Skip to content Skip to footer

This AI Tool For Diagnosing Autism Could Hit The Market ‘In The Second Half Of 2021’

A-sign-for-the-Food-And-Drug-Administration-is-seen-outside-of-the-headquarters-

A tool that helps pediatricians diagnose autism with artificial intelligence and smartphone video passed a key FDA test and could hit the market as soon as next summer, according to device-maker Cognoa, which recently announced its plans to submit the tool for final approval.

In a trial to measure its safety and effectiveness, doctors diagnosing with the Cognoa tool made the same decision as an expert panel more than enough times to beat its FDA benchmarks, said CEO Dave Happel. He expects the device, which Cognoa says will be the first of its kind, to be approved in the second half of 2021.

If approved, Cognoa could offer a faster path to diagnosis. Parents notice signs of autism at an average of 14 months, but it takes three years for the median child to get diagnosed at 51 months. One reason for the gap is the wait to see a specialist, which can take months to more than a year.

In contrast, Cognoa’s process can provide a diagnosis as soon as two to six weeks after a pediatrician suspects autism and orders the diagnostic, said Happel. That means children could get treatment earlier while their brain is more easily rewired, he said.

Cognoa’s AI works by processing the data of three questionnaires: from the pediatrician, the parents, and an autism specialist who watches two short videos of the child filmed by parents in the Cognoa app. Then the pediatrician receives a result from the AI program, which they use to make a diagnosis.

Currently there is no way to verify the tool’s claims of effectiveness. “Cognoa is not sharing the study data until after we have completed our submission to the FDA,” the company said in an emailed statement, though it will be published in a peer-reviewed journal “in the coming months.” The FDA cannot verify or comment on product applications, said agency spokeswoman Stephanie Caccomo in an email. The agency granted Cognoa Breakthrough Device designation status in 2019, which is a program to speed up development if the device is more effective in treating or diagnosing “life-threatening or irreversibly debilitating human disease or conditions.”

Part of autism business boom

Cognoa, which has raised $60 million since its founding in 2013, is part of a broader infusion of capital into companies for autism products and services. That growth has been fueled by an increase in the number of children being diagnosed with autism and the passage of laws, in all 50 states, that require insurance companies to cover autism services, said Ronit Molko, a consultant to private equity on investments in behavioral health companies and cofounder of the Autism Investor Summit. In the past five years, said Molko, there have been more than 100 deals involving autism-related companies, including the $400 million purchase of an autism therapy provider. “It’s kind of like a feeding frenzy,” she said.

2017 report by Research and Markets projected that the US market for autism treatment will grow to $2.23 billion by the end of 2021. The majority of autism companies are service providers that teach social, language, and behavioral skills, said Molko, who cofounded one such business, Autism Spectrum Therapies, before selling it in 2014. Companies making diagnostic tools like Cognoa are comparatively rare.

Molko understands the skepticism by some autism clinicians that products like Cognoa could diagnose as effectively as a specialist who observes a child for as many as 12 hours. At the same time, she sees the promise of AI to cut down the waiting time for diagnoses and make them more accurate. That would help children get services sooner, she said. “The impact on life, and the family, and that individual child is huge.”

Insurance companies also see the promise, according to Happel. They are interested in covering Cognoa, he said, because it will be less expensive than the many specialist visits required for a traditional diagnosis. Insurers are also interested in the device’s potential to get children into treatment earlier, which could improve their diagnosis and lower their need for costlier long-term services. Happel also expects Medicaid, which pays for diagnosis and treatment, to cover Cognoa. The company is still working on pricing the product, he said.

‘Not the answer’

Catherine Lord, an autism clinician and professor at UCLA, agrees that a shortage of specialists like herself is behind the long wait for diagnoses. But products like Cognoa’s will not fix the long wait time, said Lord, the creator of a widely-used diagnostic tool, because they do not provide as many details about the child’s condition as a traditional diagnosis.

Without knowing the severity of autism or the presence of mental disorders, such as ADHD, that often accompany autism, parents can’t make informed decisions about their child’s treatment and education, she said. “The label is an important start, but it is not the answer,” said Lord, who regards Cognoa as a screening device. “It might increase knowledge and referrals,” she said, “but this is just going to send more kids to me.”

In an emailed statement responding to Lord, Cognoa said the device “provides pediatricians with the information needed to give a detailed diagnosis of autism … and prescribe specific individual early interventions” based on criteria from the American Psychiatric Association. For more complex treatment, pediatricians can make referrals to “local specialists and therapists who can provide more specific guidance when needed. A diagnosis of autism is often needed for eligibility to these resources.”

To Lord, there is a straightforward solution to cutting wait times and addressing the shortage of providers: Hospitals should hire more specialists to make evaluations, she said. They have not done that, she argued, because the procedures don’t make money. “The hospital pretty much breaks even,” she said — “or loses money.”

Graison Dangor Contributor

How AI Can Transform the Autism Services Industry

How-AI-Can-Transform-the-Autism-Services-Industry

Imagine providing parents with an app that enables them to upload videos of their children, analyzed by computers applying algorithms to diagnose autism earlier in their development than currently possible.

Imagine equipping autistic individuals with smart glasses that whisper instructions for navigating social situations in the user’s ear.

Imagine two-foot-tall humanoid robots that could serve as therapists for use in treating children with autism.

Artificial intelligence (AI) has made all this possible today. AI, equipped with machine learning, is improving constantly, and is positioned to become a critical adjunct to the human element in the diagnosis and treatment of autism. The operational and clinical advantages these advances convey are creating added value propositions for those enterprises that embrace it.

AI Disruptors in Other Industries

The ground for these paradigm-shifting advances has been well-trod in other industries. Consider how Waze disrupted the navigation business less than a decade ago. When other mapping apps stumbled to diagnose traffic jams and redirect travelers, Waze crowd-sourced traffic reports and provided real-time data about traffic incidents, speed limits, speed traps, and other valuable trip information to navigate the fastest possible route. As a result, the navigation app captured market share from Google, Mapquest, and other providers, with 25.6 million monthly users by 2018.

Waze has exploited that unique value add by establishing a ride share component that allows users to find, connect and ride together at a fraction of the cost of Uber and Lyft, themselves industry disruptors. Waze is also being accessed by governments and researchers seeking real-time data collected by its users. In Israel, researchers are asking Waze users to report roadkill, so they can track animal collisions and reroute animal movements to avoid them. The city of Hampton Roads, VA is mining Waze traffic data to determine where flooding is most acute as they develop their flood maps.

The Story of AI in Autism Services

The frame-breaking opportunities are similarly ripe in the autism services industry.  The platform created by Cognoa, a Palo Alto-based pediatric behavioral health program, uses machine learning and predictive analytics to analyze parent input of behavioral information to diagnose autism earlier. This allows intervention and customized education to begin in the critical early years of a child’s brain development.

Another AI solution developed by Brain Power, helps autistic individuals manage their emotions with a wearable digital coach that recognizes social situations and guides the user towards appropriate responses. This facilitates increased functional independence, also enabling some to interact more effectively in a standard workplace setting.

Monitors and wearable devices can also deliver a plethora of health, activity, and geographic information about clients in real time. With some analysis, staff could correlate patient meltdowns with changes in physiological ratios such as a spike in blood sugar levels, increased heart rate and sweating.

As an example, for individuals whose sugar intake is correlated with aggression, making dietary changes and collaborating with parents to monitor sugar and carbohydrate intake could reduce the frequency of outbursts and the required interventions, and would improve client outcomes and quality of life.

AI’s Impact Is Clinical and Operational

The business imperative for artificial intelligence covers the operational aspects of the autism services business as well. Employing this new business intelligence would enable service providers to tackle issues such as scheduling in a more targeted manner. If traffic on Friday afternoons in one location is causing sessions to start late, for example, modifications can be made to maximize session time.

This avoids the current practice of implementing policy and process changes across the board that may have negative impacts on service delivery in a different location where Friday afternoon traffic is not an issue. It is an ongoing process: continuing use of these tools could help determine which modifications were working and whether they were creating cancellation bulges elsewhere in the schedule.

The artificial intelligence that drives these revelations is the business imperative of the future of autism services, if not the present. It will eventually be a normal part of every autism service enterprise, as it already is in many industries. The early adopters will exploit a market inefficiency to deliver services that their competitors lack.