Rutgers researchers: Tests for autism aren’t reliable enough
Rutgers researchers say a standardized test widely used to diagnose whether children have autism is not as reliable as they once thought.
Associate psychology professor Elizabeth Torres said the Autism Diagnostic Observation Schedule test is the gold standard of autism diagnosis used in both clinical and research domains. IT was developed by psychologists led by Cathy Lord, a psychologist at UCLA, and is owned by Western Psychological Services.
Torres said ADOS tests social interactions and communications skills between a child and a clinician. There are toys and tasks that the child has to perform in front of the clinician. Toress said a score determined by the clinician determines where a child lies on the autism disorder spectrum.
The problem with the test, she said. is that it depends on the clinician -- switching ADOS-certified clinicians can change a child's score and dramatically influence the diagnosis.
"For research purposes, it is not really adequate and leaves a lot to be desired," Torres said.
She said researchers found that digitizing the test, making the child wear biosensor devices such as Fitbits or Apple Watches, increased the level of precision when it came to diagnoses. The signals that the biosensors put out help find what clinicians may miss with the naked eye.
The tests might find a child is anxious because of his or her heart rate, or show the child is experiencing stress, she said
Torres said with biosensor data, it does not matter what lab a child is observed in for the autism test. Torres said this kind of data collection is less invasive, lowers the rate of false positives, shortens the time to diagnosis and makes diagnoses more objective and more reliable for clinicians.
More from New Jersey 101.5: