Flaws in the Design of a Randomized Controlled Trial Comparing Short-Term versus Long-Term Mentalization-Based Therapy for Borderline Personality Disorder

Dear Editor,

In the article authored by Juul et al. [1] and recently published in Psychotherapy and Psychosomatics, the authors conducted a randomized trial comparing two treatment lengths of mentalization-based therapy (MBT) for borderline personality disorder (BPD). However, there are some shortcomings in the trial design and the drawn conclusions that deserve further consideration.

The treatment the authors call “long-term MBT” had substantially shorter duration and frequency than standard MBT. In the MBT manual for BPD [2], the treatment is scheduled to last 18 months with individual therapy weekly and group therapy weekly, a format which was used in the largest and most well-conducted trial of MBT for BPD [3]. However, “long-term MBT” in the study by Juul et al. [1] lasted 13.5 months with individual therapy every second week and group therapy weekly. Furthermore, patients were absent from many therapy sessions. In “long-term MBT,” patients attended in average 17 individual sessions and 24 group sessions, while patients in short-term MBT attended in average 12 individual sessions and 14 group sessions. In the Bateman and Fonagy [3] study, the mean number of clinical meetings in MBT was 92 (SD = 38). Thus, it would be more appropriate to name the treatments by their actual content, i.e., “6-month MBT versus 13-month MBT.”

The within-group effects of the two treatments, especially those of “long-term MBT,” are relatively low. Using data from Table 2 to calculate Cohen’s d for within-group effects (using SD pre as denominator), we found the following effect sizes for “long-term MBT” versus short-term MBT: ZAN-BPD d = 0.40 versus d = 0.70; WSAS d = 0.63 versus 0.82, SF-36 d = 0.36 versus 0.35; GAF d = 0.53 versus d = 0.70. There are substantial differences to the within-group effects of MBT for BPD in the large trial by Bateman and Fonagy [3], for example, d = 2.34 for GAF, and to a study with 3-year MBT with d = 4.60 for GAF [4]. The relatively small within-group effect sizes and high dropout rate (35%) suggest that the 13-month MBT condition was not adequately implemented as an efficacious treatment.

From the result of no significant differences between treatments, Juul et al. [1] conclude that administrators and clinicians “…may give serious consideration to short-term MBT as a desirable option for outpatients with BPD.” However, Juul et al. [1] did not design the study as a non-inferiority trial, as McMain et al. [5] previously did for DBT, but as a superiority trial. From a nonsignificant difference between treatments in outcome of a superiority trial, non-inferiority or equivalence of short-term MBT to long-term MBT cannot be concluded [6].

As noted by the authors, results are not representative for full-criteria BPD patients. Including patients with subthreshold BPD may have artificially enhanced the efficacy of short-term MBT.

To sum up, we are concerned about the flawed design of this trial, especially about the insufficient implementation of “long-term” MBT, and also the authors’ questionable conclusions. This article could guide healthcare administrators to make unfortunate decisions about future supply of treatments for BPD, reducing the possibilities for BPD patients to receive adequate therapy.