Difficulties In Assessing Outcome Research

The following section summarises the main obstacles to assessing the outcome data in this field identified by Bateman & Fonagy (2000).

Assessment and Identification of Cases

Since the late 1980s a number of semi-structured interviews and self-report measures have been developed (see, for example, Clark 1993; Hyler, 1994; Hyler et al., 1988; Loranger et al., 1987; Millon et al., 1994; Spitzer et al. 1990; Tyrer et al., 1988) but there is little agreement between measures and not all of them are comprehensive. Some of the usual problems - such as respondents supplying socially desirable answers - arise with these standardised measures and reliability also depends on respondents having adequate insight into their longstanding difficulties. Another problem, linked to the question of comorbidity discussed below, is that the behaviours indicative of personality disorder may fluctuate in the presence/absence of Axis I disorders, obscuring the contribution of the personality traits (Klein, 1993). Accurate diagnosis is therefore assisted by assessing or re-assessing individuals once the Axis I disorder has been adequately treated. Zimmerman (1994) has reviewed the issues that arise in identifying personality disorder and suggests that reliability of diagnosis can be increased by using informants in addition to self-report/interview.

Compounding the difficulty of assessment is the dilemma of classification. Within the literature, subjects are most frequently designated by the ICD or DSM categories yet there is poor cross-classificatory reliability (Sara, Raven & Mann, 1996) and a lack of empirical evidence to support the clustering of categories used in DSM-IV (Mulder & Joyce, 1997). Some studies have instead grouped subjects according to a theoretical approach to personality disorder - for example, Beck, Freeman & Davis (2004) (cognitive) and Kernberg (1984) (psychoanalytic), but this prevents meaningful comparison of results across models. Dimensional approaches have so far been equally problematic in outcome research because investigators disagree about which traits to measure and whether or not these are on a continuum with normal personality characteristics.


In clinical practice it is rare to meet an individual who only meets ICD-10/DSM-IV criteria for a single personality disorder. As noted above, many individuals meet criteria for more than one personality disorder and others meet criteria for both Axis I and Axis II disorders (Oldham et al., 1995). This makes it difficult to decipher improvements observed clinically or reported in the literature. For example, a person with BPD and major depression may become more impulsive and destructive as the depression lifts (appearing to become more 'borderline') or this behaviour may diminish with the anergia of depression, creating the impression of improved impulse control. Despite the confounding effect of comorbidity when measuring outcome, published studies have generally failed to report this clearly.

Outcome Measures

Measurement of outcome is inherently problematic without reliable case identification. Most recent studies of personality disorder do not attempt to measure change in the syndrome itself but instead measure change in symptoms, behaviour and social adjustment. Studies of personality disordered offenders usually rely on reconviction rates as a measure of outcome although researchers in this field identify obvious problems with this approach (Dolan & Coid, 1993). A few investigators have attempted to measure change in the syndromes more directly by assessing whether or not individuals meet fewer diagnostic criteria over time (Dolan et al, 1997; Monsen, Odland & Eilertsen, 1995; Stevenson & Meares, 1992) but this approach does not prevent the confounding of personality change and symptomatic improvement in coexisting Axis I disorders. Bateman & Tyrer (2002) go so far as to argue that 'any measured change in personality should be regarded in the first instance as an artefact related to improvement or deterioration in mental state.' These authors observe that there is no standardised method for recording global outcome measures in long-term follow-up studies and note that some individuals may appear to improve over time purely as a consequence of changing their circumstances so that their personality disorder no longer generates interpersonal conflict. Finally, it is worth considering that categorical diagnoses of personality disorder permit individuals to change status from cases to non-cases without any greater personality change than those who remain cases throughout a study because their personality disturbance is more significant at baseline assessment (Tyrer et al., 1997).

Other Research Issues

In addition to the usual challenges of conducting randomised controlled trials to assess treatment efficacy (Roth & Fonagy, 2005), the main difficulty facing researchers who are interested in personality disorder is the length of time required to conduct an adequate treatment trial, with the associated problems of cost, increased risk of drop-outs and confounding effects of other treatment received during the study period. Dialectical-behaviour therapy, for example, was developed for treatment of BPD and 'Stage 1' was designed to be delivered intensively over a one-year period. Bateman & Fonagy (1999) evaluated a psychoanalytically informed day hospital programme for treatment of BPD that ran for 18 months. These study periods are considerably longer than the norm of three to four months in most drug trials or RCTs comparing psychological therapies for Axis I disorders. Furthermore, long-term follow-up beyond the active treatment phase is particularly important for adequate evaluation of efficacy in personality disorder (H0glend, 1993).

With these caveats in mind we will proceed to examine the evidence base for psychological therapies in the treatment of personality disorder. Recent systematic reviews (Bateman & Fonagy, 2000; Bateman & Tyrer, 2002; Binks et al., 2006; Perry, Banon & Ianni, 1999; Roth & Fonagy, 2005; Roy & Tyrer, 2001; Shea, 1993) have varied in their inclusiveness. For reasons of space the following section is limited to consideration of large-cohort and controlled studies in which patients were selected on the basis of Axis II disorders, treatments were clearly described and adequate measures were used. The approaches considered are dynamic psychotherapy, cognitive therapy, interpersonal group psychotherapy, behaviour therapy and dialectical behaviour therapy delivered through outpatient, day hospital or inpatient programmes. Therapeutic communities will be briefly considered at the end, together with treatment programmes for personality disordered offenders.

Personal Power Perfection

Personal Power Perfection

Get All The Support And Guidance You Need To Be A Success At Your Personal Power. This Book Is One Of The Most Valuable Resources In The World When It Comes To Loving Yourself Enough to Protect Your Personal Power.

Get My Free Ebook

Post a comment