TCPR: In 2013, the American Psychiatric Association (APA) plans to release the fifth edition of the DSM. We are all aware that there may be some significant changes in personality disorder criteria in the DSM-5, which has created some controversy. Where do the proposed changes stand right now and how might this affect us in the long-term?
Dr. Zimmerman: I don’t know what the final version will look like, but there has been an outpouring of criticism from the personality disorder research community. There are questions about changes in the criteria for the existing disorders and concerns about the rationale for deleting certain disorders.
TCPR: Could you summarize the major changes?
Dr. Zimmerman: There is a new system to describe personality dysfunction, and the APA proposes removing schizoid, paranoid, histrionic, and dependent personality disorders, as well as personality disorder NOS. The DSM-5 retains six personality disorder types: schizotypal, avoidant, borderline, antisocial, narcissistic, and obsessive-compulsive.
TCPR: Of course, if you delete a personality disorder from the DSM, it doesn’t mean that the person is cured. The person still has a disorder and the difficulties that come along with it. Aren’t practicing doctors likely to just continue diagnosing patients as they have been doing all along?
Dr. Zimmerman: I think that clinicians will embrace the changes only when they see data demonstrating why changes should be made. An important clinical piece of information is prognosis. To what degree will changes in DSM-5 allow us to better predict how individuals will do? Will they give us some guidance that we can use to treat patients more effectively?
TCPR: Let’s start with one of the more radical changes, which is a proposal to switch from a categorical approach to diagnosis to a dimensional approach. Can you define what this means, and why it is important?
Dr. Zimmerman: The categorical approach to personality disorders is what we are used to, and it follows the medical model approach towards diagnosis and classification—somebody goes to a doctor and gets a diagnosis, such as diabetes or cancer. However, the fact is that most things in medicine—and in mental health—fall along a dimension. For example, take blood pressure. When you measure blood pressure you come up with a number, and based on that number, you make a judgment about whether intervention is warranted. But as human beings, we like to categorize, and as physicians we like to diagnose and treat. So we take that dimensional score and then we categorize it as “hypertension” or “normal.” Over time, as we collect more data, we may well change the cut-off point that we use to say something warrants treatment or not. Similarly in psychopathology, the constructs that we evaluate are not all-or-nothing, but they follow along a dimension of severity scores—whether depressive symptoms, anxiety symptoms, levels of substance use, or likewise dimensions of personality.
TCPR: Where in the assessment of personality disorders for a given patient would “dimensions” come into play?
Dr. Zimmerman: That can be done in different ways. Some argue that we should measure dimensions that have been shown to reflect normal personality. For example, we could rate a person’s level of introversion, extraversion, or rigidity, or other dimensions that personality researchers have found to characterize human experience. Particularly high or low scores on these dimensions could then be mapped onto the personality disorders identified in the DSM. This is essentially the approach being suggested for the DSM-5. Another approach would be to take the personality disorder constructs as we currently know them in the DSM and to dimensionalize them. For example, rather than saying that someone does or does not have borderline personality disorder or schizotypal personality disorder, you would make a rating of “how” borderline they are or “how” schizotypal they are.
TCPR: It certainly seems that in the real world of the clinic, psychiatrists are not spending a lot of time figuring out if a patient meets the formal criteria for each personality disorder.
Dr. Zimmerman: I agree that most clinicians are not rigidly applying all of the personality disorder criteria to their patients. In fact, they are rarely assessing them. Instead, we are listening to individuals relate their histories, tell their stories, and we consider patterns that emerge over time. We are thinking things like, what are the typical stressors that come up over and over in my patients’ lives? How do they react? How do they cope with those stressors? How are they processing information? How are they relating to the external world? What difficulties characterize their interpersonal relationships? And based on that, we are identifying certain personality traits that we find have an impact on their relationships, and on how they are coping with and dealing with the world. So in a sense, we naturally think in terms of traits and severity of those traits—which is precisely why the DSM-5 committee has come up with their dimensional system. It makes intuitive sense, though we have very little actual research to either support or reject this approach.
TCPR: Although your group at Brown University has begun to do some of this research.
Dr. Zimmerman: Yes. We recently wrote a paper called, “Does the Presence of One Feature of Borderline Personality Disorder [(BPD)] Have Clinical Significance? Implications for Dimensional Ratings of Personality Disorders.” We wrote it because we had the following question: If dimensional scoring is so important how come there has never been a study that has compared individuals with one criterion versus zero criteria? (Zimmerman M et al, J Clin Psychiatry 2012;73(1):8–12). Clinically, this question can have major consequences. For example, what will happen in a child custody battle when the lawyer for one parent accuses the other parent of being a little bit borderline? Does that have any meaning? We did a study on several thousand psychiatric outpatients who were evaluated with a semi-structured diagnostic interview for both Axis I and Axis II disorders. We then searched for patients with any of the criteria for borderline personality disorder. We did an analysis comparing patients who had no criteria for BPD vs those who had exactly one criterion. We had hypothesized that there would be no difference between individuals with zero and one criterion. So we were surprised when the results showed that those with a single criterion had significantly more psychosocial morbidity than those with no criteria.
TCPR: So this study implies that counting up the number of criteria might be useful in determining how ill a patient is. Did you compare patients with other numbers of criteria for BPD?
Dr. Zimmerman: We did that in a prior paper, in which we limited our analysis only to patients who had met criteria for BPD. We wanted to know if the severity of BPD, reflected by the number of criteria met, was associated with psychosocial morbidity. Surprisingly, we found that there was no association—there was no difference between groups in terms of pathology, whether you met five or six or seven or eight or nine criteria. Once you met the threshold for borderline personality disorder all the groups looked the same (Asnaani A et al, J Personal Dis 2007;21:615–625). The bottom line is that there is evidence for both the dimensional view and the categorical view of personality disorders.
TCPR: What is the role of a structured interview in personality disorder assessment? Should we be using such instruments in our practices?
Dr. Zimmerman: Four or five studies have compared what happens when clinicians use semi-structured interviews vs standard unstructured clinical interviews (Zimmerman M and Mattia JI, Am J Psychiatry 1999;156:1570–1574). And every study finds the same thing, which is no big surprise: when you administer a semi-structured interview you make many more diagnoses. When you ask questions for two to three hours you tend to get more positive information than if you are asking questions for 45 minutes to an hour. Time constraints in the real world mean that diagnostic information gets missed.
TCPR: The current DSM-5 criteria for personality disorders will require that we look for “self and interpersonal deficiencies” in order to diagnose PDs. Can you explain how we can best assess for these?
Dr. Zimmerman: There’s really no standardized way to measure these deficiencies, so clinicians should listen to patients tell their stories and describe what is going on in their lives over time. Patients describe their problems and how they deal with them. In the context of those descriptions, you hear about the impact on interpersonal relationships and the level of distress that is caused by their approach of dealing with others and dealing with the world. But the key is that they cause distress. Sometimes we see personality traits that should not be misconstrued as personality pathology.
TCPR: Finally, can you describe your work in the MIDAS project, which provides much of the data for your studies, and some of the take-home messages from your experience?
Dr. Zimmerman: MIDAS (Methods to Improve Diagnostic Assessment and Services) is an ongoing study at Rhode Island Hospital in which more than 3,500 psychiatric outpatients have been evaluated with semi-structured diagnostic interviews. Though we have not yet collected data to demonstrate this, our experience suggests that we can better treat our patients, and achieve better outcomes, the more comprehensively we assess them. I think, increasingly, the psychiatry field has farmed out psychotherapy to other mental health disciplines, and increasingly—certainly not exclusively—psychiatrists are doing pharmacologic management rather than more comprehensive management. And when you are doing psychopharmacology—with or without psychotherapeutic intervention—I think the more you know about the individual the better the care you are able to deliver to the individual. The other significant component of the MIDAS project has been the development of measurement tools for clinicians to better evaluate their patients upon initial presentation and follow-up visits.
TCPR: Thank you, Dr. Zimmerman.