How a tech startup is using AI to find better treatments for autism

By Sy Mukherjee



June 20, 2022

How a tech startup is using AI to find better treatments for autism

Ling Shao’s 3-year-old autistic son struggles to communicate. “My youngest son, he has almost a calcified part of his forehead because he bangs it on the floor so much because he’s frustrated, and he can’t communicate what he wants,” she says. “It’s heartbreaking.”

Shao, who is the mother of four children with autism spectrum disorder, is a health industry veteran. She has held executive leadership roles at insurer UnitedHealth, healthcare services giant Optum, and digital healthcare startup Buoy Health. But she grew increasingly frustrated with what she sees as some of the biggest barriers to more widely accessible autism care: a lack of data transparency that makes defining—and, consequently, seeking out—proven, high-quality care difficult for parents.

How a tech startup is using AI to find better treatments for autism

[Photo: SpectrumAi]

Last year Shao launched SpectrumAi to tackle some of these fundamental shortcomings in autism care, using a software platform and artificial intelligence to capture autism treatment data with an eye toward a future in which care providers are paid for the value they bring rather than the bulk of services they provide, and encouraging best practices.

The company announced a $9 million funding round in early June led by F-Prime Capital and Frist Cressey Ventures, alongside investments from the Autism Impact Fund. Its mission, in brief, is to bring transparency to the notoriously opaque and poorly measured practice of applied behavioral analysis (ABA), a form of autism therapy. 

About 1 in 44 American children have autism spectrum disorder, according to the Centers for Disease Control and Prevention, and ABA is by far the longest-standing and most commonly used type of autism therapy, with multiple studies showing its promise in children. Types of ABA may include positive reinforcement, where a patient earns rewards for accomplishing certain communication goals or answering questions from their therapist correctly. Therapies might include things like picture exchange communication, where a patient may begin asking for an object by exchanging a picture of it for the real thing, eventually working his or her way up to verbally asking for the object. 

These are extremely valuable tools for people with autism spectrum disorder, especially children who fall behind in their communication skills. But therapy providers who apply ABA incorrectly can be actively harmful and reinforce negative behaviors, Shao says. And one basic shortcoming in the U.S. health system, where behavioral and mental health care is often given short shrift, is that there are “no universally accepted objective data standards to evaluate outcomes and improve quality of care,” she says.

Some parents may flounder to find an effective ABA care provider because the data simply doesn’t exist in convenient form. SpectrumAi aims to bring more clarity around what works and produce better treatment outcomes by measuring it in an empirical way via a digital platform. Rather than having to scribble largely unquantifiable and subjective paper notes that interrupts the flow of ABA therapy, which requires deep engagement by the therapist, SpectrumAi’s software helps automate the data collection process and provides a widely accessible information hub that can also help therapists zero in on the most effective ABA techniques for different patient groups. Shao did not divulge SpectrumAi’s specific customers, but said they include major players across the U.S. health insurance and ABA provider industries.

The SpectrumAi software’s purpose is twofold: give insurers a simpler way to glean exactly what kind of treatment goals and ABA techniques are being used in sessions (the behavioral health equivalent of seeing whether an MRI and blood test were administered during a hospital visit), and give ABA therapists an interface that makes it much simpler to record the kinds of techniques they’ve employed over the course of a session.

“We’ve given these great providers a unicycle, and meanwhile they should be getting a self-driving car that’s actually helping them versus preventing them from doing their jobs better,” Shao says.

As SpectrumAi analyzes the treatments and outcomes, Shao hopes to share best practices with ABA therapists to help identify proven approaches for each child’s individual needs. “One of the other ways we think you can drive the quality is really that shared learning, by using that data and providing it back to the providers saying, ‘Hey, here’s what works better than other methods,’” Shao says. “Maybe this works on this kid, and this other thing works on this other kid, given their kind of assessments and their background and the challenges they may face.”

Sy Mukherjee has reported on the healthcare industry for a decade. He is a consultant and communications architect at IDEA Pharma.