Psychiatry's Diagnostic Manual Faces Major Scientific Overhaul

The American Psychiatric Association has announced big upcoming changes to psychiatry’s big book of mental disorders, the DSM. A massive reference book of mental health conditions known as the “bible” of psychiatry is going to change. The fifth and current version of the Diagnostic and Statistical Manual of Mental Disorders (DSM) lists nearly 300 distinct conditions that mental health care providers can diagnose and treat. But this dictionary of disorders has long been a lightning rod for criticism—in particular, about the way it classifies mental illnesses, which experts have said is not scientifically valid.

Today the DSM’s publisher, the American Psychiatric Association (APA), announced plans to address these problems by changing how the book works. The Future DSM Strategic Committee is proposing that the DSM change its guidance for diagnosis and increase its focus on ostensibly more objective measures of disease—“biomarkers” that may indicate mental illness. The changes would completely reshape future iterations of the manual.

“We have to do it right, so it might take a little bit of a time, but we’ll try and do it as fast as possible because the field is ready for it,” said APA psychiatrist Nitin Gogtay at a recent press conference about the changes.

The overhaul, outlined in five papers published today in the American Journal of Psychiatry, reflects the APA’s optimism that the DSM can be retooled to become more scientific. The committee has even suggested to change the DSM’s name so that the “S” stands for “scientific” rather than “statistical.”

But some experts don’t think the changes will do much to improve the manual. “I’m not sure [this new model for diagnosis] will have any strong utility at this point,” says psychologist Ashley Watts. “I’m worried that by trying to please everyone, we please no one.”

The proposed changes would allow mental health professionals to give more nuanced diagnoses. Currently, people are diagnosed with named conditions, such as major depressive disorder or bipolar I disorder, with very specific criteria—partly because health insurance needs these for accurate billing. But providers might not have enough information available to make a correct diagnosis; an emergency room doctor treating someone experiencing a psychotic episode likely doesn’t have the capacity to know if their diagnosis should be schizophrenia, bipolar I or something else.

“Clinicians often see themselves as being compelled to give a specific diagnosis, even when there’s very little certainty,” and that is ultimately unhelpful to patients, said psychiatrist Maria Oquendo, who leads the committee, at the press conference.

The plan for the new DSM will allow doctors to give people diagnoses at varying levels of specificity. It will also make space to gather more information about individual patients. This includes contextual details about their life (such as their socioeconomic status, medical conditions and quality of life), other mental health symptoms (such as anxiety or lack of pleasure) and biological features (such as genetics).

The DSM committee has been clear that the model it is proposing will evolve based on feedback from clinicians, scientists, patients and their families before a new version of the manual is released.

Including biomarkers is perhaps the most controversial possible change. Scientists haven’t yet found reliable biological signatures that reveal if someone has a particular mental illness. The closest we have come to that is for Alzheimer’s disease; doctors can now screen people for it with blood tests.

The APA’s researchers plainly state in the new papers that there are no other established biomarkers for DSM diagnoses but that they want the manual to be able to incorporate them if and when they become available.

“The question is really no longer whether biomarkers belong in DSM but really how to introduce them in a way that is rigorous, transparent, ethical and clinically useful,” said psychiatrist Jonathan Alpert, a member of the DSM subcommittee on biomarkers, in the press conference.

Watts, for one, is skeptical that biomarkers will ever be useful for doctors in diagnosing mental illness, assuming they can even be found. Pinpointing biomarkers would likely involve expensive and sometimes invasive tests that wouldn’t necessarily offer people anything better than doctors can now using their current practice of diagnosing based on behavior, she says.

Steve Hyman, former director of the National Institute of Mental Health and a vocal critic of the DSM, doesn’t think scientists will ever find biomarkers for the conditions listed in the manual. Part of the reason is that the DSM’s disorder categories may not reflect how mental illness actually works.

The manual’s third edition, DSM-III, published in 1980, drew borders across the landscape of mental illness based on how people’s symptoms seemed to cluster together. Going into the 1990s, psychiatrists were optimistic that these borders would also show up in brain scans and genetics research. But that didn’t happen.

There are very few obvious “clusters” of illnesses around which to draw borders, leading many experts to suggest that our models of mental health conditions should be based on spectrums of traits rather than specific named disorders. Such a model would be challenging to implement in doctor’s offices, however. Even Watts, who has helped develop a dimensional alternative to the DSM called the Hierarchical Taxonomy of Psychopathology (HiTOP), has “some misgivings” about how this would work in practice.

Both Watts and Hyman think the DSM’s categories have hindered scientists trying to understand what causes mental illness. For example, when studies recruit participants based on the DSM’s criteria for schizophrenia, they can miss real links to bipolar disorder that may be crucial for understanding what’s actually going on.

Scientific research of mental illness has already started to diverge from the DSM. Once we know enough about the underlying biology, the two can be integrated again, Hyman says. “It’s going to take a long time—I always say, I’ll be dead—but in the fullness of time,” the two can be brought back together, he says.