Autism Rates Are About To Drop

Under the new rules for diagnosing autism, fewer people may be diagnosed with the condition than before, a study in the Jan. 22 JAMA Psychiatry suggests.

For example, in a group of 6,577 8-year-olds classified as autistic under the old guidelines 18.8% would probably not be classified as autistic under the new guidelines.

The rules actually changed on May 22nd, 2013, when the American psychiatric community switched to a new handbook for diagnosing mental disorders, the DSM-5.

The book sets guidelines for diagnosing every psychiatric condition, including ADHD, depression, schizophrenia, autism, and more.

The findings are significant because the DSM is also used by schools to determine a child's special education status, insurance companies to assess eligibility for services, and the legal system to evaluate defendants.

People are only beginning to use the DSM-5, so we won't really know what the effect of the new autism guidelines are for a while. But the new study offers some early clues.

The new guidelines

Autism involves some subtle behaviors and can be difficult to diagnose. But most people diagnosed with autism under the DSM-IV would still be diagnosed with ASD using the DSM-5, the new study and previous research have found.

The diagnostic criteria for autism/ASD are too long and complex to catalogue here, but there are a few key changes between the DSM-IV and the DSM-5:

• Subcategories for things like Asperger's have been eliminated. A new category, "social communication disorder," has been added.
• Symptoms don't have to be present at the time of assessment; they count if they were exhibited in the past, too.
• Patients must show problems with three specific types of social communication, including nonverbal communication. (Previously a diagnosis could include just two types, and there was more of a specific emphasis on language.)
• Patients must show "restricted, repetitive patterns of behavior, interests, or activities" of at least two types. One type can be "hyper- or hyporeactivity to sensory input," something often associated with autism but not previously included.

No one who was diagnosed with autism under the old guidelines will have their diagnosis taken away; the DSM-5 guidelines are meant to shape assessments going forward. But that creates a problem when looking at how many people have autism over time - there might be a drop in rates because of the change in diagnostic criteria.

The rate of autism in the U.S.

As we struggle to get a handle on autism, tracking how many children are affected is important. Over the last few decades, rates have skyrocketed, and we don't know why. Is it just that the syndrome is getting more attention, or is there something in our environment, our genes, or both that's driving the increase?

For example, the prevalence of autism rose by 78% between 2002 and 2008. The prevalence of autism in 2008 - using the old guidelines - was 1 in 88; using the new guidelines, the researchers conclude, it would have been 1 in 100, a significantly lower rate.

But the DSM-5 authors weren't trying to lower the autism rate with the new rules. They were simply trying to make it easier and clearer to diagnose. They actually didn't expect any "significant changes in the prevalence of the disorder."

The JAMA Psychiatry study, which used data from the CDC's Autism and Developmental Disabilities Monitoring Network, is the largest and most diverse effort to see if they were right.

Testing rates

To count the number of autistic kids in the study population, the researchers applied both the old and the new diagnostic criteria to a large pool of medical and school records.

Only 81.2% of those who were classified as autistic under the DSM-IV would still be classified as autistic under the new guidelines. Interestingly, an even larger majority - 96.1% - of that first group were within just one criterion of meeting the terms of the DSM-5 diagnosis.

Study author Matthew J. Maenner, an epidemiologist at the CDC's National Center on Birth Defects and Developmental Disabilities, says that as clinicians adjust to the new criteria they may watch for relevant symptoms that could have been overlooked with the old rules, closing the diagnosis gap.

Even with a lower rate of in autism diagnoses using the DSM-5, the upward trend in the prevalence of autism was clear whether the researchers applied the DSM-IV criteria (diamonds) or the DSM-5 criteria (circles):

The study's limitations

"I applaud the CDC's effort to work on this," Young-Shin Kim, a child psychiatrist and epidemiologist at the Yale University School of Medicine's Child Study Center, tells Business Insider. "But because of the way they did the study, they raised more questions than answers." (Kim is set to publish a study on the same topic soon.)

Analyzing the paper records of a large and diverse group of people let the researchers draw conclusions at a population level, but it did not capture how autism is diagnosed in real life: in person.

Those methods might explain, suggests Matthew S. Siegel, of Spring Harbor Hospital and the Tufts University School of Medicine, why so few children met the DSM-5 criterion for "nonverbal communication" deficits: That's a hard symptom to detect in paper records.

The study also compared the rate of DSM-IV and DSM-5 diagnoses side by side, even though they calculated them differently. To make a DSM-IV diagnosis, researchers relied on the final judgement of a clinician who looked over the patient records. But for the DSM-5 diagnosis, they tried to match up that same data with the DSM-5 criteria automatically, without a clinician making the final call.

"It's like comparing apples and oranges," says Kim.

The study actually came to two different conclusions for the the two years of data analyzed. In 2006, using the DSM-5 criteria would have reduced prevalence by 18%, but for 2008, the reduction was just 12%. (This could be due to greater awareness, screening, and more detailed records, Maenner suggests, making it easier for the researchers to match a 2008 child to the DSM-5 criteria.)

Ultimately, while the JAMA Psychiatry study is an important analysis, we probably will not know the full effect of the new diagnostic criteria until they have been in effect for many years.