TCPR: Is borderline personality disorder (BPD) underdiagnosed or overdiagnosed?
Dr. Zimmerman: I think underdiagnosis is the bigger problem. We interviewed 500 patients who presented to a private outpatient clinic with a general, unstructured interview. We also interviewed a separate sample of 409 patients with a structured interview: the Structured Interview for DSM-IV Personality (SIDP-IV). With the structured interview, which is the gold standard, the rate of BPD was 14.4%, but it was only 0.4% with the unstructured interview (Zimmerman M and Mattia JI, Am J Psychiatry 1999;156(10):1570–1574). Other groups have come up with similar findings.

TCPR: What gets in the way of recognizing BPD?
Dr. Zimmerman: One reason is that mood, anxiety, and substance use disorders are common in patients with BPD, and they usually present with symptoms associated with those disorders as their chief complaint. They don’t come in and say, “I’m here because I have real problems with abandonment” or “I don’t know who I am.” Another problem is that there is no obvious gate. The DSM-5 lists nine criteria for BPD, but none are identified as required in the way that “sudden attacks of anxiety” is for panic disorder.

TCPR: How do you screen for BPD?
Dr. Zimmerman: I ask patients about affective instability. Among the DSM-5 criteria, this one had the highest sensitivity for diagnosing BPD in our sample of over 3,500 outpatients who were evaluated with a semi-structured diagnostic interview, and other studies back this up. Not only does this item have a sensitivity of over 90%, but it has a negative predictive value of 99%, which means if it’s absent you don’t have to worry about the person having the disorder (Zimmerman M et al, Br J Psychiatry 2017;210(2):165–166). It’s more sensitive than self-harm, which only occurs in about one-half to two-thirds of individuals with BPD.

TCPR: Is self-harm not as reliable of a criterion?
Dr. Zimmerman: Correct. Depending on the setting, about one-half to two-thirds of individuals with BPD have a history of self-harm. It can also lead to overdiagnosis. About half of patients who self-harm do not have a diagnosis of BPD.

TCPR: How do you ask about affective instability?
Dr. Zimmerman: I ask, “Do you often have days where your mood changes, meaning you’ll go from depressed to angry to happy all in the same day? How typical is that for you? Have others commented that your mood changes a lot—that you’re often irritable and that you have mood swings?” I include that in all psychiatric assessments, and if they answer yes, I ask about the other eight criteria for BPD.

“The foundation of treatment for bipolar disorder is medication. For borderline personality disorder, it is psychotherapy.”

Mark Zimmerman, MD 

TCPR: How do people with bipolar disorder (BD) respond to that question?
Dr. Zimmerman: We actually published a study of this screening in a mood disorder sample, and the psychometrics were very similar to what we found in a general psychiatric sample (Zimmerman M et al, J Clin Psychiatry 2019;80(1):18m12257).

TCPR: So it helps identify people with mood disorders who also have BPD?
Dr. Zimmerman: Yes. About 20% of people with BD have BPD, and 20% of people with BPD have bipolar (more often bipolar II).

TCPR: For the rest who only have one diagnosis, how do you distinguish BD from BPD?
Dr. Zimmerman: There is some overlap in the features, but this is a superficial resemblance. Both have “mood swings,” but in BPD these are time-limited reactions to how they perceive other people are treating them. In BD, mood symptoms may be stress-related, but they are more sustained and accompanied by other characteristic symptoms. Impulsivity is common in both, but in BD impulsivity is out of character and occurs along with other manic symptoms. The best way to tell the disorders apart is simply to apply the diagnostic criteria of each. Admittedly, that is not always easy, but you don’t have to get it right on the first assessment.

TCPR: Why is it important to distinguish BD from BPD?
Dr. Zimmerman: The foundation of treatment for BD is medication. For BPD, the foundation of treatment is psychotherapy. Medications are also helpful for comorbidities in BPD, and psychotherapy is helpful in BD, but the core treatment is different.

TCPR: People with BD have long periods of sustained recovery. Can we see that in BPD as well?
Dr. Zimmerman: Absolutely. Mary Zanarini’s group did a prospective follow-up study on 290 patients who were admitted to McLean Hospital with a diagnosis of BPD. About half of them had a sustained recovery, meaning their functioning improved and their symptoms resolved. For many others, the symptoms fluctuated, often going away for two to four years and returning again (Zanarini MC et al, Am J Psychiatry 2010;167(6):663–667). Another 10-year follow-up study, the Collaborative Longitudinal Personality Disorders Study, found similar results. In my own practice, I look at the last five years when asking patients about symptoms of BPD.

TCPR: Should we tell people with BPD about the diagnosis?
Dr. Zimmerman: I do, but I understand that a lot of clinicians do not. In my group at Brown, we routinely surveyed patients about their satisfaction with the assessment, and there are no differences between patients who were told they had BPD and those who were told they had other disorders. In fact, many patients with BPD are reassured to know that there’s a name for their problems and hope for recovery. There is one small subgroup who does not like being diagnosed with BPD: mental health professionals.

TCPR: Suppose I work in an area without a DBT program. How will this diagnostic information help my patient?
Dr. Zimmerman: Well, one of the benefits of COVID has been the rapid expansion of telehealth services, so you might find that DBT is more available than before.

TCPR: What if my patient can’t afford therapy?
Dr. Zimmerman: John Gunderson and Lois Choi-Kain developed a treatment model for BPD that can be used outside of formal psychotherapy: “Good Psychiatric Management.” Patients work toward realistic goals, like gaining financial independence or having stable relationships, rather than just symptom reduction. As part of this, they are encouraged to participate in a structured social group outside of treatment. It may be a sports team or a religious or community group (Editor’s note: See our interview with Dr. Choi-Kain in TCPR June/July 2020).

TCPR: If you could include a screening instrument for every new patient at an outpatient practice, what would it be?
Dr. Zimmerman: I’m biased here. Our group developed the Psychiatric Diagnostic Screening Questionnaire (PDSQ), which screens for 13 common disorders in outpatient settings. Admittedly, it doesn’t assess for BPD. The PDSQ has questions like, “Over the past two weeks, did you worry obsessively about dirt, germs, or chemicals? Did you think that you were in danger because someone was plotting to hurt you?” It has been validated by independent groups, and it was used in the STAR-D trial.

TCPR: Does the PDSQ make a diagnosis or just screen for them?
Dr. Zimmerman: It is a screening instrument, not a diagnostic instrument. Screening instruments are designed to have high sensitivity so you don’t miss the disorder, as well as a high negative predictive value so you can be confident that when the screen is negative, the patient doesn’t have that disorder. The PDSQ has a sensitivity of 90% and a negative predictive value of 97% across all 13 disorders, based on a study of more than 2,000 psychiatric outpatients (Zimmerman M and Mattia JI, Compr Psychiatry 2001;42(3):175–189).

TCPR: Are there pen-and-paper screening instruments for BPD?
Dr. Zimmerman: Yes. The McLean Screening Instrument is one of the best out there, but it should not be relied on. Its positive predictive value is around 50%, which means a lot of people who screen positive do not actually have BPD (Zimmerman M and Balling C, J Pers Disord 2021;35(2):288–298).

TCPR: What is the gold standard for a BPD diagnosis?
Dr. Zimmerman: Many say it is a semi-structured interview like the SCID, but Robert Spitzer, who was editor of the DSM-III, has suggested a LEAD standard: Longitudinal Evaluation of All Data. That includes the semi-structured interview, follow-up over time, mental status, and collateral information from people who know the patient well (Spitzer RL, Compr Psychiatry 1983;24(5):399–411).

TCPR: Some of those LEAD elements are not in the DSM.
Dr. Zimmerman: I view the DSM as a guide. It’s a test, and like the other tests we’ve talked about, it can yield false positives and false negatives. For example, I have seen patients with depression who describe one or two past hypomanic episodes. They met the DSM criteria for bipolar II disorder, but I did not think they had it. The episodes were too infrequent, and the rest of the LEAD information did not fit. So I didn’t treat them with a mood stabilizer, and they have done well. On the other hand, I’ve seen patients who have clear and frequent hypomanias that only lasted two to three days. By the DSM’s four-day duration criterion, they did not technically have BD, but when I looked at the whole picture, I thought they did. I diagnosed bipolar unspecified and treated them as if they had BD.

TCPR: Structured interviews like SCID and MINI have licensing fees. Are there more affordable options?
Dr. Zimmerman: Not really, but keep in mind there are no magic questions on structured interviews. They simply inquire about the DSM criteria using everyday language. Their real advantage is in providing a structure for a thorough assessment, but the DSM itself provides a similar structure. I have written a book that suggests DSM-based questions, Interview Guide for Evaluating DSM-5-TR Psychiatric Disorders and the Mental Status Examination. It is not a structured interview, but if used with the DSM, it could serve as a low-budget structured interview.

TCPR: What mistakes do clinicians make when they use a structured interview?
Dr. Zimmerman: Being overly rigid. You can go beyond the questions that are printed to make sure the patient understands them.

TCPR: You’ve conducted hundreds of studies on diagnosis. Did any of your findings surprise you?
Dr. Zimmerman: Yes. When the revisions to the DSM-IV were first proposed, they suggested changing the personality disorders section to a dimensional system. I thought this would not work, so I did a study comparing patients who met zero criteria with those who met only one for BPD. My intention was to kill the dimensional idea by showing that people who met only one did not differ from those who met zero. Much to my surprise, there were rather robust differences between the two, such as in rates of comorbidities, suicidality, and overall functioning. So meeting one criterion makes a difference, but ultimately, they stuck with the categorical system.

TCPR: Thank you for your time, Dr. Zimmerman.