News The latest developments in autism research.
Profiles Portraits of scientists who are making a mark on autism research.
Toolbox Emerging tools and techniques that may advance autism research.
Spotted A roundup of autism papers and media mentions you may have missed.
Opinion Conversations on the science of autism research.
Viewpoint Expert opinions on trends and controversies in autism research.
Columnists Dispatches from experts on various facets of autism.
Crosstalk Debates and conversations about timely topics in autism.
Reviews Exploring the intersection of autism and the arts.
Q&A Conversations with experts about noteworthy topics in autism.
Deep Dive In-depth analysis of important topics in autism.
Special Reports Curated collections of articles on special topics in autism.
Webinars Presentations by leading experts on their latest research.

Proposed guidelines won’t miss autism cases, study says

by  /  1 October 2012

Better measures: The new guidelines for diagnosing autism are as accurate at identifying the disorder as are the existing guidelines.


The proposed changes to the diagnostic criteria for autism are unlikely to exclude many people currently diagnosed with Asperger syndrome or pervasive developmental disorder-not otherwise specified (PDD-NOS), according to a large analysis published today in the American Journal of Psychiatry1.

The study of more than 5,000 people found that the new criteria accurately identified 91 percent of people with autism-related diagnoses under the existing criteria, and suggests that the new version can better distinguish autism from other developmental disabilities.

The new criteria are outlined in the DSM-5, the forthcoming edition of the Diagnostic and Statistical Manual of Mental Disorders, due out in May 2013.

The reports generated a great deal of controversy in both the general media and scientific journals because in many cases, losing a diagnosis could also result in a loss of benefits and services.

Careful collection:

A major criticism of some previous studies of the DSM-5 is that the data they used, often collected from brief questionnaires, do not include information required by the DSM-5 criteria. For example, the DSM-5 criteria include sensory problems, which are not part of the DSM-III or -IV.

“Some of the studies in which children fail to meet criteria are poor in quality,” says David Skuse, chair of behavioral and brain sciences at University College London, who was not involved in the new study. “They are nowhere near as carefully done as this one, in terms of the number of cases and the range of clinical material they used.”

The new study analyzed 4,453 individuals with an autism spectrum disorder and 690 people with other diagnoses, such as language disorders or attention deficit hyperactivity disorder.

“What we were trying to do is systematically look at any large samples we could get our hands on,” says lead investigator Cathy Lord, director of the Institute for Brain Development at New York-Presbyterian Hospital in New York City.

The study includes 1,021 children from the Collaborative Programs of Excellence in Autism, a multi-site effort funded by the National Institutes of Health, and 1,992 children from Lord’s former clinic at the University of Michigan. It also includes 2,130 children from the Simons Simplex Collection — a collection of families that have one child with autism but unaffected parents and siblings, which is sponsored by’s parent organization.

“In combination, it’s a much bigger sample than anyone else had and gives us more data,” says Lord, who is also a member of the committee revising the DSM-5 autism guidelines.

Each of the children had undergone the Autism Diagnostic Interview-Revised, or ADI-R, a parent report of autism symptoms, and the Autism Diagnostic Observation Schedule, or ADOS, an hour-long clinical observation tool. (Lord developed these tests.)

Though these are detailed tests, neither one was developed explicitly to collect the information needed for diagnosis using the DSM-5.

“It’s still not the same thing as taking the new criteria and testing them out, which is why we didn’t do this analysis before,” says Lord. “But clearly people have been analyzing much more restricted datasets, so we thought we better get in here and do it.”

Lord and her colleagues found that the DSM-5 is as sensitive as the DSM-IV, meaning it accurately identifies those who have autism. The DSM-5 criteria also have better specificity than those in the DSM-IV, meaning they can better distinguish between people who have autism and those who have other developmental disorders, the study found.

Real world:

In the DSM-5 criteria, symptoms of autism are split into two main categories: social communication deficits and restricted and repetitive behaviors. One of the major concerns with the DSM-5 has been whether it will exclude some people with PDD-NOS. Some studies suggest that people with PDD-NOS have fewer restricted and repetitive behaviors than those with other autism diagnoses2.

The new paper refutes these concerns, showing that only a small percentage of people with PDD-NOS fail to meet the new guidelines, and only because of fewer social communication symptoms.

“We didn’t see any evidence that there would be dramatically lower diagnosis of people with Asperger’s or PDD-NOS,” says Lord.

Skuse says he is surprised by this finding, because in his own clinical sample, he finds more people who have social communication deficits but not restricted and repetitive behavior problems than the reverse pattern.

Fred Volkmar, director of the Yale Child Study Center and a vocal critic of the DSM-5, says he is concerned that the study’s setting does not represent the real world.

“The trouble is, these are very well-characterized cases in research studies with high-level experts, so you’re getting a rosy scenario,” says Volkmar. “In that scenario, the [guidelines] work well. But a book like the DSM is meant for clinical use as well.”

For example, not all clinicians use the ADOS and ADI-R — time-intensive tests that require extensive training to administer — to diagnose autism.

Lord responds that the same concern applies to the existing guidelines as well. “I think it’s true that we do need to think about how we expect non-specialist clinicians to make diagnoses — that’s a whole other issue, but was just as relevant for DSM-IV,” she says. “Even very skilled clinicians had difficulty using the DSM-IV, in terms of the distinctions among the different types of [autism].”

The best solution to these concerns is to educate providers and improve tools, others say.

“I don’t think the answer to [Volkmar’s] criticism is that [the criteria] have to be relaxed,” says Frazier. “Finding a tool that doesn’t take so much time and expertise is important.”

The debate is likely to linger until the DSM-5 criteria have been evaluated in so-called prospective trials — studies that collect data expressly for the purpose of comparing the DSM-IV and DSM-5 — rather than relying on previously collected data.

Autism Speaks, a research and advocacy organization based in New York, is funding one such study. Researchers at the Medical University of South Carolina plan to screen 8,000 school children for autism risk, and then use both DSM-IV and DSM-5 criteria to assess those considered to be at risk.

“It will be really helpful to see what the Autism Speaks prospective study will turn out, but that will be years from now,” says Lord.


1: Huerta M. et al. Am. J. Psychiatry 169, 1056–1064 (2012) PubMed

2: Mandy W. et al. Autism Res. 4, 121-131 (2011) PubMed

  • Sarah

    I think a very large percentage of kids dxed with autism are actually physically ill. Their immune systems are overwhelmed and the condition is chronic. Neuroinflammation/ over activation in the brain results in autistic behavior. Autism is primarily a medical condition specifically an immune dysfunction not a psychiatric one. The physical illness manifests with OCD-like behavioral symptoms. There is plenty of scientific evidence to support that autism is a physical illness. Our kids often present with physical problems like inflammation in the GI and brain, dysbiosis, malabsorption, diarrhea etc. It’s time for professionals in the medical community including psychiatrists to ackowledge the underlying physical illness that results in autism. Autism belongs in the medical realm. Our kids need medical treatment to help them function better.

    • Carmen

      I am the mom of a PDD-NOS child and he is not physically ill. Actually we ran all kinds of tests to ensure that. I was given the dxed by four doctors in different fields. I have to wonder if you have a child or know a child with this disorder to base your opinion on. I tell you this in hopes that you will understand that not all parents jump to accepting this diagnoses at face value.

  • RAJensen

    There are two Diagnostic Manuals currently under revision, DSM-5 and ICD-11. In the past the working groups for autism in both groups collaborated closely to insure that both manuals would be consistent with respect to diagnostic criteria and reach agreement on the publication dates for both manuals. This has not happened this time. The announced publication dates are not the same, DSM-5 will be published a year or more earlier than the new ICD-11 manual. There is a dispute between these two groups. Sir Michael Rutter, the chair of the ICD-11 working group on autism has explained the difference:

    ‘At the moment there are important differences between proposals for DSM-5 and ICD-11. For the most part, there is broad agreement on the overall concepts but there are difficulties with respect to the details. That arises most especially because the DSM-5 starts with dealing with research criteria before considering the concepts and WHO does it the opposite way round. That is to say, the starting point with WHO is the clinical conceptualisation and the clinical criteria. At a later point, of course, research criteria have to be developed but that comes secondarily. In my view, that is the most appropriate way round’

    The DSM-5 working group is reflecting the views of the majority who are behavioral geneticists who see autism as dimensional rather than categorical. Unlike ICD which is a medical manual published by the World Health Organization, DSM is a manual of mental disorders. Many members of the DSM5 working group on autism are not medical doctors but rather psychologists (Francesca Happe, Catherine Lord, Sally Rogers, Sarah Spence, and Amy Wetherby) whose backgrounds suggest a bias towards seeing autism through their own research.

    Thomas Frazier who also commented on this study is also not a medical doctor but holds a PHD.

  • Harold L Doherty

    This article appears to suggest that persons who would meet DSM-IV PDD-NOS and Asperger’s will “only” be reduced by approximately 10% under DSM5 criteria. The focus, as always, is on the HF end of the spectrum with no mention made of the intellectually disabled who will be excluded under the wording of mandatory criterion A of the DSM5.

    “We didn’t see any evidence that there would be dramatically lower diagnosis of people with Asperger’s or PDD-NOS,” says Lord.”

    Catherine Lord has previously acknowledged that the real targets for exclusion from the DSM5’s New Autism Spectrum Disorder are the intellectually disabled. Persons with ID represented “the vast majority” of persons with autistic disorder according to CDC autism expert Dr. Yeargin-Allsopp. The DSM-IV addition of PDD-NOS and Aspergers reduced that figure to 41-44% according to recent CDC surveys. The DSM5 exclusion under Criteria A for social communication even where EVEN if all Critera A categories are otherwise exhibited will result in a further significant reduction in numbers of person with autism and ID. And that is the real aim of the DSM5 as Catherine Lord again confesses:

    “Lord and her colleagues found that the DSM-5 is as sensitive as the DSM-IV, meaning it accurately identifies those who have autism. The DSM-5 criteria also have better specificity than those in the DSM-IV, meaning they can better distinguish between people who have autism and those who have other developmental disorders, the study found.”

    The real targets for exclusion from the autism spectrum under the DSM5 autism do-over are the intellectually disabled who are targeted by the addition of the “not accounted for by general developmental delay” disqualifying criterion in mandatory criterion A. Studies by J Matson have confirmed that substantial numbers of intellectually disabled will be excluded. That exclusionary criteria exists in slightly different language in the DSM-IV for Asperger’s.

    In the DSM5 the evolution of autism into Aspergers is under way with the exclusion of the intellectually disabled. But no one cares. Not Dr. Lord, not Dr. Geraldine Dawson of Autism Speaks, nor the New York Times and other major media. The intellectually disabled apparently are not worthy of concern.

  • Shree Vaidya

    Thank you.Yes. One of the criterias that causes to Autism is due to low quality sensors of the physical body.

  • Chris Autistic

    The new DSM-V criterias seems to more fit the diagnosis originally observed by Leo Kanner (1943) and Hans Asperger (1944). Kanner’s work resulted in autistic disorder wherease Asperger’s work resulted in Asperger’s syndrome. Their findings, at the level of criterias, was nearly the same and it is questionable why it exist as two distinct diagnosis; questions which have been in research since the findings, and which DSM-V seems to have taken into account.

    Leo Kanner did not directly describe the language in autism as delayed. He described it as deviant in a social way: “Language – which the children did not use for the purpose of communication – was deflected in a considerable measure to a self-sufficient, semantically and conversationally valueless or grossly distorted memory exercise” (Kanner, 1943, p. 243). Neither did he describe them as intellectually disabled: “They are all unquestionable endowed with good cognitive potentialities” (Kanner, 1943, p. 247).

    To have an autism diagnosis with no base in it’s etymology or ontology is to have an diagnosis which could happen to be named with any word. It does not describe what it was, and it will clearly lead to confusion when someone use the word autism in its ontological view where others do not. The DSM-V is the best I’ve seen of criterias who has an ontological base, and which include me with rich description of my disorder through both historical and modern research.

    Asperger, H. (2001). ‘Autistic psychopathy’ in childhood. In Uta Frith (Ed.), Autism and Asperger syndrome. (p. 37-93). UK: Cambridge University Press.

    Kanner, L. (1943). Autistic disturbances of affective contact. The Nervous Child 2(3), 217-250.

  • Katie Weisman

    The new study on the proposed DSM-5 ASD criteria is only mildly reassuring to me. While the methodology is sound, the population on which it was based was not reflective of the full spectrum of autism. The study used well–characterized existing pools of children and adolescents, but the pools contained 72% individuals who met the DSM-IVR criteria for autistic disorder, those typically most challenged on the spectrum. 22% of cases had PDD-NOS and 6% had Asperger’s Syndrome. By comparison, the CDC’s ADDM analysis from April showed 44% with autistic disorder, 47% with PDD-NOS and 9% with Asperger’s. Since the previous five studies this spring all showed more significant drops in diagnoses for higher-functioning individuals, one would expect that the actual impact of the new criteria will be more than a 9% decrease. And, of course, we still have no data on adults.

  • ASD Dad

    The use of DSM continues to be a crude and almost irrelevant tool as we move to a closer understanding that psychology, physiology and genetics are generally interwoven in most circumstances.

    We need a far better and more sophisticated approach to autism. Unlocking the varying aspects that strongly suggest differing phenotypes of autism ie as referred to recently by Thomas Insel as The Autisms.

  • Natasa

    I completely agree with Sarah. What we call ‘autism’ is just a surface manifestation of underlying biological pathology. In an ideal world (in not too distant future I hope!) autism would be completely removed from DSM and ‘autism symptoms’ would simply be viewed and TREATED as consequences of pathopysiological processes.


Log in to your Spectrum Wiki account

Email Address:



Request your Spectrum Wiki account

Spectrum Wiki is a community of researchers affiliated with an academic or research institutions. To be considered for participation, please fill out this form and a member of our team will respond to your request.


Email Address:

Title and Lab:

Area of Expertise: