The DSM 5 Follies, As Told in Its Own Words

Allen Frances

Professor Emeritus, Duke University

Nothing can illustrate how far DSM 5 has gone off track better than the words spoken in its defense by DSM 5 leadership. Dayle Jones and Suzy Chapman have assembled (from among many others) the 10 most wrong-headed quotations, and I have annotated them with my own thoughts.

1.      "And that's what the DSM is -- a set of scientific hypotheses that are intended to be tested and disproved if the evidence isn't found to support them."

No! DSM 5 is most decidedly not "a set of scientific hypotheses"; it is an official manual that will influence people's lives for better (and sometimes for worse) in enormously consequential ways. DSM 5 is not a research agenda "intended to be tested" after its publication; it must hit the ground safe and scientifically sound. Any DSM 5 proposal for change that lacks strong evidence should be considered "disproved" right now, without further ado before publication, not after. Our patients are not guinea pigs for reckless DSM 5 experimentation.

2.     "However, a lot of this has not been tested as well as we would like... Our plan is that these will be immediately tested once the DSM is official, and then one will be able to see if revisions can be made."

Again a resounding no! Allow entrance into DSM 5 only for changes that have already passed muster. Exclude everything that "has not been tested as well as we would like." DSM 5 is not a place for the untried speculations and the pet theories of the assembled experts. Add new suggestions only if they are well studied and of proven worth and safety.

3. "Our intent with the DSM-5 is to make it a much more fine-tuned, fine-grained system so that we have a better assessment with repeated research, like we have for blood pressure or cholesterol estimates."

DSM 5 is introducing a variety of dimensional measures. Dimensions are more accurate than categories in describing continuous variables (e.g., IQ, height, weight, blood sugar). Providing dimensions could have been a useful advance if done well. But the measures suggested by DSM 5 are ad hoc, untested, and so impossibly time-consuming and unnecessarily complex that they will never be used in clinical practice. A premature and poorly done dimensional system will have no result other than to give dimensional diagnosis an undeserved bad name. And it makes no sense to compare assessing mental disorders to assessing cholesterol or blood pressure so long as psychiatry has no objective biological tests. Our diagnoses are now, and will remain for some time, necessarily far less precise than those in the rest of medicine.

4."The revision should be a living document. That's so we can convene expert panels more frequently in the future."

Please spare us a future filled with frequently convened expert panels; their fickle suggestions will likely do far more harm than good. Expert panels are largely responsible for the mess DSM 5 is in now. Why so? Experts live in an ivory tower world and don't appreciate how suggestions that may work for them can be so badly misused in the real world. We should avoid making changes arbitrarily and instead wait until scientific discovery makes it clear that change is necessary. There are no prospects for quick breakthroughs in psychiatric diagnosis in the near future except for Alzheimer's, where diagnostic tests may be available in several years.

5. "The idea of medicalising normality comes from a perspective that there are no psychiatric disorders, and you need to avoid stigmatizing people by giving them one."

Not so! The fear that DSM 5 is conducting a "wholesale imperial medicalization of normality" originated with me, and I am a great defender of psychiatric diagnosis and treatment -- but only when it is done well and within its proper sphere. Most of the widespread criticism of DSM 5 is not directed toward psychiatric diagnosis in general, but toward the careless way DSM 5 has been prepared, and toward its disregard for the harmful unintended consequences it will most certainly cause.

6. "Our intent is not to increase or decrease prevalence, but to make something that is more accurate and scientifically based."

DSM 5 will dramatically increase prevalence by adding five new, very common "disorders" that together will mislabel many millions of people now considered normal. Also, DSM 5 plans to lower thresholds for many of the existing disorders, turning normal grief into depression and dramatically increasing rates for attention-deficit disorder and generalized anxiety disorder. It is simply irresponsible not to be concerned about or measure the major impact this will have on the over-use of medication, on stigma, and on the misallocation of scarce resources.

7: "The manual itself is not invested in representing any particular point-of-view or theory and is intent on reflecting the latest findings from scientific study."

Would that it were so. DSM 5 literature reviews are remarkably variable in quality; often one sided, incomplete, and unsystematic; and sometimes giving undue weight to unpublished papers or papers authored by DSM 5 work-group members. The DSM 5 changes are mostly unsupported by scientific evidence.

8. "A kappa for a DSM-5 diagnosis above 0.8 would be almost miraculous; to see a kappa between 0.6 and 0.8 would be cause for celebration. A realistic goal is kappa between 0.4 and 0.6, while a kappa between 0.2 and 0.4 would be acceptable."

DSM 5 is not-so-subtly warning us that the reliability results from its field trial came in so low that we should accept a level of diagnostic agreement far below the universally accepted minimum standards. It wants to include some definitions whose diagnostic agreements are barely better than chance -- which means they are useless. The unacceptably low reliability is caused by the remarkably imprecise DSM 5 writing. This was supposed to be corrected in a second stage of field testing, but DSM 5 so badly missed its deadlines for the first stage that Phase 2 had to be stealthily cancelled. Low reliability is unacceptable; DSM 5 should rewrite and retest its poorly performing criteria to ensure that they do not cause great confusion.

9. "When asked if he thinks the APA can adjust revisions by the end of this year, Regier says 'there is plenty of time.'"

Perhaps this casual attitude toward time explains why every DSM 5 deadline has been missed -- some by a whopping 18 months. The poor planning and execution of DSM 5 have already forced a postponement from its original publication in 2012 to 2013, and yet another postponement is now necessary if it is to meet even minimal standards of quality.

10. "There is in fact no outside organization that has the capacity to replicate the range of expertise that DSM-5 has assembled over the past decade to review diagnostic criteria for mental disorders."

This statement is in response to the request for independent scientific review contained in a petition endorsed by 47 mental health professional associations. Because it is internal, secret, and porous, DSM 5's own scientific review process has no credibility. But contrary to the quote, there is a very clear and highly desirable alternative. The Cochrane group is expert in conducting independent, evidence-based scientific reviews to guide medical decision making. Cochrane should be contracted by APA to review the most contentious diagnoses. There is no other way DSM 5 can possibly gain the public trust.

* * * * *

This is just a select sample from among any number of equally self-incriminating quotations that consistently hoist DSM 5 on its own petard. The public statements of the DSM 5 leadership consistently reveal just how insulated they are -- far out of touch with the proper purpose of their task and unable to see serious risks that seem perfectly apparent to everyone else. DSM 5 is probably stuck on its disastrous course unless it can finally be restrained by outside forces --some combination of press shaming, public and professional opposition, and/or governmental intervention. Time is running out.