AI Tool Helps Parents Diagnose Children with Behavioral Conditions
Raji Sahota posted on October 09, 2020 |
The Cognoa app will enable parents to help diagnose and learn about autism.
Cognoa, the diagnostic tool, just passed all its FDA targets. (Image courtesy of Cognoa.)
Cognoa, the diagnostic tool, just passed all its FDA targets. (Image courtesy of Cognoa.)

A California-based company is breaking the barriers between autism resources and those who need it most.

Cognoa, a medical technology company, has created an app to diagnose and provide information on behavioral treatments. The app analyzes answers from a questionnaire and includes short videos to provide an early diagnosis using AI and proprietary machine learning. The company hopes the app will help the almost 64,000 general pediatricians diagnose children using the American Academy of Pediatrics (AAP) updated Autism Spectrum Disorder (ASD) guidelines as of January 2020.

According to the company, its app delivers accurate data about each patient’s diagnosis and how it will likely develop in the future, especially regarding low-functioning autism. The app combines data-driven predictive models and clinical validation to supply standard-of-care diagnostics and therapeutics.

Currently, children who have suspected developmental delays are referred to a specialist who can provide a diagnosis and future treatments. However, the waiting period to receive the diagnosis and get treatment can be up to several months or even years. While 85 percent of parents notice developmental concerns in their children by age three, the median age of diagnosis is around 4.3 years old.

There are also few clinics specializing in behavioral therapy. A clinic and a family-in-need may be 50 miles apart in states like Wyoming and Nevada and as much as 100 miles in Alaska.  

Cognoa’s method is based on data-driven predictive models and clinical validation. (Image courtesy of Cognoa.)
Cognoa’s method is based on data-driven predictive models and clinical validation. (Image courtesy of Cognoa.)

In the current system, the specialist will conduct a developmental screening, which includes a questionnaire. Though these tests are inexpensive and flag some at-risk children, they may not always be accurate as they often give false positives. A comprehensive diagnostic evaluation is more accurate but is also expensive. 

The average age of a patient’s autism diagnosis has remained at 4.3 years of age for over 15 years. (Image courtesy of Cognoa.)
The average age of a patient’s autism diagnosis has remained at 4.3 years of age for over 15 years. (Image courtesy of Cognoa.)

Early intervention is key to better lifelong outcomes for children and their families living with autism as it allows for treatment during a key window of developmental plasticity. With the app, parents can obtain a diagnosis during critical neurodevelopmental windows. 

Cognoa provides opportunities for earlier treatment during the development of a child’s neural pathways. (Image courtesy of Cognoa.)
Cognoa provides opportunities for earlier treatment during the development of a child’s neural pathways. (Image courtesy of Cognoa.)

Families can now get help all from the comfort of their homes while their child is behaving in a normal state.

The app directs parents through a developmental screening test, which includes a question about the child’s behavioral patterns that may indicate that the child may be at risk. The questions are phrased so that parents can understand without needing clinical assistance. There are 17 questions for children less than four years old and 21 questions for children over four.

Parents are directed to upload two to three, two-minute videos of their child during play or mealtime at home. Cognoa then uses AI and three trained analysts to identify key behaviors that can lead to a diagnosis. In addition, analysts review the questionnaire. 

 The Cognoa process. (Image courtesy of Cognoa.)
The Cognoa process. (Image courtesy of Cognoa.)

Analysts must make observations from the short videos provided by the parents and skip any questions if they cannot be answered by watching the videos. The machine learning module will weigh the questionnaires from the analysts and parents the same but use the analysts’ answers in cases that they give contradictory answers to the same question. This can result in a discrepancy in the app’s accuracy unless steps are taken to enable better observations. 

Once the app determines it has enough information to provide a diagnosis, a pediatrician will receive the results. The pediatrician will then review the results as well as patient context, history, and experience to make a final decision.  

The Cognoa app. (Image courtesy of Cognoa.)
The Cognoa app. (Image courtesy of Cognoa.)

From July 2019 to May 2020, an American scientific research group tested the app using 425 children from the ages of 18 to 72 months who had developmental concerns but were never formally evaluated or diagnosed with autism. 

The study was conducted at 14 sites across the U.S. to evaluate how Cognoa’s ASD Diagnostic device helped provide a diagnosis while meeting the clinical reference standard. One or more reviewing specialist clinicians had to determine how often the app identified a patient on the ASD and how often it identified someone who was not on the spectrum. 

In the study, participants went through the entire process on the Cognoa app and then completed an appointment with a primary care physician and pediatric specialist. Most of the process was completed remotely. 

The study concluded that the app should extend its service to identify the severity of the condition as well as to build algorithms for children, adults and seniors with other behavioral conditions. 

The company has passed FDA targets due to the study, which means Cognoa can now submit the app to the FDA for clearance.  

Cognoa is already working on using the app to observe how each treatment is affecting a child and modifying their therapy for better results using AI.

Recommended For You