour affect the chances of achieving these. Patients are encouraged to develop different approaches to solving problems, to test them out both within and outside of sessions, and to continually review whether the solutions they choose help them achieve their goals (Reference McMurran, Fyfee and McCarthyMcMurran et al, 2001). Results of a non-randomised evaluation of social problem-solving for a group of 52 out-patients with borderline personality disorder in Iowa in the USA demonstrated improved mood and reduced self-harming behaviour over the course of a 20-week programme (Reference Blum, Pfohl and JohnBlum et al, 2002).

In this new study Huband and colleagues randomised people with personality disorder to either three sessions of psychoeducation followed by 16 sessions of group-based social problem-solving therapy, or to a waiting-list control. Two-thirds of those offered the intervention attended at least eight sessions and almost half were still in treatment at 15 weeks. Improvements in self-rated ability to cope with social problems were greater among those offered the intervention than among the waiting-list control group. Active treatment was also associated with a slight improvement in social functioning.

As a pilot study the trial had a range of limitations, such as a relatively short follow-up period, resulting from the limited resources that are usually available for such studies. Interventions delivered in groups are known to result in clustering of outcomes resulting from both therapist factors and group dynamics. Such factors limit the power of studies and should be taken into consideration when analysing the impact of complex interventions such as this (Reference Lee and ThompsonLee & Thompson, 2005). Improved social functioning among those who received social problem-solving therapy is noteworthy; however, the level of improvement was small, equivalent to less than two points on the social functioning scale, and further research is needed to establish whether this intervention results in sustained improvements in social functioning that are clinically as well as statistically significant.

Nonetheless, findings from this trial are important for several reasons. First, it included people with a range of personality disorders: 41% met diagnostic criteria for borderline disorder, 40% for avoidant disorder, and 14% for antisocial disorder. Nearly all previous trials have focused exclusively on those with a primary diagnosis of borderline disorder. This trial, together with other recent studies, is therefore important in highlighting the potential impact of outpatient treatment for people with a range of other personality disorders (Reference Emmelkamp, Benner and KuipersEmmelkamp et al, 2006).

TRAINING IN INTERVENTIONS FOR PERSONALITY DISORDER

What gives the study by Huband et al (Reference Huband, Duggan and Evans2007) special significance is the manner in which active treatment was delivered. Rather than examining the efficacy of an intervention provided by experts in social problem-solving therapy, the study team trained mental health professionals who had no previous experience of delivering this intervention. Mental health workers with experience of working with people with personality disorders were given 2 days of training before the start of the study.

With a community prevalence of approximately 5%, it is clear that even if specialist services for people with personality disorders were greatly expanded, they would not have the capacity to provide services to all those with such disorders. In addition to exploring the effectiveness of social problem-solving therapy, this trial also provides important evidence that brief training for healthcare workers might be sufficient to enable non-specialist staff to deliver psychosocial interventions to people with personality disorders.

Pragmatic studies which evaluate the impact of interventions delivered by nonspecialists may have other advantages as well. It has been a feature of the development of complex interventions ranging from home treatment to cognitive–behavioural therapy for psychosis that large effect sizes found when the interventions are delivered by pioneers tend not to be found when attempts are made to replicate them. By evaluating the impact of social problemsolving therapy delivered by people with no previous experience of this treatment, the study team have generated outcome data that provide a better estimate of the impact the intervention might achieve in a real-world clinical setting (Schoenwald (Reference Schoenwald and HoagwoodSchoenwald & Hoagwood, 2001).

FUTURE RESEARCH

The use of three sessions of psychoeducation prior to the delivery of the main intervention is another noteworthy aspect of this trial. Levels of drop-out from treatment services for people with personality disorder are notoriously high. As with previous studies, these data show that those with the greatest level of personality disturbance are those least likely to engage in treatment. Examining ways to increase retention is therefore important, and the psychoeducational approach used in this study has intuitive appeal. It is not possible to work out whether the use of psychoeducation prior to the delivery of social problem-solving made a difference to drop-out rates in this study, but this hypothesis is amenable to experimental evaluation and should be tested.

Concerns have rightly been expressed about the gap between the evidence base for psychosocial interventions for people with psychosis and the extent of their delivery (Reference RowlandsRowlands, 2004). Financial constraints are the main factor responsible for this ‘therapy gap’, and new psychosocial interventions will need to demonstrate cost-effectiveness if they are to be implemented in clinical practice. This pilot study did not demonstrate statistically significant reductions in service utilisation, but an important trend towards reduced contact with emergency medical services was seen. These findings clearly provide a basis for further investigation of this potentially valuable intervention. Should such studies demonstrate the cost-effectiveness of social problem-solving, they would not only improve the evidence base for treatment of personality disorder, but they might also go some way towards challenging the ambivalence that some healthcare professionals continue to have about working with people with personality disorders.