Only 35% of a sample of children diagnosed with ASD before age 3 based on DSM-IV criteria retained the diagnosis when DSM-5 criteria were applied.
"The strict nature of the criteria in a population whose symptomatology may be emerging is at odds with an early diagnosis model," said Dr. Lisa H. Shulman, director of the Rehabilitation, Evaluation and Learning for Autistic Infants and Toddlers program at the Albert Einstein College of Medicine, New York.
Studies of the impact of the new criteria on diagnosis of ASD have focused on school-age children. Dr. Shulman and her coinvestigators looked at younger children, examining all the children who had been diagnosed with ASD by 3 years of age based on a multidisciplinary evaluation at their center during 2003-2010. Diagnoses were based on DSM-IV criteria. Childhood Autism Rating Scale (CARS) scores also were examined, however, and cognitive testing was completed in some of the children.
To determine how many of the children would retain the diagnosis using DSM-5 criteria, an algorithm was used to map data from the DSM-IV and the CARS scores onto the DSM-5 criteria. Of 237 children who had been diagnosed with ASD by age 3 at the inner-city early intervention program, only 84 children (35%) met the criteria for ASD using the DSM-5 criteria.
The children whose diagnosis was retained using DSM-5 criteria were significantly more likely to have had higher CARS scores (38.3 vs. 33.5), which are indicative of more severe social impairment. They also had a trend toward lower cognition, with 34% of those with an IQ of less than 70 diagnosed under DSM-5 criteria, compared with 12% with IQ greater than 70, Dr. Shulman reported.
The potential of the DSM-5 criteria to diagnose the more impaired children preferentially might present an obstacle to diagnosis for the children most likely to benefit from early intervention – those with the mildest symptomatology, she said