The FDA has granted authorization to the Cognoa ASD Diagnosis Aid, an AI-based software aimed to help doctors diagnose autism spectrum disorder.
Autism Spectrum Disorder (ASD) is a developmental disorder that leads to significant social, behavioural, cognitive and communication delays. According to the CDC, it affects 1 in 59 children in the United States. Although symptoms usually appear within the first two years of life, diagnosis is often delayed because of varying severity. Moreover, many children, especially girls, present with less obvious signs of autism. Thus, delaying treatment and early interventions.
There is currently no medical test for ASD. In the past, researchers have tested eye scans for an early diagnosis of the disorder. However, current methods involve screening children for early signs of ASD. These often lead to misdiagnosis and delayed diagnosis.
But a digital health startup, Cognoa is looking to change all that. A team of researchers at the company have designed an AI-based software that can help paediatricians diagnose cases. Called the Cognoa ASD Diagnosis Aid, the device combines data from questionnaires and home videos to evaluate patients at risk of ASD.
Last month, the US Food and Drug Administration (FDA) authorized the marketing of the device. They based their approval on a study of 425 patients that compared assessments made by the device and that made by current standard diagnostic criteria.
How Does it Work?
The Cognoa ASD Diagnosis Aid is made up of three components: a mobile app, a video analysis portal, and a health care provider portal. Parents and caregivers can use the mobile app to upload videos of their child and answer questionnaires regarding the child’s behavioural problems. Specialists then analyze these uploaded videos using the portal. Next, a healthcare provider answers pre-loaded questions on behaviour problems and review the information provided by parents or caregivers. Lastly, the algorithm processes all the provided information and gives a positive or negative diagnosis. In case of insufficient information, the software gives no result.
According to the results, the device made an accurate diagnosis in 98.4% of patients with the condition, and 78.9% of patients without the condition. Moreover, it only gave one false-negative result out of the 122 kids with ASD. Researchers believe that such a device can lead to early interventions and help combat the challenges associated with ASD.