Skip to main content

Transforming the understanding
and treatment of mental illnesses.

Celebrating 75 Years! Learn More >>

 Archived Content

The National Institute of Mental Health archives materials that are over 4 years old and no longer being updated. The content on this page is provided for historical reference purposes only and may not reflect current knowledge or information.

Rethinking BPD: A Clinician’s View

Dr. Marsha Linehan was featured in a series of lectures on Borderline Personality Disorder.


Marsha Linehan: So I want to start something only because I am afraid that I'll talk too long and won't get to tell you this but it fits this anyway so I am going to tell a little story just to start, which is this fabulous patient that I had for many years. She had been in and out of hospitals forever and was in New York at a hospital for two years and they sent her to me on an ambulance plane of all things and she came and I said okay I'd see her as long as she was in a half-way house in Seattle because she was very suicidal and many other things and running down streets in night gowns with knives and being found behind dumpsters and alleys almost dead. So I treated her for quite some time and finally we got her through college, which was a big thing. This was more than my usual research one-year treatment, but she actually got through college and she moved and got a job in San Francisco, sorry Los Angeles as a teacher and would call me because she wanted to teach some of the skills to her students. I love her so much and so I was down not too long ago.

She calls me every once in a while and we talk and so I'd gone down and we were having lunch and she was telling me about how awful her life because she also was not only Borderline Personality Disorder but she was also schizoaffective and so she was saying how hard it was to be psychotic and teach and that it was problematic and so she had to take a leapand was tutoring now. I was so sad that I was in tears because it was just so sad this had happened to her and she said "Marsha don't be so upset, this is a lot better than being borderline was" and I think that says a lot. You know that being psychotic schizoaffective beats out Borderline Personality Disorder.

So I'm going to talk a little bit about, people have talked about my treatment dialectical behavioral therapy just to give you, you know someone said to me, Marsha your job is to give us hope so I'm going to give you hope because I'm not alone anymore, I was for years. The only person doing treatment research on Borderline Personality Disorder and most of my life the only person doing it on suicidal behavior, but now we have managed to get other people out there doing data, there is mentalization treatment, all these young people are doing things, so there is a lot of hope, but I know this the best so I am going to talk about. Dialectical behavioral therapy, where it started and tell you what I ran into. So I started, I always wanted to work with suicide and so I decided that I was going to work with suicide at NIMH and those days helped young investigators enormously. They talked to you and you would run your research by with them. I really feel to be honest with you; NIMH is like the co-author of most of what I've done. They have been so helpful all the time. So I wanted to deal with high-risk for suicide with multiple suicide attempts and self-injury and I called all of the hospitals in Seattle and I said I wanted the very worst that you got because I wanted to be sure that if my people got better that I would be able to tell the difference between them and their control condition and so I got a little grant to do treatment development for suicidal behavior and I just learned behavior therapy and I was totally sure it would cure everybody and I had not the slightest doubt, but I ran into a bunch of problems.

My first problem was that the people I was treating had extreme sensitivity to rejection and invalidation and that made a change focused treatment, which definitely behavioral therapy in 1980 was untenable so I switched to an acceptance based treatment. That was a disaster too though because of the extreme suffering made an acceptance approach untenable. The patient would say, "well you're not hearing my suffering? You're not going to help me change?" And I'd say "no no I'll help you change, I'll help you change" and they'd say, "You're saying I'm wrong? You're saying that?" And I'd say "no I'm not saying that!" So this is sort of how it went. So I came up with a approach really by accident just working with all these people, which ultimately got called dialectical but dialectal is walking the mental path in many ways, where it's a synthesis of change strategies, this is on the part of the therapist, the change on the part of the therapist and acceptance strategies on the part of therapist, where that's my job.

The next problem though was I was dealing with people who had an extraordinary low distress tolerance and it made focusing on any one problem area or one part of a problem or one disorder, as you know they have multiple disorders or one therapy topic impossible with frequent crisis overtaking any ability for sustained work on change. So we are constantly every topic, you go through ten topics because of avoidance and inability to tolerate. So the solution in the end was to develop an approach that taught the patients an effect acceptance skills which included within there, distress tolerance and there is a whole teaching I do, module, called distress tolerance. How to tolerate crises and not make them worse but also how to radically accept the life that may not be the life you want but that doesn't mean you have to miserable your whole life and change skills, which all came from behavior therapy. I stole every one of them.

But then I had another problem, which was the ever changing clinical presentation together with frequent crises resulted in all the therapists being confused and chaotic and therapy got very chaotic. You've got to remember this was back in the 80's when all the behavioral treatments were protocol based treatments you do this the first session, this the second session, this the third session, that was all of them. There was only one treatment manual, literally one and that was the psychodynamically oriented treatment manual and that didn't do that and so I modeled after that. I had patients who are meeting criteria both for loads of Axis I disorders but also Axis II. Although I didn't know that they were meeting criteria for that stuff at that time because remember I was focusing on suicide.

So my solution was to develop an approach to a treatment that combined protocol-based interventions, in other words all of our patients get skills training, which is this week, you get the skill that we are teaching this week and next week is the one we are teaching next week, combined that with a target-based agenda, where the target-based said we are going to organize ourselves and do this problem first, this problem second, this problem and life threatening behavior of course was first. Life threatening at that time was suicide but since BPD is in so many prisons you have to look at homicide, also as a life threatening behavior. So that brought those two things together. But then the next problem I had was that therapists got emotional dysregulated treating these patients and that led to excessive fear, excessive anger, hostility, the therapists would get angry at the patients. They tried to control the patients. You know it's really frightening to have a person that could be dead at any moment and that you can't keep them from being dead. So the therapists would then try to get in control, they would reject the patient or attack the patients. This is group one of therapists and group two therapists were the ones who had excessive empathy and they would fall into the pool of despair with the client and just abandon therapy altogether and everyone would sort of cry together.

So I had to solve that and the way I solved that was by defining a treatment team as part of the treatment. I didn't really create a treatment team; I defined a treatment team as a part of the therapy because I am going to say something to you may not be aware of, I don't think many people are. To my knowledge there has never been one single randomized trial that has ever been conducted without a team as part of the therapy with the role of the team is to keep all the therapists doing the treatment that they are studying in that clinical trial.

As far as I know, there are two things that are always done assessment of outcomes and sum group, you either have one supervisor or team that keeps you doing the therapy and then we wonder why when you take a therapy to a community it does not work. Why? Because we don't take the entire therapy to the community. We take all kinds of things but we leave the team that keeps you doing the therapy and assessment out of what's translated so I just said to myself, fine I'll just redefine therapy and I will say that if we did a team and the research we'll call the team part of therapy since I developed the therapy you can't say you're doing the therapy if you don't have a team and so that's essentially how that was dealt with and the role of the team is to treat the therapist and to keep the therapist in the model. Do I believe that it's important all the time? Possibly not. Do I believe it's important with high-risk, out of control, difficult to treat, multi-diagnostic suicidal patients? Yes I do, therapists get burned out.

The next problem to solve was that I wanted to get another grant to study what I was doing but I couldn't get grant unless I had a mental disorder and I was doing suicide and I didn't have a mental disorder so I figured okay, someone told me at a review actually that I was studying Borderline Personality Disorder. My entire manual was written without the word Borderline Personality Disorder in it even once because I had never heard of the disorder. So then I heard of the disorder and I had to choose Borderline Personality Disorder or Major Depression. So since most the people meet the criteria for borderline, I said I would do borderline and my advocate at the NIMH told me, I'll never forget it because I said okay I will go with borderline and not depression. He said, "Marsha you are making the biggest mistake of your entire career." I'll never forget it. I said, "Well I'm doing it anyway." It turned out not to be a huge mistake but that was obviously a problem. Once I did that I had to have a model of the disorder and there wasn't any model that behaviors could live with because of course I had found out at that point that the only people who thought there was Borderline Personality Disorder were psychoanalysts who of course good behaviors do not read psychoanalytical stuff and when I read it did not make a lot of sense because the model, the theory was so completely divergent from any theory that I could deal with or what I think of as behavioral science so I, but you do have to have a theory to do treatment. I had a theory of suicide but now I needed a theory of borderline and I needed one capable of, I had three criteria only three, capable of guiding effective therapy, has to be non-pejorative, engendering compassion, has to be compatible with current research data so my theory does all of that because when the data changes, I just change the theory.

The basic model which I fought all my career for this model but I think I'm really winning the data but I'm not going to give you data on it because it's not part of this talk but I could. Borderline Personality Disorder is a pervasive regulation disorder basically and it's a pervasive disorder of emotion regulation system where you look at an emotion as a system not as feelings but as a total system that is a full person response but it's an emotional response and therefore it has action as part of thought and physiology etc. and that Borderline Personality Disorder criterion behaviors if you actually look at them and really study them, function to regulate emotions, a number of them suicidal behavior for example, is extraordinarily effective at regulating emotion and the average person, I kid you not actually believes that if they kill themselves, that they will not feel as bad when they're dead. I pointed that out to my clients there is no evidence of that particular point of you. Or suicidal behaviors have natural consequences of emotional dysregulation. You cannot have good relationships if you are not regulated, if you hate a person one day and love them the next, want to go to the movie with them one night, don't want to go the next etc. etc. In other words, relationships, a sense of self most of the criteria that are in there, have to do with the fundamentals of emotion. So the solution was to development a treatment of pathogenesis. This I have to say is a biological regulation disorder; I don't have the slightest doubt that there is a biological basis to this disorder and I don't believe you can create a Borderline Personality Disorder without the biology. I could be wrong on that but that's my experience in just watching all this time.

Combined with what I call invalidating the social environment, now whether there are other characteristics of the environment besides that, invalidating covers a lot of territory so far I haven't had to change the theory but that's because I change the definition of invalidation as needed, but my real point is we need more research on this, we are very very limited on our research here. But it's much like the Patterson, like other research on the development of other disorders where it's the same and that therefore the disorder is transactional. Transactional and dialectical are incompatible with blame just to be clear because it is that person A creates person B, but person B creates person A so it's not like the family creates the disorder, the disorder creates the family and the family creates the disorder, not the family but the environment. So you have this transactional overtime, creation of a disorder based on both characteristics not necessarily the family by the way. So what did I add to standard kinds of behavioral therapy, I'm just going to go because I just looked at the time and I was supposed to be finished in a minute. Synthesis of acceptance to change, just remember I started way back, most of this stuff is in there now but it wasn't that. Principle-based integration of evidence based treatments, I'm a big believer in bringing protocols into a principle-based approach as suppose to developing new therapies all the time. Focus on in-session behaviors, stages targeting by severity and threat. BPD is organized in stages of treatment based on control. Suicide risk and assessment protocol, there is reason to believe now that this actually maybe controlling a lot of our findings much to my disbelief. Skills-based evidence based treatments, in other words I stole all of my skills from evidence based treatments and mindfulness. Definition of team as a part of therapy and therapists self-disclosure. So what do the clinical trials say? No one really disagrees right now that this treatment is effective. So that's not the fight at the moment, that was the fight for years but that's no longer the fight everybody agrees the treatment works. The question now is why it works? We have nine randomized control trials on Borderline Personality Disorder. DBT is going so far away from borderline now to other disorders that it's unclear that it's actually a treatment for borderline as it is a treatment for substantial dysregulation.

For low-level, non-serious stuff, you would never need DBT. There are plenty of good treatments for that. DBT's real strength is when you have really serious dysregulation, DBT works then, or high-risk, difficult to treat, multiple diagnosis, multiple stuff. That's when DBT sort of comes into its own, otherwise it becomes something else. Our data is that we do suicide attempts, self-injury, depression, hopelessness, anger, substance dependence, impulsiveness. Now the psycho analytic people fought for years with me saying that only treated symptoms, behaviors of course don't even use the construct but nonetheless. So I used one of the psychoanalytic measures and made them promise that they would stop saying that if I got the findings and I did. So it also changes the introject, which is the psychoanalytic construct. This is just the suicidal behavior plus self-injury, the yellow is the people who didn't get DBT. There has been a lot internal validity controls in these studies, this is the percent attempting suicide. What I want to talk about for a moment is what matters. In other words, do you need the whole treatment? The standard treatment is individual DBT, plus skills training, plus phone calls to the individual therapist for behavioral coaching, plus a team. All the research I am going to talk to you about now, every single one of them have a team and they have coaching of some sort but not necessarily with the individual therapist. The first thing I want to, oh oops. Okay what happened was the slides didn't get put on here, I am going to tell you our data. This a fabulous study, which NIMH funded. This is very important. What we did was we did a study looking at DBT Standard Dialectical Behavioral therapy; sort of the ordinary one and we took highly suicidal people with a suicide attempt in the last year and self-injury or suicide attempt and the eight before they came in and Borderline Personality Disorder.

Most of the meet between 8 and 9 criterion. Okay most of them are working etc. so we did that. One-third of them went into standard behavioral therapy, the same stuff that I tell everyone they have to do to say they are doing DBT. The other one went in to a treatment where we wanted to see are skills training important? Okay so we gave one group the entire treatment except they didn't get skills training, we gave them a psycho education activity's group like you get in communal health, so that's group two. We'll call that the individual therapy only group because they didn't get skills, the no skills group and then we took another group and we wanted to say, is the individual DBT, the way its run is that important? Maybe we can get rid of that and just give people skills training, so we did but because people were so highly suicidal, I thought I had to have somebody who could deal with crisis so we gave them intensive case management Washington State case management Treatment Manual, this is important to get that part. Washington State was manualized and we had a supervisor and they had a team. So they had skills training, intensive case management but the intensive case managers only took calls during office hours, this big ticket item but we gave a really good suicide, crisis management plan to the crisis clinic in Seattle, which is one of those phones lines that every town has. Now the other thing that happened in the study was I very worried, I just thought I can't let people die just because I want to do a study, so I gave the University of Washington, which is really DBT actually, risk assessment and management protocol, which is a treatment known designed to get people to want to treat suicidal people by giving you something that says after a session you check off whether you did it. It makes you verify, do you have risk and should you think of hospitalization but the way its set up is that it tells you every reason not to hospitalize, which you can check off to say I didn't do it and I'd give you your reasons, I didn't do something else, and I'd give you your reasons.

So the whole is sort of set up to get therapists not to hospitalize and to do top of the line risk management, mainly because I didn't think the treatment was working. If I‘d known what I know now, I might not have done this. So what happened? One, all the treatments were effective for reducing suicidal behavior and they were equally effective suggesting more than likely that the commonality across is what? Team and risk management protocol and some characteristics of having someone who can handle a crisis okay, sort of crisis management and it was top of the line. However, when you look at mental health outcome something different happened, this is the biggest one of my bigger life shocks is what counts is the skills. Has a big effect on reducing depression, big effect on reducing anxiety and the individual the therapy is useful for that, not useless but not as good. So in other words, since I was certain standard therapy would be the best, it never occurred to me to think anything else but the only place that its good is that it has half the drop out. So if you want to keep people in therapy it's great, but why keep them in if you can get them better without keeping them in, that's the way I look at it, so we're still analyzing the data. The outcome of this is two things: one the skills probably are critical and you have heard this from your colleagues over here who said teach me how to live and that's exactly what it's designed for. They are designing it to go into eighth grade now. We've got the skills translated down from five to twelve year olds, there is kindergarten mindfulness training now, we run friends and family skills training, so I think I underestimated this enormous value of skills training. Now the other thing that we do in DBT is we take them off psych what medication and we keep them out of the hospitals and so it's possible really the treatment that we have for suicidal behavior has been reducing hospitalizations. You know there isn't a single shred of evidence that any hospitalizations ever kept anybody alive five extra minutes, but we all hospitalize and all the treatment is usually therapists are terrified not to hospitalized because they are afraid of being sued.

So we have this unbelievable fear in our therapists and when you can get therapists secure, which I obviously have been able to do, you can get that down. Now I'm not saying standard might not be better in some ways, just looking at it during follow-up maybe you know it's coming back alive again, but the point I'm making to you is without rigorous research that focuses on component analysis and tearing treatments apart, we're not going to get anywhere and horse race therapy isn't going to help us. Managed care should help; insurance companies should do it, but science has to figure out how to improve and make the treatments better. I have a lot here on training but I'm going to skip it because I think my time is up and I don't want to go over, but we're now looking at skills training as being used as treating resistant depression, we have studies on eating disorders or skills training so it's going widely across and so we need to look at that because even in our study with borderline drug addicts, what we found was that DBT was a little bit better on drugs but I don't know if it's worth the money on drugs, but it's really better on depression and anxiety. So let me tell you one thing, one last little thing. Here is where we're not good, just to be clear in case you thought we were good at everything, DBT standard DBT is really good at most disorders except it's terrible at anxiety disorders unless you target anxiety disorders specifically and DBT does that but only after you get suicidal behavior down. So we're now at our lab, Melanie Harned, a research scientist working me and Martin , were collaborating and working on bringing prolonged exposure into DBT and the data we have so far suggests that we can move way up and deal with anxiety disorders and so that says also that we got to have these principle-based treatments and the research on how to bring a module in for this. So thank you, thank you very much.