Psychotherapy for Borderline Personality Disorder: The Evidence

Psychotherapy for borderline personality disorder (BPD) is the core treatment, and dialectical behavior therapy (DBT) has the largest evidence base, but no single approach has proven better than the rest^[6][2]. A 2020 Cochrane review pooled 75 randomized controlled trials with 4,507 participants, most of them testing DBT or mentalization-based treatment (MBT)^[1].

In December 2024, the American Psychiatric Association (APA) updated its practice guideline and gave a 1B recommendation to "a structured approach to psychotherapy that has support in the literature and targets the core features of the disorder"^[2]. Announcing the guideline, its chair, George Keepers, said several structured psychotherapies were effective and "no therapy emerged as a ‘gold standard’"^[2].

The borderline personality disorder therapies with randomized-trial support include DBT, MBT, schema therapy, transference-focused psychotherapy (TFP) and generalist models such as general psychiatric management (GPM)^[6][7]. This article covers how well each BPD treatment therapy works. Prevalence and demographic data are on the BPD condition page.

Compared with treatment as usual, psychotherapy reduced BPD symptom severity by an SMD of −0.52 across 22 trials and 1,244 participants, rated moderate-quality evidence^[1]. That equals about 3.6 points on the 36-point Zanarini Rating Scale for BPD, above the 3-point threshold the authors set for a clinically relevant change^[1].

Self-harm (SMD −0.32, 13 trials), suicide-related outcomes (SMD −0.34, 13 trials) and psychosocial functioning (SMD −0.45, 22 trials) also improved more with psychotherapy than with usual care^[1]. None of those three results reached the authors’ threshold for a clinically meaningful difference, and all rested on low-quality evidence^[1].

Psychotherapy did not reduce dropout compared with usual care, and adverse effects were unclear because too few trials reported them^[1]. Treatment in the included trials lasted anywhere from 1 to 36 months^[1].

An earlier meta-analysis in JAMA Psychiatry reached a similar but more cautious conclusion. Across 33 trials and 2,256 participants, psychotherapies beat control conditions on combined borderline outcomes (g = 0.32 in stand-alone designs)^[8]. The authors judged the effects small, "inflated by risk of bias and publication bias, and particularly unstable at follow-up"^[8].

DBT was developed by Marsha Linehan and starts from the premise that intense emotional reactivity and unhelpful ways of regulating emotion are central to BPD^[6]. It pairs cognitive, behavioral and problem-solving strategies with acceptance strategies, taking a "dialectical stance" that balances the need for change against the need for acceptance^[6].

DBT was the subject of about one-third of all trials in the 2020 Cochrane review^[1]. Compared with usual care, it reduced BPD severity (SMD −0.60, 3 trials) and self-harm (SMD −0.28, 7 trials) and improved psychosocial functioning (SMD −0.36, 6 trials), all on low-quality evidence^[1]. Data on self-harm in the wider population are on the non-suicidal self-injury page.

The best-known test of DBT compared it with expert therapists rather than usual care. In a 2006 trial, 101 women with recent suicidal and self-injurious behavior received either a year of DBT or a year of community treatment by nonbehavioral psychotherapy experts, followed by a year of follow-up^[3].

Women receiving DBT were half as likely to make a suicide attempt and had fewer psychiatric hospitalizations and psychiatric emergency visits^[3]. Dropout from the first therapist was 25% in DBT versus 59% with the community experts^[3]. The authors concluded that DBT’s results could not be explained only by the general benefits of seeing an expert therapist^[3].

Mentalization-based treatment (MBT)

MBT, developed by Anthony Bateman and Peter Fonagy, is rooted in a psychoanalytic model and trains people to read their own and others’ behavior in terms of mental states such as beliefs and intentions^[6]. Compared with usual care, MBT reduced self-harm (risk ratio 0.62, 3 trials) and suicidality (risk ratio 0.10, 3 trials) in the Cochrane review, though both findings rest on low-quality evidence^[1].

Schema therapy

Schema therapy, developed by Jeffrey Young, is an integrative therapy in a cognitive framework that targets "schema modes," patterns of thinking, feeling and behaving theorized to stem from early experiences in which core emotional needs went unmet^[6]. Its relationship-based technique is called limited reparenting^[6]. In a 2023 network meta-analysis, schema therapy ranked first for reducing BPD severity, but the authors cautioned that the result rested on very few trials^[6].

Transference-focused psychotherapy (TFP)

TFP, developed by Otto Kernberg, adapts psychodynamic psychotherapy for BPD^[6]. It works on "split" all-positive or all-negative views of self and others as they show up in the relationship with the therapist, and focuses on the here and now rather than the past^[6].

General psychiatric management (GPM)

GPM combines psychodynamically informed therapy with symptom-targeted medication management drawn from APA guideline recommendations^[7]. In a 2009 trial, 180 patients with at least two suicidal or self-injurious episodes in the previous 5 years received a year of either DBT or GPM, and both groups improved, with no significant differences between them on any outcome^[7].

Head-to-head trials are scarce. A 2023 network meta-analysis of 43 trials and 3,273 adults found DBT, MBT and schema therapy more effective than usual care and generic treatments, but concluded that "no single treatment seems to be the best choice"^[6]. DBT had the most robust base, with 19 of the 43 trials^[6].

Some therapies roughly halved the combined risk of suicide attempts and suicide deaths, with risk ratios around 0.5, but those differences were not statistically significant against usual care or other therapies^[6].

In the 2017 JAMA Psychiatry analysis, DBT (g = 0.34) and psychodynamic approaches (g = 0.41) were the only therapy types that beat control interventions^[8]. The advantage faded in well-designed trials and when the control treatment was also manualized^[8].

That pattern explains why the APA guideline calls for a structured, evidence-supported approach rather than naming one brand of therapy^[2].

Length is a known barrier: the authors of the FASTER trial noted that evidence-based BPD psychotherapies are lengthy, which limits access^[4]. That trial randomized 240 suicidal or self-harming patients in Canada to 6 or 12 months of comprehensive DBT and followed them to 24 months^[4].

Six months of DBT was noninferior to 12 months for total self-harm, and the shorter course produced faster reductions in BPD symptoms^[4]. For suicide attempts alone, noninferiority held at months 6 and 12 but not at month 24^[4]. Dropout rates did not differ between the two groups^[4].

Dropout across BPD therapy is concentrated early. A survival meta-analysis of 111 studies and 9,100 adults found dropout was highest in the first quarter of treatment^[9]. Schema therapy had the lowest dropout, individual therapy had lower dropout than group therapy, and community treatment by experts had the highest^[9].

Access also depends on the supply of trained clinicians and on price. Related data are in the therapist shortage statistics and cost of therapy statistics.

No medication has shown a reliable effect on the core features of BPD. A 2022 Cochrane review of 46 randomized trials with 2,769 participants, covering 29 different medications, found no difference from placebo on any primary outcome at the end of treatment for any drug^[5].

The review also noted that medication for BPD is off-label, yet prescription rates are high and people with BPD often take several psychotropic drugs at once for long periods^[5].

The APA guideline suggests (2C) that any medication be time-limited, aimed at a specific measurable symptom and adjunctive to psychotherapy, and recommends a medication review at least every 6 months^[2]. Its chair said no evidence was found for any drug’s effectiveness on core BPD symptoms^[2].

Long-term data show that BPD symptoms often remit, while full recovery is less common. The McLean Study of Adult Development enrolled 290 psychiatric inpatients with BPD and reassessed them every two years for 16 years^[10].

By the 16-year follow-up, 78% to 99% had achieved symptomatic remission, depending on how long the remission had to last^[10]. Only 40% to 60% achieved recovery, defined as remission plus good social and vocational functioning, and 10% to 36% had a recurrence of symptoms after remission^[10].

The study was observational, so it describes the course of BPD over time rather than the effect of any one therapy^[10]. A companion article reviews the evidence on psychotherapy for histrionic personality disorder.