Dimensional Models of Personality Disorders: Refining the Research Agenda for DSM-V

A major issue facing the developers of DSM-V is whether to adopt dimensional diagnoses, as opposed to categorical personality disorders, such as DSM-IV's categorical diagnosis of borderline personality disorder or a dimension of emotional dysregulation versus emotional stability.

Thirty international authorities addressed this issue at a conference in December 2004, which laid a firm foundation for DSM-V developers. At the conference they considered almost twenty different dimensional models, many of which follow from studies of normal personalities. One five-dimensional example included in the book is extraversion versus introversion, antagonism versus compliance, constraint versus impulsivity, emotional dysregulation versus emotional stability, and unconventionality versus closedness to experience.

Unlike categorical diagnoses, these five would pertain to both the well and the sick, and these are five traits, not clusters. Also unlike categorical diagnoses, five scales of these characteristics could be far more exact than DSM-IV's categorical personality disorders, and could therefore provide the potential for greater reliability. Borderline personality disorder, for example,with its five or more of nine signs, has 256 different configurations of those nine signs.

Furthermore, there is very little evidence that the psychopathology pie is divided up as DSM-IV has proposed. One of us warns medical students and residents—in "Peele's dictum" —that if their presentation of a patient fits DSM-IV criteria perfectly, the student will know that he or she has not talked to that patient.

Why not adopt these dimensions? The dimensional approach seems consistent with today's concepts of personality; consistent with medicine's interest in the impact of traits on medical illnesses, considering that the rest of medicine has no interest in DSM-IV's personality disorders; more reliable; easier to adjust for cultural differences and more culturally sensitive than the categorical; and very rational. Several problems exist.

First, the proposed benefits are only rational, not empirical, in that we never meet a single patient in this book who has been observed to benefit from any of the nearly 20 proposed dimensional models. Not one.

Second, cutoff points are going to be a huge challenge. What is normal, and what is pathological? Taking one example, on the dimension of antagonism versus compliance, where would the cutoff be to demarcate psychopathology? Would it be the same in all settings, or would it have to be defined in relation to the environmental need? Even at the extremes, are there not some situations in which extreme antagonism is lifesaving, and others in which extreme compliance is lifesaving?

Third, what about the current categorical personality disorders? They could continue in DSM-V —hopefully without the word "personality" in the title, a word that unnecessarily hurts patients and postpones treatment. Most of this book implies that the dimensional would replace the categorical.

Fourth, what about the worthy experts who have given their careers to contributing to our understanding of DSM-IV's categorical personality disorders? If DSM-V retires the personality categorical disorders, the American Psychiatric Association should give each of these experts a golden parachute.