What to conclude from a non-randomized clinical trial comparing dialectical behavior therapy and mentalization-based treatment in patients with borderline personality disorder?
Bibliographic record
Abstract
The study by Barnicot and Crawford (Barnicot and Crawford, 2018) comparing clinical outcomes of Dialectical Behavior Therapy (DBT) and Mentalization-Based Treatment (MBT) in patients with Borderline Personality Disorder (BPD) in the context of a non-randomized study in the United Kingdom represents a major step forward in identifying effective treatments for BPD patients.Indeed, there is a lack of direct head-to-head comparisons of current evidencebased treatments of patients with BPD (Cristea et al., 2017;Fonagy et al., 2017).This study therefore provides important information concerning the relative effectiveness of both types of treatment, particularly because it was conducted in routine clinical care, increasing the ecological validity of its findings.Given the paucity of comparative studies, the use of appropriate analysis strategies and correct reporting of clinical trials in this area is all the more important.However, several problems with the data analysis and reporting make it unclear what conclusions can be drawn from this trial for both research and clinical practice.Barnicot and Crawford highlight in the abstract, results and discussion section of their paper that 'reductions in self-harm and improvements in emotional regulation at 12 months were greater amongst those receiving DBT than amongst those receiving MBT' (p.1), suggesting this was the major finding of their study.However, this conclusion seems not to be supported by the data.As noted by the authors themselves in the results section, their study found 'no differences between participants receiving DBT and those receiving MBT in number of incidents of selfharm, BPD severity, emotional dysregulation, relationships with others or dissociation' (p.4).Hence, no significant differences were found on any of the clinical outcome measures in this study.Still, in the abstract of their paper and in the discussion section the authors argue that reductions in self-harm and improvements in emotional regulation were greater in DBT.This erroneous conclusion appears to be based on the finding that in adjusted multilevel models there was a steeper decline in self-harm and emotional dysregulation in DBT compared to MBT.Yet, there is a clear difference between the rate of change during treatment and outcomes at the endpoint of a study.If there were no differences at the study endpoint, but there were differences in the rate of change, then patients simply followed different trajectories toward the same endpoint.If valid, these findings may have implications for clinical practice, even when DBT and MBT are equally effective at the study endpoint, as it would suggest that self-harm and associated features may improve faster in DBT.Furthermore, Barnicot and Crawford adjusted for baseline differences in several clinical variables between patients in MBT and DBT, of which some were significant, and others were not.Thus, the basis for their selection of potential confounding variables is not clear.Moreover, it is well-known that if covariates overlap with the experimental effect, adjusting for these variables does not balance out these differences, as is often wrongly assumed, but may instead obscure treatment effects if the covariates show a significant correlation with the outcome (Field, 2013).The authors did not report whether there was an association between any of these variables and outcomes.These considerations are relevant to the authors' analyses of several outcome measures.DBT showed significantly higher drop-out rates, hospitalization and emergency department attendance at 12-month follow-up.These differences disappeared after including covariates in the analyses.Whether this result is valid or not is not clear due to the problem of including covariates described above.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.014 | 0.116 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.030 | 0.014 |
| Insufficient payload (model declined to judge) | 0.008 | 0.006 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".