he study focussed on comparing three manualised therapies for treating people with borderline personality disorder (BPD). The authors reported using a randomised-controlled trial design but did not control some variables as tightly as might be expected. For example, therapy was conducted in the community, medication usage varied considerably across participants and there were between-group differences in the number of therapy sessions attended (which, according to the authors, increased the ecological validity of the study). Although an a priori hypothesis was not stated explicitly, the design resembled a non-inferiority trial comparing treatments to a “gold standard” (dialectical behaviour therapy; DBT).
Over a five-year period, 90 adult participants with BPD were recruited from routine clinical practice (98 were excluded for reasons such as current substance dependence). Participants were randomly assigned to one of three, year-long outpatient therapies (which served as the experimental groups), conducted by 19 clinically-experienced therapists. According to the authors, one group received DBT; an approach focussed on learning emotion-regulation techniques in a validating therapeutic setting; another group received transference-focussed psychotherapy (TFP), an approach based on object-relations theory; and the final group received supportive treatment which involved therapists offering emotional support and working with transference but not offering interpretations (unlike in TFP).
The authors reported their primary outcomes were suicidality (measured using the Overt Aggression Scale-Modified), aggression (measured using the Anger, Irritability and Assault Questionnaire) and impulsivity (measured using the Barratt Impulsivity Scale-II). Secondary outcomes were anxiety, depression and social adjustment. Data was collected at baseline and at 4 months, at 8 months and at the end of therapy (at 12 months).
Data was analysed using growth-curve analysis and hierarchical linear modelling, primarily to examine the change within each group over time. Based on the results, the authors concluded that all treatments resulted in broadly-equivalent, significant positive change but that some treatments were differentially effective on certain outcomes. For example, TFP was associated with significant improvements in anger, irritability and assault, whereas BPD was not. Analysis also indicated that attrition and the consistency of medication did not have a significant impact on the results.