You wouldn’t think that a new edition of the Diagnostic and Statistical Manual of Mental Disorders would create the kind of buzz surrounding a hotly anticipated film or novel. But when it comes to controversy, the DSM-5 is a blockbuster.

Though the latest edition of the manual is not due out until 2013, draft criteria for autism have been posted online, spurring furious debate.

Critics argue that the proposed criteria may lead to some individuals with Asperger syndrome losing their diagnosis and, as a result, their access to services, such as specialized programs in public schools.

A new epidemiological study shows that those concerns are justified. Of 26 children with intelligence quotients (IQs) above 50 diagnosed with autism by criteria in the DSM-IV-TR, the current edition of the manual, the proposed criteria miss more than half. Those children are clustered at the higher functioning end of the spectrum, just as critics fear.

In fact, the study found that the more severe the autism and the lower the IQ, the more likely it is that a child will meet DSM-5 criteria.

A few months ago, when I talked with Darrel Regier, vice-chair of the American Psychiatric Association’s DSM-5 task force, he said that the objections to merging Asperger syndrome into the autism spectrum are “in some ways more sociological than scientific.”

Regier was referring to the fact that some individuals with Asperger syndrome are proud of their identity and don’t like the idea of being lumped together with lower-functioning individuals. He also said that some in the Asperger syndrome focus group thought that they might have an easier time accessing services with an autism diagnosis.

The new study calls both of those assumptions into question.

The study was carried out in Finland, where researchers asked parents of more than 4,000 8-year-old children to fill out the Autism Spectrum Screening Questionnaire. About three percent, or 125 children, screened positive and entered the diagnostic phase of the study.

After excluding those with physical or intellectual disabilities, researchers assessed 82 children using the Autism Diagnostic Interview-Revised, the Autism Diagnostic Observation Schedule, the Wechsler Intelligence Scale for Children, clinical records — and, in the case of 24 children, in-school observation.

When the researchers used DSM-IV-TR criteria, they diagnosed 37 of the 82 children with autism for an overall prevalence of 8.4 in 1,000, consistent with official estimates.

But when they applied the draft DSM-5 criteria, 12 of the 26 children with IQs above 50, including 11 with Asperger syndrome, no longer merited the diagnosis. The criteria also excluded another seven children with IQs below 50.

The average IQ of the children who fit the DSM-5 autism diagnosis is significantly lower than that of others in the study. This suggests that the proposed criteria are biased toward identifying lower-functioning individuals.

The researchers propose five modifications to the criteria that they say would address the problem.

For example, they say, changing “deficits in verbal communication AND non-verbal communication” to “deficits in verbal AND/OR non-verbal communication” would include more children. The children who are being missed use normal gestures, facial expressions and vocabulary, but can’t carry on a real conversation.

Changing the requirement for “unusual and abnormal idiosyncratic sensory behavior” to “unusual OR abnormal sensory behavior” would also help. It can be difficult to distinguish the unusual from the abnormal, setting the bar quite a bit higher in this category than it needs to be.

Perhaps most importantly, the emphasis on the appearance of symptoms of autism before 36 months of age excludes high-functioning individuals who don’t experience difficulties until they begin attending school and encounter greater social demands. The researchers suggest replacing the defined age limit with a more general requirement that symptoms must be present in childhood and worsen when social demands exceed the child’s capacity to meet them.

The DSM is often referred to as the ‘bible’ of psychiatry. In our conversation last spring, Regier said that’s wrong. “It’s not inerrant scripture on stone tablets. It’s a set of diagnostic hypotheses subject to research, review and disapproval.”

Research and review on DSM-5 have begun. Apparently now is the time for disapproval.