Allen J Frances M.D.

DSM5 in Distress

It Is Time To Scrap the DSM 5 Personality Disorder Proposal

Three Strikes And You Are Out

Posted Jun 29, 2011

Humans have just these two ways of sorting things- giving them a name or a giving them a number. We have been naming things since the days of Adam- it is often quite an inaccurate approach, but one that is very familiar to us. Sorting by numbers is certainly much more precise, but it comes more naturally to computers than to humans.

The naming method is called sorting by categories. It is best for situations with clear boundaries- either you fit the name or you don't. The numbering method is called dimensional and is best suited to describe things that have a continuous distribution with no clear boundaries. Neither method is superior to the other in all situations. Which you choose depends on what your purpose is and how clear are the boundaries of what you are describing.

In everyday life, we mostly use categories. Unless we are doing a physics experiment, we describe colors broadly as blue, or green, or red- not more accurately by giving their exact wavelength. When describing ourselves we normally use adjectives not numbers, except only for those characteristics that are continuous in their distribution and readily rendered numerically (like height, weight, IQ).
Dimensions would often be more accurate but are also more cumbersome and come less naturally to most people.

DSM 5 meant well in its effort to substitute dimensions for categories in its labeling of personality disorders . The categorical description is necessarily awkward and inaccurate because personality disorders have such unclear boundaries - with normality, with each other, and with other psychiatric problems. Giving something a name carries an implicit assumption it has reasonably well defined boundary and personality disorders often just don't. So it is understandable to hope that a dimensional description of personality would be a better solution.

The DSM 5 Work Group's goal was entirely reasonable, but its method turned out to be ludicrous. It somehow developed a dimensional system so impossibly complicated and confusing that even experts in the field were unable to figure it out. Clearly the proposal was the work of researchers trying to piece together their pet approaches in a crazy quilt that had no relevance whatever to the realities of clinical practice.

Predictably, all the outside experts in personality disorder were appalled- worried that the DSM 5 Work Group's muddled suggestions would result in a downgrade in both the clinical and the research work in personality disorder. In numerous journal articles, at meetings, and in a letter to the Work Group, they detailed how completely unfeasible was the DSM 5 proposal and how lacking in empirical support.
Finally, one of the Work Group members stood in open opposition to the rest of the Work Group and wrote a brilliantly reasoned paper detailing all the fatal flaws that bedeviled their proposal.

The well meaning Work Group listened, went back to the drawing board, and returned some months ago with a proposal that was allegedly simpler and more user friendly. Simpler perhaps to its creators, but still an impossible mess to the rest of us. The Work Group was showered with continued cries of confusion and criticism. Agreeably enough, it went back yet again to the drawing board and created its third proposal, recently posted on the DSM 5 web site.

I defy anyone to make sense of this latest in a series of frankensteinish proposals. The DSM 5 Work Group has proven itself, once again and for all times, simply incapable of preparing a proposal that anyone can understand or that has any chance of ever being used. Three strikes and you are out.

What to do now? Clearly, it is long past time to give up. Going back yet again to the drawing board makes no sense- this proposal is too flawed to be rescued by any magic of editing redemption. It is also too late in the DSM 5 process to test alternative dimensional approaches that might have been sufficiently user friendly for clinicians. There seems to be no choice. The existing section on personality disorders, with all its long admitted flaws and limitations, appears to be the only possible default position. To keep dimensional hopes aflame, a different and much simpler dimensional method should be developed by a new team of outside experts and included in the Appendix.

About the Author

Allen Frances, M.D., was the chair of the DSM-IV Task Force and is currently a professor emeritus at Duke.

More Posts