Table 1 DBT

DBT vs inactive/non-specialist comparators

 RCT (n = 12) [33,34,35,36,37,38,39,40,41,42,43,44]

Sample size: 20–100 (n = 12). Date: 1990–1999 (n = 2); 2000–2009 (n = 4); 2010–2019 (n = 6).

Country: Asia (n = 1); Europe (n = 4); North America (n = 4); Oceania (n = 1); UK (n = 2).

Diagnoses: “BPD” diagnosis (n = 10); “BPD” or “personality disorder” diagnosis and self-harm (n = 2).

Demographics: 100% female (n = 5); 50–79% (n = 2) White.

RCTs with primary outcomes: On the primary outcomes of RCTs, compared to inactive/non-specialist controls, participants receiving DBT showed improvement in self-harm in 2/3 studies that examined self-harm (in 1 study this was only the case for clinician-rated self-harm), symptoms at discharge (1/1), global distress (0/1), hospital admissions (0/1), and “BPD” symptoms (1/1).

On non-primary outcomes, compared to controls, participants receiving DBT showed improvement on approximately half of the outcomes. For some of the outcomes, differences were no longer significant at follow-up compared to post-treatment.

 Non-randomised experiments, observational studies, quasi experiment, and natural experiment with pre-post comparison (n = 15) [45,46,47,48,49,50,51,52,53,54,55,56,57,58,59]

Sample size: < 20 (n = 2); 20–100 (n = 13).

Date: 2000–2009 (n = 7); 2010–2019 (n = 8).

Country: Europe (n = 5); North America (n = 8); UK (n = 1); Oceania (n = 1).

Diagnoses: “BPD”, “personality disorder”, or emotionally unstable “personality disorder” diagnosis (n = 11); “BPD” diagnosis and self-harm or a comorbid diagnosis (n = 3); severe impairment and history of suicide attempts or crisis service use (n = 1).

Demographics: 100% female (n = 5); 50–79% White (n = 2), 80–99% White (n = 5).

Non-randomised experiments: In 1 study with two control groups, the DBT group was superior compared to TAU on the primary outcome “BPD” symptoms (1/1).

In studies with comparisons over time only, participants improved on the one reported primary outcome and on close to all secondary outcomes. Studies focusing on patients with comorbid severe mental illness, substance dependence, or an extensive history of suicide attempts or crisis service use all showed improvement in above-mentioned outcomes.

 Uncontrolled intervention development studies and single case study with multiple measures (n = 3) [60,61,62]

Sample size: < 20 (n = 1); 20–100 (n = 2).

Date: 2000–2009 (n = 1); 2010–2019 (n = 2).

Country: North America (n = 3).

Diagnoses: “Personality disorder” or “BPD” diagnosis (n = 2); severe impairment and history of suicide attempt or crisis service use (n = 1). Demographics: 80–99% White (n = 2); 100% White (n = 1).

Studies with comparisons over time only: In 1 study with a primary outcome, participants with severe impairment and an extensive history of suicide attempts or crisis service use improved on the primary outcome, but this was no longer significant one year later. Across studies, participants also improved on all secondary/other outcomes.

 Implementation studies (n = 1) [63]

Sample size: > 100 (n = 1).

Date: 2020 – (n = 1).

Country: Europe (n = 1).

Diagnoses: “BPD” or “emotionally unstable personality disorder” diagnosis (n = 1).

Demographics: no data reported.

In 1 study with comparison over time only, participants improved over time on all outcomes (1/1).

DBT vs specialist comparators

 RCT (N = 6) [64,65,66,67,68,69]

Sample size: 20–100 (n = 3); > 100 (n = 3).

Date: 2000–2009 (n = 4); 2010–2019 (n = 2).

Country: North America (n = 5); Oceania (n = 1).

Diagnoses: “BPD” diagnosis (n = 2); “BPD” diagnosis and self-harm (n = 3); “BPD” diagnosis and opiate dependence diagnosis (n = 1). Demographics: 100% female (n = 1); 50–79% White (n = 2), 80–99% White (n = 1).

RCTs with primary outcomes: DBT was not different or inferior to the specialist comparator in 2 RCTs (including a follow-up study) on the primary outcomes (suicidal episodes (0/2)). In 1 RCT, DBT was superior to the specialist comparator on the primary outcome (suicide attempts (1/1)). In 1 RCT no direct comparisons were made between specialist comparators. Across RCTs, compared to controls, DBT did not show improvement on non-primary outcomes, except for service use (1/1).

 Non-randomised experiments, observational studies, quasi experiment, and natural experiment with pre-post comparison (n = 2) [70, 71]

Sample size: 20–100 (n = 2).

Date: 2010–2019 (n = 2).

Country: Europe (n = 1), UK (n = 1).

Diagnoses: “BPD” diagnosis (n = 2). Demographics: 50–79% White (n = 1).

Non-randomised experiments without primary outcomes: Participants in the DBT group showed no improvement compared to controls in the MBT group on any outcome. Participants in the combined DBT group showed no improvement compared to controls in the individual DBT group on outcomes.

Studies of partial/modified DBT

 RCT (n = 6 including 1 pilot) [72,73,74,75,76,77]

Sample size: 20–100 (n = 6).

Date: 2000–2009 (n = 1); 2010–2019 (n = 5).

Country: Asia (n = 1); Europe (n = 3); North America (n = 2).

Diagnoses: “BPD” diagnosis (n = 4); “BPD” diagnosis and self-harm (n = 2).

Demographics: 100% female (n = 6); 50–79% White (n = 2); 100% White (n = 2).

RCTs without primary outcomes: In 3 RCTs, there was no difference between the adapted versions of DBT on all or most outcomes. In 3 other RCTs, compared to controls, participations receiving (adapted) DBT showed improvements on all or most outcomes.

 Non-randomised experiments, observational studies, quasi experiment, and natural experiment with pre-post comparison (n = 10) [78,79,80,81,82,83,84,85,86,87]

Sample size: < 20 (n = 2); 20–100 (n = 6); > 100 (n = 2).

Date: 2000–2009 (n = 4); 2010–2019 (n = 2); 2020 – (n = 4).

Country: Europe (n = 2); North America (n = 3); Oceania (n = 3); Republic of Ireland and Northern Ireland (n = 1); UK (n = 1).

Diagnoses: “BPD” diagnosis and/or experiencing emotional dysregulation (n = 8); self-harm (n = 1); “BPD” and self-harm (n = 1) Demographics: 100% female (n = 2); 80–99% White (n = 1).

In 1 non-randomised experiment, compared to standard DBT, participants in the DBT skills training group showed no improvement in “BPD” symptoms, symptom severity, and suicidality (0/1). In studies with comparison over time only, participants improved on the primary outcome service use (1/1) and most secondary outcomes.

 Uncontrolled intervention development studies and single case study with multiple measures (n = 3) [88,89,90]

Sample size: < 20 (n = 2); 20–100 (n = 1).

Date: 2000–2009 (n = 1); 2010–2019 (n = 2).

Country: Europe (n = 1); Oceania (n = 2).

Diagnoses: “BPD” or cluster B diagnosis (n = 3).

Demographics: no data reported.

In 3 studies with comparisons over time only, participants improved on all secondary/other outcomes, except of anxiety symptoms (0/1).

Studies of adapted DBT

 RCT (n = 3 including 1 pilot) [91,92,93]

Sample size: 20–100 (n = 2); > 100 (n = 1).

Date: 2010–2019 (n = 2); 2020- (n = 1).

Country: Asia (n = 1); Europe (n = 1); North America (n = 1).

Diagnoses: “BPD” diagnosis (n = 1); “BPD” diagnosis/criteria and PTSD diagnosis (n = 2).

Demographics: 100% female (n = 2); 80–99% White (n = 1); 100% male, 18–50-year-olds and married (n = 1).

RCTs with primary outcomes: In 1 RCT, compared to Cognitive Processing Therapy (CPT), participants with comorbid PTSD receiving DBT-PTSD showed improvement on primary outcomes (diagnostic and symptom remission of PTSD: 1/1) and secondary outcomes. In 1 RCT focusing on married men, compared to waitlist controls, participants receiving Couple-DBT showed improvement in all outcomes. One RCT did not report significance results.

 Non-randomised experiments, observational studies, quasi experiment, and natural experiment with pre-post comparison (n = 3) [94,95,96]

Sample size: 20–100 (n = 2); > 100 (n = 1).

Date: 2010–2019 (n = 3).

Country: Europe (n = 2); Oceania (n = 1).

Diagnoses: “BPD” diagnosis (n = 2); “BPD” and eating disorder diagnosis (n = 1).

Demographics: 100% female (n = 2); only 18–25-year-olds (n = 1); only primary caregivers of child younger than 3-years-old (n = 1).

Non-randomised experiments: In 1 study, compared to CBT, participants with a comorbid eating disorder receiving DBT showed improvement on some primary outcomes and most secondary outcomes. In 1 study, compared to the general DBT group, participants in the young adult only DBT group showed improvement in non-primary symptom outcomes. In 1 study with comparisons over time only, participants, i.e. caregivers of young children, improved on all outcomes.

 Uncontrolled intervention development studies and single case study with multiple measures (n = 2) [97, 98]

Sample size: < 20 (n = 1); 20–100 (n = 1).

Date: 2010–2019 (n = 2).

Country: Europe (n = 2).

Diagnoses: “BPD” diagnosis or criteria (n = 2).

Demographics: 100% female (n = 2); only 18–25-year-olds (n = 1).

In studies with comparisons over time only, participants improved over time on all outcomes.