The Shifting Goalposts of Autism
Jo Craven McGinty at the Wall Street Journal:
The latest report by the U.S. Centers for Disease Control and Prevention uses two sets. One shows a growing number of children with autism. The other, when strictly applied, identifies 18% fewer cases.
For now, the report, which was published last week, gives priority to the looser standard, but when the CDC issues its next report two years from now, the more stringent measure will take precedence.
The current report tracked the estimated prevalence of autism in more than 300,000 children who were 8 years old in 2014. Statistically, the children, who reside in 11 communities in as many states, aren’t nationally representative, but the findings inform treatment and policy decisions across the country.
The conclusions are based primarily on criteria drafted in 1994 for the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders. [DSM] Using those standards, 1 in 59 children had autism, compared with 1 in 68 just two years earlier.
For the sake of comparison, the report assessed 263,775 children using both the old and new standards, which were drafted in 2013.
Based on the old standards, 4,658 were autistic. With the new criteria, 4% fewer, or 4,498, were—but that number includes 681 on the basis of a previous diagnosis. Without them, only 3,817 children would have been diagnosed, or 18% fewer than the number identified with the old rules.
It’s possible some of those kids would meet the new criteria, but the documented record of their symptoms didn’t support that conclusion.
I
write in The Politics of Autism:
Dr. Allen Frances, the lead editor of previous version of the manual, says that the diagnostic instructions are confusing, so the diagnosis will vary from rater to rater and place to place. “It will be even more impossible than it is now to determine rates of autism and why they shift so much over time.” A more recent study says that prevalence estimates are likely to decline under DSM-5 – but it cautions that policies for service eligibility and diagnostic tools will change in response to the revised criteria. Clinicians may react by looking for more symptoms to back up ASD diagnoses.
And DSM-5 is not the last word. Whereas the titles of previous editions included Roman numerals (DSM-III, DSM-IV), the new one instead has an Arabic numeral to make it easier to designate revisions. The future will bring DSM-5.1, DSM-5.2, and so on, which may include further changes in the way we describe and diagnose autism.
