The newest version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), to be published in May 2013, has sparked much controversy. Critics say the new version may miss many people now diagnosed with autism, but studies in the past few weeks have found otherwise.
A study published 26 September in the Journal of Autism and Developmental Disorders finds that which tool a clinician uses to diagnose a child with autism may determine whether that child meets the DSM-5 criteria.
The new study is part of an ongoing effort to determine how well the DSM-5 criteria detect autism in those diagnosed using the current version, the DSM-IV.
Two tools commonly used to diagnose autism are the Autism Diagnostic Observation Schedule (ADOS), a 45-minute test in which a psychologist observes a child for specific behaviors, and the Autism Diagnostic Interview-Revised (ADI-R), which asks a parent or caregiver about the child’s behaviors.
In the new study, researchers analyzed data from nearly 500 people with an autism diagnosis, ranging in age from 5 to 61 years. All had participated in previous studies, completing the ADI-R and relevant sections of the ADOS. The study focused on people with high-functioning autism, and the group had an average intelligence quotient of 105.
According to the findings, using the ADOS alone, only 33 percent of participants met the criteria, whereas 83 percent met them when using the ADI-R alone. Using both tests together boosted that number to 93 percent.
As the researchers note, this discrepancy between the ADOS and ADI-R isn’t new — it’s evident for the DSM-IV as well, although it appears to be greater for the DSM-5.
Lead investigator Carla Mazefsky, director of the Center for Excellence in Autism Research at the University of Pittsburgh, says the low numbers for the ADOS alone may result from the fact that it’s often difficult to detect repetitive behaviors — one of the two main categories of symptoms in the DSM-5 — during a 45-minute observation.
The new figures are somewhat lower than those reported last week by Catherine Lord and her collaborators in the American Journal of Psychiatry. That study found that 91 percent of participants meet DSM-5 criteria using data from parent reports, and 99 percent meet them when both parent and clinical measures are used.
The discrepancy may stem in part from the fact that the new study focused entirely on people with strong language skills, whereas participants in the Lord study span the range of verbal abilities.
When Lord’s team focused on people older than 10 years with fluent language, they found 81 to 90 percent accuracy, depending on the clinical population they studied. For people with a verbal intelligence score greater than 70, similar to the population in the new study, sensitivity ranged from 86 to 91 percent.
The two studies also used slightly different methods of determining whether an individual meets DSM-5 criteria.