Reading this chapter will have given you an understanding of some of the difficulties associated with the use of an exclusively categorical diagnostic system for personality disorders. For instance, you may have had some difficulty in developing a clear, distinctive picture of each of the personality disorders. You may also have recognized that the characteristics and attributes of some disorders, such as schizoid personality disorder, seemed to blend with other conditions, such as the schizotypal or the avoidant personality disorders. It is also the case that people frequently do not fit neatly into any one specific diagnostic category. They may also qualify for a diagnosis of more than one personality disorder (e.g., Clark, 2007 ; Grant, Stinson et al., 2005 ; Widiger et al., 1991 ). Indeed, a common diagnosis is “personality disorder not otherwise specified” (e.g., Livesley, 2007; Krueger & Eaton, 2010 ; Verheul & Widiger, 2004 ; Verheul et al., 2007 ), a category reserved for people who exhibit features from several different categories but do not cleanly fit within any of them.
In the past, many studies of personality disorder categories were conducted in an effort to find discrete breaks in such personality dimensions—that is, points at which normal behavior becomes clearly distinct from pathological behavior. None were found (Livesley, 2001 ; Widiger & Sanderson, 1995 ). Moreover, changes in the cut-points, or thresholds for diagnosis of a personality disorder, can have drastic and unacceptable effects on the apparent prevalence rates of a particular personality disorder diagnosis (Widiger & Trull, 2007 ). For instance, when the DSM-III was revised to the DSM-III-R, it was noted that the rate of schizoid personality disorder increased by 800 percent and narcissistic personality disorder by 350 percent (Morey, 1988 ).
Such issues are much less problematic when a dimensional (or continuous rating) system is used because it is expected that across individuals there will be many different patterns of elevation of scores on different facets of different traits. As noted earlier, the personality traits classified for the personality disorders are dimensional in nature. For example, everyone is suspicious at times, but the degree to which this trait exists in someone with paranoid personality disorder is extreme. A dimensional system would allow people to be rated on the degree to which they exhibit each facet and trait dimension—not on whether they do or do not have a given personality disorder. Each individual would also be rated on numerous dimensions, and highly personalized patterns of scores would thus be expected rather than problematic.
There has been lively debate among psychologists and psychiatrists over the best way to design a more dimensional system (Clark, 2007 ; Livesley, 2011 ; Skodol et al., 2011 ). Although it has long been clear that a more dimensional system is needed, actually implementing this has proven very difficult. A major challenge has been creating a scientifically valid diagnostic system that is not overly complicated and does not render the substantial research on existing categories useless. As we noted earlier, the proposal that was offered for inclusion in DSM-5 was not accepted. Although moving to a dimensional system makes a lot of sense for many reasons, the complexity of the model that was proposed was no doubt a major issue. All of this speaks to the difficulty of creating a valid yet utilitarian diagnostic system that satisfies the different needs of both clinicians and researchers.