The groundwork for the preparation of the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM) began in 1999, under the direction of David Kupler. In A Research Agenda for DSM-V,1 Michael First, in his chapter on personality disorders, announced a shift towards dimensional classification in response to growing user dissatisfaction with the DSM's diagnostic categories. Following discussions at the APA/WHO/NIH Personality Disorders Conference held in Arlington in December 2004, Thomas Widiger et al published a monograph on the 18 main dimensional models describing normal and pathological personalities.2 The 27 members of the DSM-5 Task Force then drew up a first plan for the new revision of the DSM. The initial recommendations of the personality disorders working group chaired by Andrew Skodol included several major innovations, which were posted on the DSM Web site (www.dsm5.org) on 10 February 2010. These were principally a new general definition of personality disorders, new diagnostic criteria (W. John Livesley), a 5-point assessment of the level of personality functioning (Donna Bender), the introduction of a dimensional model inspired by the 5- factor model, with six domains covering 37 clinical facets (Lee A. Clark and Robert Krueger), and a reduction in the number of personality disorder categories from 10 to 5: antisocial, avoidant, borderline, obsessive, and schizotypal. The other disease entities figure in the DSM as a personality disorder with, depending on the case, specific traits: histrionic, narcissistic, paranoid, schizoid, dependent, depressive, or passive-aggressive.

The main argument that Skodol et al3,4 put forward for limiting the number of categories was the inadequacy of published empirical justifications of the validity of the other categories. And by limiting the number of categories it was possible to reduce the number of comorbidities, which were deemed far too numerous in personality disorders. Lastly, the use of traits rather than behaviors in diagnostic criteria was an acknowledgement of the continuity between personality and personality disorders.

For the diagnostic procedure itself, a prototype-matching approach, already used in psychology in the 1980s, was recommended. In this approach, the clinician is asked to use a 5-point scale to assess how well the subject matches clinical vignettes representing either clinical types or prototypical character traits.

Most reactions to the proposed changes were highly critical. In a signed comment,5 eight internationally renowned specialists considered that the new system was too complex, and that there was no proof of the validity of the dimensional model chosen. Mark Zimmerman criticized the proposed prototype rating system.6 A few months later, Andrew Skodol, in a Letter to the Editor,7 simplified the proposed system, restored narcissistic personality disorder as a category in its own right, and affirmed that the validity and clinical value of the proposed hybrid system would be verified in planned Held trials.

In the January 2011 version, the essential features of a personality disorder are impairments in Identity and sense of self and in the capacity for effective interpersonal functioning. To diagnose a personality disorder, the impairments must meet all of the following criteria:

A. A rating of mild impairment or greater in self and interpersonal functioning on the levels of personality functioning.

B. Associated with a “good match” or “very good match” to a personality disorder type or with a rating of “quite a bit like the trait” or “extremely like the trait” on one or more personality trait domains.

C. Relatively stable across time and consistent across situations.

D. Not better understood as a norm within an individual's dominant culture.

E. Not solely due to the direct physiological effects of a substance (drug of abuse, medication) or a general medical condition (eg, severe head trauma).

The simplification also concerns the number of domains considered: 5 (negative emotionality, detachment, disinhibition [vs constraint], antagonism, and schizotypy-psychoticism) and the number of facets: 25.

These simplifications did nothing to silence the criticisms, some of which emanated from the working group itself. Two members of the working group questioned the validity of a hybrid system in assessing personality, mirroring the problem diagnosing hypertension (Livesley8), and warned against (Verheul9) a complete break with the past, which would cast aside over 30 years of research on the diagnostic criteria of DSM-III and -IV, with no guarantee that the proposed new system had any advantage over the old one! Lastly, a survey of two personality disorder associations showed that most members were hostile to the proposed deletion of diagnostic categories.10

The APA's final decision was made public on December 1, 2012, confirming the deletion of Axis II and the maintenance of the 10 categories of the DSM-IV, and the addition of'”new trail-specific methodology in a separate area of section 3 to encourage further study.“ This decision constitutes an about-turn, but is also a novelty inasmuch as it places personality disorders among other mental disorders, which had in the past been strongly contested. This decision was taken despite recent publications on the predictive validity of the hybrid model and the heuristic value of the proposed new model.11,12

It is likely that field survey results disclosed by Allen Frances,13 indicating poor agreement on diagnoses (kappa coefficients) among experts, were also responsible for this last-minute change.

The current position, which bears witness to physicians' attachment to diagnostic categories, also, points out the inherent limits of the very principles of the DSM. Indeed, it is unrealistic to expect a single instrument simultaneously to prove useful in daily practice, to be reliable, and to have a heuristic value likely to promote understanding of normal and pathological psychological functioning.