Does DBT Work for Suicidal Behavior? What the Randomized Trials Found, With Sources

If you’re thinking about suicide right now, or you’re worried about someone who is, call or text 988, or chat at 988lifeline.org. It’s free, it’s confidential, and someone is there 24 hours a day. Text and chat are available in Spanish, and there are options for people who are Deaf or hard of hearing.

People ask us this in different words. Does therapy do anything when someone keeps thinking about ending their life? Is DBT worth the time? What’s the proof? The honest way to answer is to go to the trials, so that’s what this page does. Dialectical Behavior Therapy has been tested in randomized controlled trials since 1991, and the main ones are listed below with what each one measured and what it found.

The short answers

What is DBT, and who was it built for?

DBT is a form of cognitive behavioral therapy that Marsha Linehan developed at the University of Washington for women who were repeatedly harming themselves and attempting suicide. The first randomized trial, published in Archives of General Psychiatry in December 1991, ran for a year and compared DBT with treatment as usual in the community. The women who got DBT had fewer episodes of self-harm and suicidal behavior, the episodes were less medically severe, they were more likely to stay in individual therapy, and they spent fewer days in psychiatric inpatient care.

That starting point matters when you read the rest of the evidence. Most therapies were designed for depression or anxiety and later tried on people at risk of suicide. DBT began with suicidal behavior as the outcome it was meant to change.

Does DBT reduce suicide attempts?

Yes, in the trial designed to test exactly that. Linehan’s team published a two-year randomized controlled trial in 2006 with 101 women who had recent suicidal and self-injurious behavior and met criteria for borderline personality disorder. One year of treatment, one year of follow-up.

The comparison arm is what sets this trial apart. The other half of the women were treated by experienced non-behavioral therapists chosen for their reputation, and the two groups of therapists were matched on sex, availability, expertise, belief in their own approach, training, experience, access to consultation, and the standing of their institution. The point was to find out whether DBT beat good therapy, not whether it beat no therapy.

Over the two years, the women who received DBT were half as likely to make a suicide attempt. They were less likely to be hospitalized for suicidal ideation, had fewer psychiatric hospitalizations and fewer psychiatric emergency department visits, and were less likely to drop out of treatment. The authors wrote that the results “cannot reasonably be attributed to general factors associated with expert psychotherapy.”

What does the whole body of trials show?

The same direction as the 2006 trial. A 2019 meta-analysis in Behavior Therapy pooled 18 controlled trials of DBT. Against the comparison treatments, DBT reduced self-directed violence, meaning suicide attempts and self-harm counted together, and reduced the use of psychiatric crisis services.

The 2020 Cochrane review took a wider view. It pooled 75 randomized trials with 4,507 participants across more than 16 psychological therapies for borderline personality disorder, and DBT accounted for about a third of the trials, more than any other treatment. Compared with treatment as usual, DBT reduced the severity of the disorder, reduced self-harm, and improved psychosocial functioning. On the count of trials alone, DBT has more controlled evidence behind it for this problem than any other therapy.

DBT has been tested in younger people as well. A 2018 trial in JAMA Psychiatry randomized 173 people aged 12 to 18 with a recent suicide attempt and repeated self-harm across four academic medical centers. After six months, 90.3 percent of the DBT group had made no suicide attempt against 78.9 percent in the comparison therapy, and 75.6 percent completed treatment against 55.2 percent. The authors called DBT the first well-established, empirically supported treatment for reducing repeated suicide attempts and self-harm in youths. Our program is for adults, and that trial is included here because it is part of the evidence for the treatment itself.

What is Comprehensive DBT, and does the full program matter?

Comprehensive DBT, also called standard DBT, is the version that was tested in the trials above. It has four parts that run at the same time: weekly individual therapy, a weekly skills training group, phone coaching between sessions, and a weekly consultation team for the therapists.

Linehan’s team tested whether all four parts are needed. Their 2015 component trial randomized 99 women at high risk of suicide to standard DBT, to DBT skills training with case management, or to DBT individual therapy with an activities group in place of skills training. All three produced similar reductions in suicide attempts, with every therapist using the DBT protocol for assessing and managing suicide risk. The versions that included skills training did better on non-suicidal self-injury and on depression. Standard DBT had the lowest dropout rate, 24 percent against 48 percent for individual therapy alone, and in the follow-up year 3 percent of the standard DBT group used the emergency department for a crisis and 3 percent were hospitalized, against 13 percent and 13 percent for individual therapy alone.

The same trial tracked suicidal thoughts, and they came down too: across the treatment year, suicidal ideation fell in every DBT condition, alongside the fall in attempts and in the use of crisis services. Other adult trials report the same. In a 2016 randomized trial at a veterans’ medical center, 91 veterans at high risk of suicide, 61 of them men, were assigned to six months of standard DBT or to usual care, and suicidal ideation, depression and anxiety improved over the six months of DBT. In a 2012 randomized trial of 63 suicidal college students aged 18 to 25, DBT produced greater decreases in suicidality, depression and self-injury than an optimized comparison treatment, and most of those gains were still there at the 18-month follow-up.

Staying in treatment is most of the job for a person at risk, and the 2015 trial is the clearest evidence that the full program does it better than a part of it. What “the full program” looks like in practice, and how to tell it apart from DBT-informed therapy, is the subject of the fourth post in this series.

How strong is this evidence?

Strong for a psychotherapy. The two Linehan trials enrolled women only, so the direct evidence on men comes from the wider trial pool rather than from those two studies, and most of the DBT trials were run with people who met criteria for borderline personality disorder, because that is the group the treatment was developed for.

Behind those trials sit 35 years of randomized trials, a comparison against expert therapists that DBT won, and a meta-analysis and a Cochrane review that point the same way. For a treatment whose target is suicidal behavior, that is the largest body of controlled evidence there is.

Treatment for adults in Texas

North Texas DBT Associates provides Comprehensive Dialectical Behavior Therapy for adults, by telehealth across Texas. Our program has the four parts the trials tested: weekly individual therapy, a weekly skills group, 24/7 phone coaching, and clinicians who sit on a weekly DBT consultation team. Courtney Hicks is a DBT-Linehan Board of Certification Certified Clinician.

You can read about how Comprehensive DBT works, see the program, meet our clinicians, check where we work, and read our fees, including the sliding scale.

Contact North Texas DBT Associates, call (817) 214-7200, or email info@ntxdbt.com. If it’s urgent, call or text 988.

The rest of this series

Three companion posts, written for Suicide Prevention Month 2026:

Sources:

Every source above was opened and checked on 8 September 2026, apart from the 1991, 2012 and 2016 trials, which were added on 23 and 24 September, and each figure was checked against its published abstract on 23 September 2026.