In case you missed it, there’s been a bit of an uproar about the new Diagnostic and Statistical Manual of Mental Disorders (DSM). DSM is the go-to guide for diagnosing mental disorders; it’s published by the American Psychiatric Association (APA). The manual contains descriptions of mental disorders, symptoms, and other criteria to support consistency and accuracy in diagnosis; it has also been the basis for reimbursement followed by health care providers, insurance companies, and Medicare.
How could something with a title so dry and a purpose so seemingly utilitarian cause such a fuss?
Much of the fuss came before the manual was even released. And that explains some of it.
Yet the DSM often informs how clinicians, researchers, policymakers, and the public interpret mental health conditions and diagnoses, so its impact on treatment and funding decisions can be profound. The manual’s latest revision has been an arduous, contested process. One measure of how salient the DSM is: During three open comment periods in the revision process, the APA received 13,000 comments and 12,000 e-mails and letters from clinicians, researchers, and patient advocates.
The gargantuan task of revision was led by David Kupfer, who is the Thomas Detre Professor of Psychiatry at the University of Pittsburgh and chair of the APA’s DSM-5 Task Force. (See "Works in Progress" to learn more about the man behind the manual.) Kupfer, with vice chair Darrel Regier, executive director of the American Psychiatric Institute for Research and Education and director of the APA’s Division of Research, directed the task force’s efforts to revise the retiring DSM-IV, which had served as the gold standard since 1994.
Long overdue, DSM-5 itself was 14 years in the making; it represents the scientific input of more than 500 experts from the United States and abroad. It takes into account developments that could barely have been imagined 20 years ago.
Shortly before unveiling DSM-5 at the APA’s 2013 Annual Meeting in San Francisco in May, Kupfer spoke with UPMC Endowed Professor in Translational Neuroscience David Lewis, who succeeded Kupfer as Pitt’s chair of psychiatry and is Western Psychiatric Institute and Clinic’s medical director and director of research, and also Maggie McDonald, who worked as a science journalist specializing in psychiatry in the 1970s and ’80s. McDonald is Pitt’s associate vice chancellor for academic affairs, health sciences. She also holds appointments as assistant professor of epidemiology in the Graduate School of Public Health and of psychiatry in the School of Medicine. These edited conversation excerpts give a glimpse behind the scenes of the revision process, DSM’s move towards criteria based on emerging biological research, and the future of the manual. Number 5 is the first online DSM, and Kupfer imagines that it may be the last print version. He sees it as DSM 5.0—a more agile, living document that will adapt as the science behind psychiatry progresses.
—Introduction by Josie Fisher
Maggie McDonald: You have been working on DSM-5 since 2006. We’re here to talk about this journey.
DSM-II [published in 1968] was a thin, half-inch-thick book of definitions made up by prominent psychiatrists of the time. Then, DSM-III [which came out in 1980] was the first criterion-based system. It was really a sea change in defining psychiatric illnesses. How does DSM-5 differ substantially from the previous version, DSM-IV?
David Kupfer: To put it in perspective, DSM-IV did not differ much from DSM-III. So we’re really talking about what substantial changes have taken place since 1980 that need to be incorporated into DSM-5. DSM-III was influenced by a group at both Washington University and Columbia, and it represented the consensus of the research and diagnostic criteria of that time and from the early to mid-1970s. DSM-5 constitutes a much wider group of individuals involved and very different procedures used to arrive at changes.
David Lewis: What principles guided the process?
DK: The first was that [after 30 years] everything was up for grabs in terms of looking at every diagnosis. On the other hand, the thresholds and standards we used for change were quite high.
Another is that we espoused the position that development needed to be thought of across the entire lifespan. So we removed the first chapter of DSM-IV, which dealt with all of the disorders of what I would call childhood. Instead, we would work to ingrain the whole continuum of both age and development within each major cluster of disorders.
Another principle: The DSM, since it is primarily to be used by clinicians for clinical assessment, would be [designed for ease of use by these practitioners], although its [consistent application] would inform research across various fields. Another principle was to move DSM-5 closer to the rest of medicine . . . to say that whether you had a psychiatric condition or a medical condition, it was all on the same axis. Furthermore, we felt that we could do a better job of aligning the DSM to the next edition of the ICD, the International Classification of Diseases (which covers all of medicine and psychiatry), developed by the World Health Organization.
MM: Have there been particular advances in neuroscience that have broadened the base of evidence that have allowed you to bring the psychiatric disorders closer to the medical diagnostic model?
DK: The optimism was that by the time we finished DSM-5, we would have enough information for some of the major disorders—whether from genetics, neuroimaging, or cognitive neuroscience—to apply some of these biological variables as diagnostic criteria or to enhance diagnostic criteria in existing categories. We haven’t gotten where we would like to get.
The chapters are reorganized so that they are more neuroscience compatible. [For example], chapter one is neurodevelopmental, which has autism and ADHD. Chapter 2 is schizophrenia and other psychoses. And chapter 3, standing by itself, is bipolar disorders. Only in a very few disorders have [scientists pinpointed] a biological variable—for example, in narcolepsy. We have some of that in the neurocognitive areas. But not very much of it. Hopefully, we’ll get there soon.
That leads me to our changing the Roman numeral V to the Arabic 5. DSM-5 can be a living document. We don’t have to wait 20 years for the next version. And so, hopefully, in three or four years, changes in a version 5.1 or 5.2 might include [specific biological] variables relating to psychosis and schizophrenia to diagnose, say, 20 percent of the people who have psychosis in a more objective fashion than we can now. And 5.1 doesn’t have to affect all diagnoses. Updates to specific sections can be made to the online DSM as needed.
[Regarding moving the DSM closer to the medical model:] One thing we did is we reconsidered conceptually somatic disorders. . . as conditions due to . . . or associated with medical conditions. We assumed that there may be an etiology related to both a psychiatric disorder and a medical disorder.
This assumption of comorbidity relates to a discussion of categories versus dimensions in the DSM. We need categories because we need the code to be reimbursed. However, in the world of psychology and science, most of us think in dimensions. We think about continuous measures. And so, DSM-5 reflects graded levels of severity in many diagnostic areas. And some diagnoses are grouped to reflect more of a dimensional way of thinking about how [patients] got to where they are. [Dimensional assessments rate the presence and severity of symptoms in increments such as “very severe,” “severe,” “moderate,” or “mild.”]
DL: Some changes in DSM-5 have attracted controversy. Could you speak to one or two examples, and what you think about the basis for the controversy?
DK: So, let’s take a couple of them. One of the major areas of public discussion and clinical discussion was around autism. The data suggested that there weren’t such fine differences between Asperger’s, pervasive developmental disorder, and autism. For years, people have talked about putting these together and calling it autism spectrum disorder. We’d look at the major symptom clusters in a dimensional way and, therefore, be able to grade different levels of severity and need, using a label called autism spectrum disorder. We decided that if we [presented the autism spectrum diagnosis early on in the revision process], hopefully studies in the field would follow that would allay what some people feared might happen [if the previous four diagnoses went away]. People feared that the prevalence of these disorders would dramatically change [either through lack of diagnosis or overdiagnosis in light of the new classification]. Some feared that educational institutions and other institutions would deny benefits for children if they were, quote, “not diagnosed with Asperger’s” or one of the specific [previous diagnoses]. What we’ve discovered is that the diagnosis of autism spectrum disorder seems to work well. The scientific literature, as well as major associations of advocacy groups and advocacy/scientific groups like Autism Speaks, have, in general, endorsed the change.
Another [area of controversy is the] so-called bereavement exclusion. Let’s think about it in terms of primary care physicians, obstetricians, and gynecologists. According to DSM-IV, in essence, if somebody were in the early phase of bereavement (within two months of having a loss), it was not permissible to diagnose clinical depression.
So, it made some interesting assumptions: One is that everybody would always be confused about making that diagnostic differentiation and would assume that even if someone were, quote, “severely depressed and suicidal,” you shouldn’t do anything for the first two months. We were told [by a number of groups] not to get rid of the bereavement exclusion because that would be permission for, say, gynecologists and obstetricians to immediately give everybody antidepressants.
. . . If, for example, a couple had lost a child at birth, when they really needed grief counseling. So we’ve gotten rid of that exclusion; it’s not part of the criteria of major depression anymore. [Instead] we put in two different notes, carefully crafted, explaining the difference between sadness, grief, and clinical depression. One note is included right with the criteria set in the short version of the DSM that people keep on their desks. In the [longer] text is a further explanation. Now, having done all of those things, there is still a great deal of furor about what we’ve done.
DL: In The New York Times and The Wall Street Journal, and in the scientific literature, we continue to read about advances in imaging the human brain. President Obama has initiated a new process to map the human brain. To what extent does DSM-5 incorporate findings from brain imaging, or, if not, when do you think that will become part of the psychiatric diagnostic criteria?
DK: Some of us would have hoped that [even] without the new brain initiative coming up, that we would have had enough data to [include brain imaging and other biologically based evidence] in the actual [diagnostic] criteria sets.
There is mention within the text of specific disorders of some major things going on from a genetic and imaging point of view. And for the first time there are references in the DSM (online version), linking to the actual journal articles.
MM: You mentioned that one of the issues with a new classification system is its association with reimbursement for care. How are the DSM-5 and the American Psychiatric Association working with the insurance industry to be sure that people who need to be reimbursed don’t suffer from the changes?
DK: In the development of the DSM-5, and appointing of the work groups and all of the members, there were two constituencies that were purposely left out. They represented a level of bias that we did not want. One of them, not surprisingly, was the pharmaceutical industry, and the second was, basically, insurance companies. Now, with DSM-5 coming out, obviously all of us are going to have to deal with those two constituencies. We will likely see that some of the changes that we’ve made will facilitate the use of both pharmacalogic and nonpharmacalogic agents in mental disorders. But the second thing that we may see is that we may have made some quote, “changes in reimbursement,” which represent opportunities for the insurance industry to change the level of coverage for certain mental disorders and psychiatric conditions. We are already working with clinicians and physicians to explain the coding for DSM-5 and the coding to use for reimbursement.
MM: Over the last decade or so, there seems to have been a transition from the use of the word “psychiatry” to the use of the words “behavioral health.” Do you think that change has helped or hurt our understanding of the root causes of these disorders?
DK: I’m not sure. I do think that we’re still dealing with a certain level of stigma, no matter what you call it, that pervades all of medicine and therefore is also driven by the decisions that we have made as a society—which relates to reimbursement, which relates to much of the separation of mental and addictive disorders from other medical disorders, which did not work out, I think, to the advantage of patients and their families. And by doing so, we don’t understand that we’re dealing with patients who have a chronic psychiatric condition and another medical disorder. It’s not an accident that I have a strong interest, since 50 percent of the patients who have serious bipolar disorder, my own specialty area, have metabolic syndrome. We are dealing with major medical problems and major psychiatric problems in the same individual all the time.
It’s not perhaps an accident that the first commentary I wrote for JAMA [related to the new manual] was why all of medicine needs the DSM-5. And I guess it is timely that the [April 24 issue] of JAMA had a viewpoint which is basically DSM-5: The future has arrived.
All of medicine needs to understand more about mental disorders. And the root causes, as you put it, are going to be found to have a lot more common etiological features than we ever suspected.