An artificial intelligence-based medical device shows promise in diagnosing cases of autism spectrum disorder in kids 18 months to 3 years old, researchers from six U.S. states say.
The software-as-a-medical-device tool uses a machine-learning algorithm that employs data from parent and provider questionnaires and home videos. For about a third of the 425 study participants, the technology made timely and accurate autism diagnoses, the May 5 study in NPJ Digital Medicine found.
While the study authors say replication and follow-up research is needed, devices like these could be valuable in a system rife with diagnostic delays at a crucial stage of development and intervention for children with autism, they said. The average age of diagnosis is over 4 years old, the study noted.