Using YouTube videos, researchers say they’ve identified a new approach to autism screening that could significantly speed up the path to diagnosis.

Short home movies may be sufficient to accurately spot children who have the developmental disorder, according to a new study from Harvard Medical School and the Stanford University School of Medicine published this week in the journal PLOS ONE.

For the study, researchers found 100 videos on YouTube, each of which were 10 minutes or less and showed kids ages 1 to 15 playing. Of the clips, 45 were identified by their creators as depicting children with autism, while the other films showed children without the condition.

Advertisement - Continue Reading Below

A group of college students were then trained to assess the behavior of the children in the videos using a rating scale based on the Autism Diagnostic Observation Schedule, which is often used to diagnose autism. Despite having limited training, the students were able to accurately distinguish videos portraying children with autism 97 percent of the time, the study found.

“Our new paper supports the hypothesis that we can detect autism quickly in very short home videos with high accuracy,” said Dennis Wall, an associate professor of pediatrics at Stanford and the study’s senior author.

Wall indicated that it’s unlikely that videos would completely replace the traditional approach to clinical evaluation for an autism diagnosis, but said the technique could offer a speedy way to identify or monitor kids likely to be on the spectrum so that they can begin intervention while waiting for a formal diagnosis.

Experts say that spotting autism as early as possible is key since treatment is often more effective the sooner it begins. At present, however, most kids are not diagnosed until after age 4 even though the developmental disorder can be reliably detected by age 2, according to the U.S. Centers for Disease Control and Prevention.