Table 2 Cognitive and behavioural and schema therapies

Cognitive and behavioural treatment vs inactive/non-specialist comparators

 RCT (n = 18 including 4 pilot) [99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116]

Sample size: 20–100 (n = 12); > 100 (n = 6).

Date: 1990–1999 (n = 2); 2000–2009 (n = 7); 2010–2019 (n = 9).

Country: Europe (n = 4); North America (n = 6); Oceania (n = 1); UK (n = 7).

Diagnoses: “BPD” or other “personality disorder” diagnoses/criteria (n = 13); mixed clinical diagnoses including “personality disorder” diagnosis (n = 1); “BPD” diagnosis/criteria and recent or previous (repeated) self-harm (n = 3); recent and previous self-harm (n = 1). Demographics: 100% female (n = 4); 0–49% White (n = 1), 80–99% White (n = 5); 100% White (n = 5).

RCTs with primary outcomes: On the primary outcomes of RCTs, compared to controls, a greater proportion of participants receiving cognitive and behavioural therapies recovered on symptoms (1/1) and also showed improvement in “personality disorder” symptoms (3/3), symptom severity (1/2), and social functioning (1/2), but not depressive (0/1) or (social) anxiety symptoms (0/1), service use (0/1), or frequency/number of participants with self-harming/suicidal behaviour (0/4). On non-primary outcomes, compared to controls, participants receiving cognitive and behavioural therapies showed improvement in approximately half of the outcomes.

 Non-randomised experiments, observational studies, quasi experiment, and natural experiment with pre-post comparison (n = 8) [117,118,119,120,121,122,123,124]

Sample size: < 20 (n = 1); 20–100 (n = 7).

Date: 1990–1999 (n = 1); 2000–2009 (n = 2); 2010–2019 (n = 5).

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

Diagnoses: “BPD” (n = 3) or avoidant “personality disorder” diagnosis (n = 1); “BPD” diagnosis/criteria, mood disorder, history of self-harm, with our without emotional and behavioural dysregulation (n = 3); childhood sexual abuse (n = 1).

Demographics: 100% female (n = 1); 50–79% White (n = 1); 80–99% White (n = 1).

In studies with comparisons over time only, participants improved on the one reported primary outcome (self-harm: 1/1) and most secondary outcomes.

 Uncontrolled intervention development studies and single case study with multiple measures (n = 11) [125,126,127,128,129,130,131,132,133,134,135]

Sample size: < 20 (n = 8); 20–100 (n = 3).

Date: 2000–2009 (n = 2); 2010–2019 (n = 9).

Country: Asia (n = 2); Europe (n = 3); North America (n = 1); Oceania (n = 1); UK (n = 4).

Diagnoses: “BPD” (n = 4) or other “personality disorder” diagnoses (n = 4); “BPD” diagnosis/features and comorbid mood disorder (n = 2) or drug/alcohol disorder (n = 1).

Demographics: 80–99% White (n = 1); 100% White (n = 1); older age (n = 1).

In studies with comparisons over time only, participants improved over time on the primary outcomes symptoms/distress (2/2) and quality of life (1/1), and also showed no dropouts (1/1). Participants improved on secondary outcomes. Patients with a current substance misuse disorder showed a reduction in drug use (1/1). Elderly patients with a chronic mood or adjustment disorder improved in symptom distress (1/1) and some but not all aspects of schema and coping variables (1/1).

Cognitive and behavioural treatment vs specialist comparators

 RCT (n = 4) [136,137,138,139]

Sample size: 20–100 (n = 4).

Date: 2000–2009 (n = 3), 2010–2019 (n = 1).

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

Diagnoses: “BPD” features/diagnosis (n = 2) or other “personality disorder” diagnosis (n = 2)

Demographics: 100% White (n = 1).

RCTs with primary outcomes: In 3 RCTs comparing cognitive and behavioural therapy with specialist comparators, there were no between-group differences on primary outcomes (symptom improvement: 0/1; symptoms severity: 0/1; interpersonal problems: 0/1) or secondary outcomes. In 1 RCT, significantly more participants receiving Schema Focused Therapy (SFT) recovered on the primary outcome (“BPD” symptoms: 1/1) as well as on three secondary outcomes compared to cognitive therapy.

 Non-randomised experiments, observational studies, quasi experiment, and natural experiment with pre-post comparison (n = 3) [140,141,142]

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

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

Country: Europe (n = 3).

Diagnoses: “Personality disorder” diagnosis (n = 2); cluster B “personality disorder” diagnosis with comorbid Axis I disorder (n = 1).

Demographics: no data report.

In 3 non-randomised experiments, there were no differences between the cognitive behavioural treatment and specialist comparators on primary outcomes (personality functioning: 0/1; symptom severity: 0/1) and all or most secondary outcomes (0/3).

 Uncontrolled intervention development studies and single case study with multiple measures (n = 1) [143]

Sample size: < 20 (n = 1).

Date: 2010–2019 (n = 1).

Country: North America (n = 1).

Diagnoses: NSSI disorder (n = 1). Demographics: 50–79% White (n = 1).

One study with comparisons over time only did not report significant results for outcomes on patients with NSSI disorder. However, 8/10 participants reported meaningful reductions in self-harming behaviour.

Studies of modified cognitive and behavioural treatments

 RCT (n = 2 including 1 pilot) [144, 145]

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

Date: 1990–1999 (n = 1); 2010–2019 (n = 1).

Country: North America (n = 1); UK (n = 1).

Diagnoses: “BPD” diagnosis (n = 1); previous suicide attempts, antidepressants taken as part of an overdose, and suicidal behaviour (n = 1).

Demographics: 80–99% White (n = 1).

RCTs with primary outcomes: On the primary outcome of 1 RCT, findings for differences between the cognitive Behavioural Problem Solving and TAU group on suicidality were mixed (0/1). Findings were mixed or showed no between-group differences for non-primary outcomes (0/2).

 Non-randomised experiments, observational studies, quasi experiment, and natural experiment with pre-post comparison (n = 1) [146]

Sample size: 20–100 (n = 1).

Date: 2000–2009 (n = 1).

Country: Europe (n = 1).

Diagnoses: “Personality disorder” diagnosis, excluding borderline, schizotypal, schizoid, antisocial, or NOS “personality disorder” diagnoses (n = 1). Demographics: no data reported.

The 1 study utilised a crossover design and showed significant improvements over the treatment period as a whole, but no between-group differences.

Studies of adapted cognitive and behavioural treatments

 Uncontrolled intervention development studies and single case study with multiple measures (n = 1) [147]

Sample size: < 20 (n = 1).

Date: 2010–2019 (n = 1).

Country: Oceania (n = 1).

Diagnoses: “Personality disorder” diagnosis (n = 1).

Demographics: no data reported.

In 1 study with comparisons over time only, no statistical analysis was conducted. However, 5/8 patients no longer met criteria for an avoidant “personality disorder” at end of follow-up.

Schema therapy vs inactive/non-specialist comparators

 RCT (n = 1) [148]

Sample size: > 100 (n = 1).

Date: 2010–2019 (n = 1).

Country: Europe (n = 1).

Diagnoses: Avoidant, dependent, obsessive–compulsive, paranoid, histrionic, or narcissistic “personality disorder” diagnosis (n = 1).

Demographics: no data reported.

On the primary outcome of the 1 RCT, compared to controls, a greater proportion of participants receiving schema therapy recovered (1/1). Compared to controls, participants also improved on some non-primary outcomes.

 Uncontrolled intervention development studies and single case study with multiple measures (n = 4) [149,150,151,152]

Sample size: < 20 (n = 4).

Date: 2000–2009 (n = 1); 2010–2019 (n = 3).

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

Diagnoses: “BPD” (n = 3) or other “personality disorder” diagnosis (n = 1).

Demographics: 100% female (n = 3); old age (n = 1).

In the 1 study that reported significant results with comparisons over time only, participants improved on “BPD” symptoms (1/1) and most other outcomes.

Studies of modified schema therapy

 RCT (n = 1) [153]

Sample size: 20–100 (n = 1).

Date: 2000–2009 (n = 1).

Country: Europe (n = 1).

Diagnoses: “BPD” diagnosis (n = 1). Demographics: 80–99% (n = 1).

RCTs with primary outcomes: On the primary outcome of 1 RCT, there was no difference between participants receiving schema therapy with and those without phone support on recovery from “BPD” (0/1). There was also no significant difference on non-primary outcomes (0/1).