Time: Redefining Crazy: Researchers Revise the DSM

Wednesday, Mar. 11, 2009
By John Cloud

If you wanted to make a list of important books you should read, what would you choose? Anna Karenina, maybe? The Bible? How about the Diagnostic and Statistical Manual of Mental Disorders?

It may not be at the top of your list, but the DSM, as it's usually called, is one of the most important books in the world. It attempts to categorize, describe and give a code number to literally every problem that can occur in your mind, from schizophrenia to borderline personality disorder to something called mathematics disorder, which is essentially being so bad at math that it amounts to a mental problem.

The DSM is important not only because it is wildly ambitious but also because mental-health professionals around the world have adopted its classification system. In the U.S., it is virtually impossible to get reimbursed by an insurance company for treatment unless a mental-health professional identifies your condition by a DSM diagnosis number. (The DSM code number for mathematics disorder, if you were wondering, is 315.1. The code for Tourette syndrome is 307.23; the code for sexual sadism is 302.84. As I said, the DSM tries to cover everything.) (See the top 10 medical breakthroughs of 2008.)

The American Psychiatric Association (APA), which owns the DSM, is in the process of rewriting the book, which was first published in 1952. The DSM-V, as the fifth edition will be called, is set to be published in 2012. But the process of researching it began way back in 1999 — five years after the publication of the last major revision, the DSM-IV — meaning the new book's production will take 13 years overall. (Read "How We Get Labeled By the DSM.")

Why so long? Last week, a research organization called the American Psychopathological Association (which goes by the acronym APPA, to distinguish it from the APA) brought many of the key players in the development of the DSM-V to a New York City conference to discuss some of the reasons the writing of the book is so complicated.

One obvious reason is that so many people have a stake in what the world defines as crazy and what it calls normal. Famously, homosexuality was listed as a DSM condition until a 1974 vote among APA members to remove it. Other groups of mental-health professionals and patients want certain disorders to be added (and covered by insurance): sexual compulsivity, for instance, is not in the DSM, even though "sexual aversion disorder" (302.79) — the persistent and distressing avoidance of genital contact not explained by another disorder such as depression — is included. (Read an interview with an author who has bipolar disorder.)

Debates about what should and shouldn't be in the DSM are fascinating and often bitter, and as I have pointed out before, the book makes at least one fundamental error in the way it conceives of mental problems: it ignores causes almost entirely. If you feel sad and tired for a couple of months, have trouble sleeping and making decisions, and gain weight, you can be given a DSM diagnosis of depression (296.31 or 296.32, mild or moderate, recurrent) and prescribed drugs for it — even if the reason for your funk is that you just lost your job. Such physiological responses as insomnia are evolutionarily natural (and sometimes helpful, in a jump-starting sort of way) when you go through a trauma like losing your job. But according to the DSM, only perfect is considered normal. Another basic problem with the DSM: it tries to reduce the vastly complex experiences of your mind to a single number.

At last week's conference, there were tantalizing hints that the DSM-V might fix some of these problems. Dr. Steven Hyman, provost of Harvard, a former psychiatry professor at its medical school and former director of the National Institute of Mental Health, agitated at the meeting for a new DSM framework that would stop trying to divide mental problems into discrete all-or-nothing categories. That method works for some medical problems — you either have leukemia or you don't — but depression, for instance, doesn't work like that. (Read "Why Do the Mentally Ill Die Younger?")

Rather, Hyman argued that many mental illnesses are problems that lie along a continuum from normal and functioning to disordered and tragic. To the annoyance of some old-fashioned DSM defenders, he made the case that the DSM should regard mental illness as "continuous with normal": less like leukemia and more like hypertension. You don't get diagnosed with hypertension until you meet a cut-off point for high blood pressure that takes into account other extenuating factors: your age, for instance, or the conditions under which the blood-pressure reading is taken. Depression should be the same: if you are sad because you just got divorced, the DSM shouldn't necessarily consider you to have an illness.

Such a diagnostic model wouldn't be simple, though, which is one reason the DSM is taking 13 years to rewrite. And in the meantime, the book still has to be useful to everyday clinicians seeing patients who need a code number for insurance companies. "It's like wondering how you repair the airport while the planes are still flying," Hyman said at the conference.

Hyman noted that medical problems, whether in the mind or in the body or both, are usually caused by some combination of genes, environment, behavior and chance. Despite the comforting modern notion that severe psychological illnesses are simply due to an unfortunate genetic inheritance, it is the exceedingly rare mental condition caused only by genes (Rett syndrome is one example.) Rather, if you take something like generalized anxiety disorder (300.02), there may be a variety of causes that set it off: genes that cause excessive activity in the fear-producing part of the brain called the amygdala; a stressful job that stimulates that activity; engaging in a dumb behavior like having an affair that exacerbates your anxiety; then randomly getting into an anxiety-heightening situation like a car accident. The DSM has to try to account for all of that complexity — causes, effects, unintended consequences — and still be definitive.

Hyman said in an interview that one way the DSM currently handles this complexity is to have what he called a "wastebasket" diagnosis called "not otherwise specified" (NOS), which captures just about anything that doesn't easily fit the categorical model. One major problem with the NOS diagnosis: pretty much anyone can qualify for a diagnosis that, by definition, is not specified. A 2005 American Journal of Psychiatry paper found that nearly half of a group of 859 people who sought psychological help in Rhode Island could be considered to have a DSM personality disorder if diagnosticians were allowed to include the "not otherwise specified" option. Another problem: how do you adequately treat patients whose illness is unspecified?

A continuum model like the one Hyman proposes could help solve this problem by recognizing that people aren't always one thing or another. They're sometimes just a little depressed, or a little anxious. To avoid medicalizing normal stress, the DSM-V would set a cut-off point within the spectrum. Of course, determining the right cut-off point for the DSM's 350 illnesses would take an enormous research effort, one that has begun for some disorders like depression but likely hasn't even been thought about for rare problems like sexual sadism.

Other attendees at the APPA conference indicated that the new DSM will almost certainly adopt a continuum model for mental illnesses. But don't be surprised if the book doesn't come out as scheduled in 2012. If the three-day conference came to any solid conclusion, it was that totting up all the ways our minds can fail is a lot harder than, say, explaining why your appendix might burst.