- Researchers at Stanford University are trying to determine whether an mHealth platform that combines AI technology with home movies can be used to identify children with autism spectrum disorder.
If proven reliable, the telehealth platform could be used to screen children more quickly and at home, bypassing issues with access to pediatric specialists or clinics and enabling both providers and parents to begin treatment earlier.
The process gains added value when noting that autism diagnoses have jumped 700 percent since 1996 and now affect one in every 59 US children. Coupled with a shortage of specialists, this means the average wait time to have a child properly screened can be longer than a year.
“The sharp rise in incidence of autism, coupled with the un-scalable nature of the standard of care (SOC), has created strain on the healthcare system, and the average age of diagnosis remains around 4.5 years, two years past the time when it could be reliably diagnosed,” the study, published online at PLOS, reported. “Mobile measures that scale could help to alleviate this strain on the healthcare system, reduce waiting times for access to therapy and treatment, and reach underserved populations.”
The researchers - Qandeel Tariq, Jena Daniels, Jessey Nicole Schwartz, Peter Washington, Haik Kalantarian and Dennis Paul Wall, all of Stanford’s Department of Pediatrics and Department of Biomedical Data Science – crowdsourced more than 160 home videos of children with and without ASD, averaging two minutes. They then created a web portal by which experts could asses the children for 30 behavioral features used by eight independent machine learning models for identifying ASD.
That screening process, they reported, identified ASD in children, as opposed to those with either typical or atypical development, with greater than 90 percent accuracy.
Adding AI technology, the researchers said, enables healthcare providers to create a “growing matrix of video features” that will allow them to improve ASD diagnosis from home videos over time. In addition, the platform could evolve to the point that non-specialists could run the program.
“These results support the hypothesis that the detection of autism can be done effectively at scale through mobile video analysis and machine learning classification to produce a quantified indicator of autism risk quickly,” they concluded. “Such a process could streamline autism diagnosis to enable earlier detection and earlier access to therapy that has the highest impact during earlier windows of social development.”
“Further, this approach could help to reduce the geographic and financial burdens associated with access to diagnostic resources and provide more equal opportunity to underserved populations, including those in developing countries.”