Proposed changes in the definition of autism would sharply reduce the skyrocketing rate at which the disorder is diagnosed and may make it harder for many people who would no longer meet the criteria to get health, educational and social services, a new analysis suggests.

The definition is under review by an expert panel appointed by the American Psychiatric Association, which is completing work on the fifth edition of its Diagnostic and Statistical Manual of Mental Disorders. The DSM, as the manual is known, is the standard reference for mental disorders, driving research, treatment and insurance decisions.

The study results, presented Thursday at a meeting of the Icelandic Medical Association, are still preliminary but offer the latest and most dramatic estimate of how tightening the criteria for autism could affect the rate of diagnosis. Rates of autism and related disorders like Asperger's syndrome have taken off since the early 1980s, to prevalence rates as high as 1 in 100 children in some places. Many researchers suspect that these numbers are inflated because of vagueness in the current criteria.

"The proposed changes would put an end to the autism epidemic," said Fred Volkmar, director of the Child Study Center at Yale University School of Medicine and an author of the new analysis. "We would nip it in the bud — think of it that way."

Experts working on the new definition strongly questioned the new estimate.

"I don't know how they're getting those numbers," said Catherine Lord, a member of the task force working on the diagnosis.

Previous projections have concluded that far fewer people would be excluded under the proposed diagnosis change, said Lord, director of the Institute for Brain Development, a joint project of NewYork-Presbyterian Hospital, Weill Cornell Medical College, Columbia University Medical Center and the New York Center for Autism.

Disagreement about the effect of the new definition will almost certainly increase scrutiny of the finer points of the psychiatric association's changes to the manual The revisions are about 90 percent complete and will be final by December, according to David Kupfer, a professor of psychiatry at the University of Pittsburgh and chairman of the task force making the revisions.

At least 1 million children and adults have a diagnosis of autism or a related disorder, like Asperger's syndrome or "pervasive developmental disorder, not otherwise specified" — or PDD-NOS. People with Asperger's or PDD-NOS endure some of the same social struggles as those with autism but do not meet the definition for the full-blown version. The proposed change would consolidate all three diagnoses under one category, autism spectrum disorder, eliminating Asperger's syndrome and PDD-NOS from the manual. Under the current criteria a person can qualify for the diagnosis by exhibiting six or more of 12 behaviors; under the proposed definition, the person would have to exhibit three deficits in social interaction and communication and at least two repetitive behaviors — a much narrower menu.

Kupfer said the proposed changes were an attempt to clarify these permutations and put them under one name.

Hundreds of thousands of people receive state-backed special services to help offset the disorders' disabling effects, which include learning and social problems, and the diagnosis is in many ways central to their lives. Close networks of parents have bonded over common experiences with children; and the children, too, may grow to find a sense of their own identity in their struggle with the disorder.

Mary Meyer, of Ramsey, N.J., said that a diagnosis of Asperger syndrome was crucial in getting her daughter, who is 37, access to services that have helped tremendously.

"I'm very concerned about the change in diagnosis because I wonder if my daughter would even qualify now," she said. "She's on disability, which is partly based on the Asperger's, and I'm hoping to get her into supportive housing, which also depends on her diagnosis."

Mark Roithmayr, president of Autism Speaks, an advocacy organization, said that the proposed diagnosis should bring needed clarity but that the effect on services was not yet clear.

"We need to carefully monitor the impact of these diagnostic changes on access to services and ensure that no one is being denied the services they need," Roithmayr said by email. "Some treatments and services are driven solely by a person's diagnosis, while other services may depend on other criteria such as age, IQ level or medical history."

In the new analysis, Volkmar, along with Brian Reichow and James McPartland, both at Yale, used data from a large 1993 study that served as the basis for the current criteria. They focused on 372 children and adults who were among the highest-functioning and found that overall, only 45 percent of them would qualify for the proposed autism spectrum diagnosis now under review. The focus on a high-functioning group may have slightly exaggerated that percentage, the authors acknowledge.

The likelihood of being left out under the new definition depended on the original diagnosis: About a quarter of those identified with classic autism in 1993 would not be so identified under the proposed criteria; about three quarters of those with Asperger's would not qualify; and 85 percent of those with PDD-NOS would not.

Volkmar presented the preliminary findings Thursday. The researchers will publish a broader analysis, based on a larger and more representative sample of 1,000 cases, later this year. Lord said that the study numbers are probably exaggerated because the research team relied on old data, collected by doctors who were not aware of what kinds of behaviors the proposed definition requires.

"It's not that the behaviors didn't exist, but that they weren't even asking about them — they wouldn't show up at all in the data," Lord said.