Improving child healthcare through early diagnosis of autism
Years to diagnose autism in children
Can act on early developmental concerns
Remote, equitable diagnosis
“Databricks Data Intelligence Platform plays a crucial role in helping us empower clinicians to make rapid, accurate diagnoses. The lakehouse architecture improves our ability to develop downstream products that enable early, equitable care that can change lives.”
Diagnosing autism has historically been hindered by subjectivity and inconsistency: There isn’t a clear process, existing assessments are long and in-clinic, there’s a growing shortage of specialists, and there are disparities in diagnosis based on socioeconomic status, race, gender, and geography. Cognoa aims to change all this with Canvas Dx, the first FDA-authorized diagnostic device that allows healthcare providers to diagnose or rule out autism in children as young as age 18 months and up to 72 months. Using the Databricks Data Intelligence Platform, Cognoa can help clinicians to make a highly accurate diagnosis in a shorter time frame than previously possible, which is crucial for getting children the treatment they need earlier, when it will have the greatest potential to impact a child’s lifelong trajectory.
How Databricks helps Cognoa innovate autism diagnoses
Because of how much time it takes, accurately diagnosing a child for autism can be a long and difficult journey, with parents and caregivers spending as much as 3 years seeking answers. Studies have shown that the younger a child is diagnosed, the better, as early intervention can greatly alter the course of development and may even erase the signs of autism altogether.
Cognoa, a pediatric behavioral health company, hopes to make a big difference in the diagnosis of autism with the release of Canvas Dx, a device designed to support clinicians in the diagnostic process, rather than to serve as a stand-alone diagnostic. Caregivers can get a prescription for Canvas Dx, and with a simple app, answer questions about their child and record and upload videos of themselves interacting with the child in their home environment. The child’s doctor also answers a questionnaire, and all this information produces an accurate “Positive” or “Negative” for autism. When there is insufficient information for Canvas Dx to confidently render a determination, the device abstains from producing either a positive or negative output. Diagnostic abstention in cases of high uncertainty or complexity is an important method of risk control in medical machine learning algorithms.
Powering this life-changing technology is Databricks Data Intelligence Platform. Cognoa uses the lakehouse to take in both the structured questionnaire answers as well as the unstructured videos, and then runs that data through a machine learning model to reach a diagnosis.
“With the help of Databricks Data Intelligence Platform, this critical data, analytics and AI can potentially shorten the time to diagnosis for many children,” said Valerie Jenkins, Head of Communications at Cognoa. “We take in huge amounts of unstructured and structured data, standardize it, and run ML models that help inform a provider’s diagnosis. And all this happens in the lakehouse.”
Not only is Canvas Dx highly accurate, but it also provides a remote diagnosis and one that is possible starting at age 18 months, compared to the U.S. average age diagnosis of 4.3 years. An earlier diagnosis is a first critical step to earlier intervention for children, which can result in positive lifelong outcomes.
Looking ahead to more life-changing applications in neurodevelopment
This ability to improve the journey of diagnosing autism among children is only the beginning. Cognoa’s data sets from rich multidimensional measures of a child’s social and behavioral features – data vital to an objective and unbiased assessment – can be extended to other neurodevelopmental conditions. The company plans to develop further accessible diagnostic and therapeutic devices with AI and ML techniques to allow for rapid, early, and equitable diagnoses and treatments.
All of this requires extensive experimentation, research and development. Using Databricks Data Intelligence Platform to collaborate in real-time with people around the world means the Cognoa team can conduct experiments more efficiently and at scale. And, when the well-being of children is at stake, being able to create production-ready models faster is crucial.
“Cognoa is not only a pediatric behavioral health company — we’re a data and AI company. We want to improve how behavioral health conditions are diagnosed and treated, and, with the help of Databricks Data Intelligence Platform, we’re just getting started,” concluded Jenkins.
Beyond autism, Cognoa plans to tackle a broad range of behavioral health conditions, including childhood anxiety, ADHD, and speech and language delay. With the help of Databricks, Cognoa is looking to make a lasting impact on children and their families from anywhere, everywhere.
