To the Editor: As I interpret the findings, Dr. Drew Westen’s excellent article R4915510CEBCGGFD on personality disorder diagnoses indicates that we are not yet in a scientifically acceptable place with the categories provided by DSM-IV. In his well-written editorial response to this important finding, Gabbard—perhaps a bit tongue-in-cheek—suggests that there be one diagnostic system for clinicians and another for researchers R4915510CEBCDFII. I think we need one diagnostic system, and DSM-IV axis II does not provide the needed solution to the problem of categorizing personality disorders. As just one example, someone with a histrionic personality style could be normal, neurotic, narcissistic, borderline, or fragmented in terms of coherence of identity and continuity of regard for significant others.

I completely agree with Westen and Gabbard that observation of patients’ behavioral patterns during interviews and analysis of their narratives about self-other relationships and identity experiences are key to both personality diagnosis and case formulations. The problem that leads to the multi­diagnosis of any one patient by too many personality disorder categories is with the theory of diagnoses in the DSM-IV definitions. These definitions are excessively slanted by the menu-driven approach for axis I disorders. More important, these definitions are not scientific products based on empirical validity; they do not even have much basis in empirical reality. They are expedient compromises arrived at by committees with internal disagreements and by superiors to committees trying to meld apples and oranges. Unfortunately, the operations of NIMH study sections in reviewing grants to do research in this area tend to reify rather than revise the system. What a mistake this will be if these poor diagnostic personality disorder definitions were to remain static (and I say that even though I feel my own work has been very well represented in how the histrionic and narcissistic personality disorder categories have been defined in DSM-IV).

Westen has ably demonstrated why we need to be flexible and active at arriving at new research methods and tools: the definitions as given do not lead to good research measures and do not help clinicians develop formulations that lead to treatment plans. The take-home messages do not tell us what to do, but they suggest three injunctions:

• Clinicians, do not go from diagnosis to treatment plans without an intermediate step of formulating the causes of the problems;

• Researchers, do not consider that you are on serious scientific footing if you are basing evaluation and outcome instruments only on DSM-IV axis II definitions; and

• NIMH study-section colleagues, stop insisting that subject samples be homogenized by DSM-IV categories.