Both the DSM-5 Web site1 and Psychiatry News2 have recently heralded the proposed DSM-5 revision for diagnosis of personality disorders. The end product of much dodging and weaving on the part of the Work Group (eg, narcissistic personality out/narcissistic personality disorder back in), the revision is touted to resolve problems inherent in the DSM-IV treatment of personality disorders: it recognizes the dimensional character of personality traits and disorders, and it moves beyond the dichotomous (present or absent) strictures of DSM-IV diagnosis.
Those accomplishments granted, the news is not all good regarding the new model. A good look at the DSM-5 Web site (I wanted to say a “quick” look, but a quick look will leave you deeply confused) reveals a degree of complexity worthy of the finest medieval scholastic philosophers. If the current version seems disorienting, bear in mind that it was worse. The DSM-5 site notes, remarkably, that “Since its original posting on the APA’s DSM-5 Web site in February of 2010, all parts of the model have been simplified and streamlined in response to comments received and to critiques in the published literature.”3
The site also informs us that “The goal of the new criteria is to maximize their utility to clinicians and benefit to patients.” And the Psychiatric News article quotes Drs Oldham and Skodol as saying that “the system will require some training, but that once clinicians familiarize themselves with the criteria, they will find the system easier to use than DSM-IV, more intuitive, and more reliably concordant with the clinical presentation of patients with personality disorders."2 Surely they jest—but judge for yourself.
The new list
In this posting I want to focus on the unwieldy complexity of the new Personality Disorders section, but first let me note the revised list of personality disorders. The work group reduced the DSM-IV list from 10 to 5 (and then to 6, when narcissistic personality disorder was reintroduced following objections from the rank and file).
For each of the disorders removed from the DSM-5 list the explanation given is that “The Work Group recommends that this disorder be represented and diagnosed by a combination of core impairment in personality functioning and specific pathological personality traits, rather than as a specific type.” Anyone wondering how the decisions were made is on his own. For this reader at least, paranoid personality disorder (removed) is a readily identified personality disorder, and easily as worthy of inclusion as avoidant personality disorder (retained).
Doing the assessment
The assessment begins with Criterion A, Level of Personality Function, presumably to be part of every patient’s evaluation. The Levels of Personality Function Scale is a five-point rating scale for two domains of personality functioning—Self and Interpersonal. Under Self are subdivisions of Identity and Self-direction, and under Interpersonal are Empathy and Intimacy. This criterion will let us know whether and to what degree any patient has a personality dysfunction. Aside from issues of uncertain scientific foundation and laborious implementation, this scale is notable for its reintroduction of psychological and psychodynamic features that had been banned from DSM-III. Assessing someone’s capacity for empathy is not the same as checking for sadness or insomnia, and will take considerably more time as well as psychological sensitivity. I can ask someone if he feels sad, but I’ll have to talk to him at some length to determine his capacity for empathy.
Criterion B moves from general personality dysfunction to presence of one of the six disorders. Rather than just naming the criteria for each personality disorder, this criterion uses another architectonic structure in which pathological personality traits are assessed under five broad personality trait domains—negative affectivity, detachment, antagonism, disinhibition versus compulsivity, and psychoticism, as well as component train facets like impulsivity and rigid perfectionism for each of the five trait domains. Needless to say, another scale, the 4-point Personality Trait Rating Form, is administered to score the individual on these traits, and the particular personality disorder is defined in terms of particular trait clusters. (Are you confused? Just remember the reassurance that you will find it easier than DSM-IV.)
To deal with the person with significant personality disturbance who does not reach the threshold for a particular personality disorder, a diagnosis of Personality Disorder Train Specified (PDTS) can be given along with the relevant traits.
Evaluating the evaluation
There is something quite elegant about the DSM-5 Personality Disorders diagnostic system—an architectonic of divisions, subdivisions, and sub-subdivisions. On the other hand, for all their scholastic erudition, the work group have created a monster—a bloated, pedantic, cumbersome diagnostic instrument that will never be used by anyone working in the hurly-burly of clinical practice. Just imagine doing a routine new-patient evaluation and trying to include the personality disorder assessment, each of the first two criteria with its many-item scale, each item to be scored on a 4- or 5-point rating system. It’s hard to imagine anyone having the patience or motivation to use this instrument.
As I indicated above, the DSM-5 Web site assures us that “The proposed system is designed for flexible use to maximize clinical utility.” Who could make such an out-of-touch statement about this instrument, and who could construct such an instrument under the remarkable notion that it will be clinically useful? The answer is simple: experts—or I should say, well-intentioned experts, scholarly psychiatrists who have worked intensively in this area and want to inform the new manual with the benefits of their expertise.
What they lack, of course, is common sense—the ability to limit the implementation of their knowledge to what is practical useful. If we were starting over with the Work Groups, the biggest improvement would be to appoint a number of non-experts to each group—individuals with real-world experience who could say to the experts, this is a brilliant idea, but it’s totally impractical for the ordinary working clinician.
To illustrate, let’s take an example—borderline personality disorder (BPD). Start with the ICD-10, which offers a 100-word description of the disorder, highlighting the main features. The ICD-10 definition would be adequate for most clinicians and in fact reflects how most of us work. It would of course take a fraction of the time as the DSM-5 assessment.
Now let’s move on to DSM-IV. The manual lists 9 diagnostic criteria for BPD. The patient must meet 5 or more of the criteria to warrant the diagnosis. This approach to diagnosis has been criticized for requiring a categorical, dichotomous yes-or-no decision on the diagnosis, eliminating any possibility of a dimensional assessment that would permit diagnosing subthreshold, mild conditions. This flaw in the DSM-IV approach has been the principal argument for the elaborate dimensional approach of DSM-5, which allows the clinician to make the diagnosis along a continuum ranging from mild to severe forms of BPD.
Given the sheer impracticability of the DSM-5 model, what are our alternatives? We could take the simple ICD-10 approach. We could stick with DSM-IV and its flaws. If we wanted to address the flaws of DSM-IV, we could look for a simpler dimensional model than DSM-5. As it turns out, one has just been offered us. In a recent publication4 (already noted by Allen Frances5), Mark Zimmerman and colleagues have demonstrated in a thorough study of 2150 psychiatric outpatients that one could easily use the DSM-IV NOS category (with traits specified) to create a simple 3-point dimensional assessment with categories of absent, subthreshold traits, and present.
Using this makeshift DSM-IV dimensional assessment for personality disorders, they demonstrated that “There was no difference between the 3-point, 5-point, and criterion count methods of scoring the DSM-IV personality disorder dimensions . . . The DSM-IV 3-point rating convention was as valid as scoring methods using more finely graded levels of severity.” In the case of our example of borderline personality disorder, this method would allow for BPD, no BPD, and mild, subthreshold traits of BPD.
We now have two alternatives to the unmanageable system of DSM-5. We could elect for real simplicity and use ICD-10, or if we want a dimensional component with mild, subthreshold diagnoses, we could use the Zimmermann modification of DSM-IV. Either would allow for quick assessment and spare us the extraordinary time and labor required by the DSM-5 model (The ICD-10 system, with its own NOS category, could easily be modified a la Zimmermann to produce another simple, 3-point dimensional model.).
In the face of these alternatives, it is hard to justify adopting the DSM-5 proposal. One more time, a huge effort by well-intentioned experts that has produced an unwieldy, impractical instrument that will be used by no one.
1. American Psychiatric Association. DSM-5 Revisions for Personality Disorders Reflect Major Change: Public Comment Period for Proposed Diagnostic Criteria Extended Through July 15. Press release: July 7, 2011. http://www.dsm5.org/Newsroom/Documents/DSM-5-Revisions-for-Personality-Disorders-Reflect-Major-Change-.pdf. Accessed November 7, 2011.
2. Moran M. 2011. DSM-5 Developers Propose New System for Diagnosing Personality Disorders. Psychiatr News. http://psychnews.psychiatryonline.org/newsArticle.aspx?articleid=116625. Accessed November 7, 2011.
3. American Psychiatric Association DSM-5 Development. Personality Disorders. http://www.dsm5.org/ProposedRevision/Pages/PersonalityDisorders.aspx. Accessed November 7, 2011.
4. Zimmerman M, Chelminski I, Young D, et al. Does DSM-IV already capture the dimensional nature of personality disorders? J Clin Psychiatry. 2011:72:1333-1339.
5. Frances A. An alternative to the DSM-5 personality proposals. Psychiatr Times. September 29, 2011. http://www.psychiatrictimes.com/blog/couchincrisis/content/article/10168/1959616. Accessed November 7, 2011.