Debate remains over changes in DSM-5 a decade on

Changes to the DSM-5’s diagnostic criteria for autism were meant to add clarity, but they also generated new questions.

Collage of multiple figures on a hillside looking at a large blue book in the distance.
Illustration by Julien Pacaud
Listen to this story:

Just over 13 years ago, when the draft of the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) was released, debates erupted almost immediately. Many of them centered on the manual’s folding of Asperger syndrome into the autism spectrum disorder definition and concerns that some autistic people would be excluded from services under the new guidelines.

In the years since, things on that front have mostly “settled down,” according to Catherine Lord, distinguished professor of psychiatry and education at the University of California, Los Angeles. But even as old debates die out, the past few years have seen a new one spring up: whether the term “profound autism” is a useful way to differentiate some members of the spectrum. That has parts of the autism community again eyeing the DSM-5, hoping for a revision and additional clarity for clinicians.

But the story of how the DSM-5 came to say what it does suggests a tough road ahead. It took years of collaboration between autism researchers and a fledgling group of autistic self-advocates. And to those closest to the matter, it’s not even clear there will be a DSM-6.


he American Psychiatric Association (APA) publishes and maintains the DSM. The prior major update to the manual was done in 1994, and in 2007 the APA gathered 13 work groups to overhaul diagnostic criteria and update accompanying scientific information to create the DSM-5.

When Ari Ne’eman, co-founder and then-president of the Autistic Self Advocacy Network (ASAN), and his colleague Steven Kapp heard that there would be major changes to the criteria for autism for the fifth edition of the DSM, they knew what was at stake.

Ne’eman and Kapp, who was then a doctoral student and ASAN chapter co-director, understood that whatever was laid out in that book would affect the lives of autistic people for years to come, just as it had when the DSM-IV was released.

Yet the two men also knew they were being presented with an opportunity. Updating the diagnostic guide is at its core an academic process, but also a “political” one and therefore subject to influence, Kapp and Ne’eman wrote in a book chapter in 2020. Getting a diagnosis is a critical point in an autistic person’s life, and ASAN wanted to help shape the process.

ASAN, however, was on the outside. The Neurodevelopmental Disorders Work Group’s purview included autism, and the APA appointed its members — all vetted researchers in the space. This did not sit well with Kapp or Ne’eman. The autistic community, the two men wrote in the book chapter, possessed an “indisputable moral claim to be represented in the DSM-5 process on an equal basis” as the researchers. Because they could not force the APA to accept them as contributors in an official capacity, they launched “an intelligence operation” to get autistic voices heard.


e’eman started the effort in 2009 with a member of the work group he described in the book chapter as “Member A” — a researcher he already knew through previous contact and their “common ground” thoughts about rigid behaviorist interventions. This member was Amy Wetherby, a speech-language pathologist and director of the Autism Institute at Florida State University.

Wetherby had worked with Ne’eman before and was impressed with his rhetorical skills. She mentioned ASAN to the Neurodevelopmental Disorders Work Group. But Ne’eman also knew Susan Swedo, chair of the work group and a National Institute of Mental Health researcher at the time. She invited him to meet with the group in person, and on 8 April 2010, Ne’eman stood before the them and laid out ASAN’s thoughts.

“He was able to bring many points to us, for us to consider very carefully and seriously,” Wetherby says.

The members were sufficiently intrigued, and that opened up an avenue for more robust communication with some members of the work group. Ne’eman was the one who lobbied for autism to be diagnosed “currently or by history,” says Lord, who was a work group member. That change made it into the final version of the DSM-5 and expanded opportunities for adults to get diagnosed later in life.

The relationship between ASAN and the work group was “complex,” Kapp and Ne’eman wrote, but it was also “historically significant,” as the two men knew of “no prior example of successful advocacy to refine and improve diagnostic criteria from the community subject to it.” Indeed, Ne’eman is officially listed in the DSM-5 as an adviser to the Neurodevelopmental Disorders Work Group, along with more than a dozen researchers.

The two groups did not see eye to eye on every topic, but they shared an ambitious goal: to fold the diagnoses of Asperger syndrome and pervasive developmental disorder-not otherwise specified (PDD-NOS) into the autism spectrum disorder diagnosis. Beyond that, ASAN also felt “deeply worried” about proposals to include in the DSM-5 “recovery” criteria — details on the way autistic adults might no longer meet criteria for an autism diagnosis — that could lead to autistic people losing access to services and legal protection.

Clinicians use criteria in two domains to diagnose autism, and each domain has sub-items. ASAN wanted to decrease the sub-items required in the broader social-communication domain, positing that only two of the three sub-items should need to be met to obtain a diagnosis. This, again, was meant to keep an autism diagnosis easier to achieve and to keep support and services available.

After some three years of lobbying, ASAN failed to convince the work group to reduce the number of social-communication sub-items needed for autism diagnosis. But it achieved its goal of excluding recovery criteria from the DSM-5, and, most importantly, the definition of autism was united: Gone were Asperger syndrome and PDD-NOS from the landscape.

The work group also added a new condition called social communication disorder (SCD), which some felt was meant to help address fears that the loss of Asperger syndrome and PDD-NOS would create a gap.

But the end result of these changes is that today autism is being used to “describe an incredible range,” Lord says. That in its own right has only led to more confusion — the exact opposite of the intended effect of collapsing autism into one diagnosis.

Questions of diagnosis and identity swirl in the autism community. Some, such as Alison Singer, president of the Autism Science Foundation, suggest a need to return to the days of the more divided spectrum that was in the DSM-IV. Swaths of others say this suggestion is merely “weaponizing” autism’s heterogeneity, as Mary Doherty, founder of Autistic Doctors International and Autistic Med Students, wrote in Spectrum in 2023.

“In my opinion, all autism professionals should consider carefully their own positionality, and any discussion around changing or revising the diagnostic criteria should include autistic clinicians, particularly autistic psychiatrists,” Doherty says.

The question these days, Lord says, is how to best communicate respect for autistic people “while also acknowledging that we need to seek help and accommodations for the people that we see in the clinic.”


hen a draft of the DSM-5 was released in 2010, the removal of Asperger syndrome as a separate diagnosis quickly drew attention from not just academics, but the general public. Members of the work group felt it necessary to explain their decision.

Francesca Happé, professor of cognitive neuroscience at King’s College London in the United Kingdom and a member of the work group, wrote in Spectrum in 2011 that there was not enough evidence to support the diagnostic distinction between subgroups of autism, such as Asperger syndrome. Happé also wrote that the diagnostic criteria for autism subgroups had been applied inconsistently, and the idea behind eliminating them was to acknowledge the continuity found in the autism spectrum.

ASAN agreed with this, believing that folding all three conditions into one diagnosis would help eliminate the stigma of an autism diagnosis versus an Asperger diagnosis.

But there was also a sense that something had been lost. Fred Volkmar, professor of child psychiatry, pediatrics and psychology at Yale University, led the field trial portion of the DSM-IV revision for autism. He was originally in the work group for the DSM-5, but he felt the group was putting too much emphasis on research methodology and not clinical applications, and he resigned. He was against losing the Asperger syndrome label. Some who had been previously diagnosed with Asperger syndrome agreed, expressing a sense of disconnect. Sandra Jones, writing in Spectrum in 2020, said her research found that for many, “Asperger syndrome is more than a diagnosis. It is an identity.”

There was also fear that the diagnosis of autism would become too narrow without Asperger syndrome or PDD-NOS included. In 2012, Volkmar gave a presentation at a meeting in Iceland that used previously collected data to show that many people diagnosed with Asperger syndrome or PDD-NOS would not meet the DSM-5 criteria for autism. The data looked troubling, and the media picked up on this angle, enough for a New York State assemblyman to introduce a bill (eventually stalled in the Senate) meant to ensure continued access to services for those diagnosed with autism under the DSM-IV, just in case the DSM-5 excluded them.

But the work group rejected the idea that the DSM-5 would miss large swaths of autistic people. Lord said the data Volkmar used in his presentation were from the 1990s, and because the questions being asked in the DSM-5 were different from those used back then, the comparison wasn’t valid.

The water around inclusion data is still murky. A meta-analysis from 2019 indicated that the change to the DSM-5 criteria did lead to fewer diagnoses of autism among some who would have been diagnosed with Asperger syndrome, but at the same time, the prevalence of autism in the United States has continued to climb, with experts saying one reason for the increase is better screening and more accurate diagnoses of children. (According to the most recent report from the U.S. Centers for Disease Control and Prevention, autism prevalence now sits at 1 in 36 children.)

A table of severity levels added to the diagnostic criteria also drew attention. It detailed three tiers of autism — people “requiring support,” those “requiring substantial support” and those “requiring very substantial support” — but critics said there was no clear quantitative method to determine what the levels should be. ASAN was against it, and even Lord says the APA told the Neurodevelopmental Disorders Work Group to include it.

The DSM-5 offers additional specifiers for autism diagnoses, too, including “with or without accompanying intellectual impairment” and “associated with a known medical or genetic condition or environmental factor.”

But the specifying terms have not been effective, Lord says. “Nobody uses those,” she says. “I mean, it’s just not practical to have long phrases.” In the end, the levels ended up “almost tautological,” Lord says, and should be improved to be more evidence-based.


he ongoing debates have raised hopes in the broader autism community that some issues might be resolved in a DSM-6. Kapp and Ne’eman also considered that a new version might come: On behalf of autistic advocates, they laid out a plan for future consideration in their 2020 book chapter. They wrote that if a DSM-6 rolls around, the APA “should include autistic people within the process in an explicit and acknowledged fashion.”

In the decade since the release of the DSM-5, Ne’eman and Kapp have both stepped away from ASAN leadership. Ne’eman, who could not be reached for an interview despite Spectrum’s multiple attempts to contact him, is a doctoral student at Harvard University. Kapp, who is senior lecturer of psychology at the University of Portsmouth in the U.K., declined to be interviewed.

ASAN stands by its position on the DSM-5 from 10 years ago, says Zoe Gross, ASAN’s director of advocacy. “We are still definitely in support of the unified diagnosis, which has eliminated a lot of confusion and clinician-by-clinician variation in the diagnostic process,” Gross says. Of a DSM-6, she says, “we would want to see criteria that reflect autistic people’s experiences, not just outside observations of autistic people.” An example of something that could be included, she says, is more about motor issues. “Autistic people discuss experiencing issues with movement and motor planning beyond what is described in the current criteria, including ‘getting stuck’ or catatonia,” she says.

But it isn’t clear a revision is coming, even 10 years after the previous update.

Michael First, professor of clinical psychiatry at Columbia University, was an editorial and coding consultant for the DSM-5. There were nearly two decades between the publication of the DSM-IV and the DSM-5. When asked if 6 is being considered, First said it’s “an amazingly excellent question.”

After the release of the DSM-5, the APA decided to move to a new model of updating the guide, one that is iterative and more nimble. It’s possible that the APA will consider a DSM-6 if diagnostic criteria need major changes. But for now, suggested changes are submitted online and reviewed by a small committee of experts. If the proposed change is accepted, it is posted for public comment on the APA website. Indeed, in 2022, a formal text revision of the DSM-5 was published, which included small clarity changes to the autism entry. So far, this process is “working pretty well,” First says, and there are “no plans” for a DSM-6, though he says “the door is open.”

Regardless, some, such as Volkmar, think that maybe a book isn’t the way forward anyway. He is pinning his hopes on finding viable biomarkers that have not yet been discovered, though not for lack of trying. For her part, Lord agrees that “we need ways of better acknowledging the very different set of circumstances” of each autistic person, “but I don’t know how that fits into a DSM-6.”