Performance of a Novel Software-based Autism Spectrum Disorder Diagnosis Aid for Use in a Primary Care Setting

Background: The lack of diagnostic tools for Autism Spectrum Disorder (ASD) in primary care settings contributes to a delay of 3 years between first parental concern and diagnosis. Efficient and effective tools for primary care physician (PCP) use that leverage technology have the potential to eliminate the lag between screening and diagnosis thereby streamlining the process and enabling timelier ASD-specific interventions.

Objective: This study examined the performance of an artificial intelligence-based (AI) software as a medical device to aid in the diagnosis of ASD.

Methods: This was a prospective multi-site pivotal study conducted in 6 states using a double-blind active comparator design. Device results were compared to diagnosis by DSM-5 criteria determined by independent agreement of experienced specialist clinicians in 425 completed subjects (36% female) between the ages of 18-72 months old.

Results: Comparison of the device results to specialist diagnosis found the Positive Predictive Value (PPV) for the device was 80.8% with 95%CI [70.3%, 88.8%], the Negative Predictive Value (NPV) was 98.3% with 95%CI [90.6%, 100%], sensitivity was 98.4% with 95%CI [91.6%, 100%], and specificity was 78.9% with 95%CI [67.6%, 87.7%] for subjects with determinate device results. Covariate analysis demonstrated consistent performance of the device across subject’s gender, race/ethnicity, income, and education level.

Conclusion: Using this device, PCPs could efficiently and accurately diagnose at least 50% of children 18- 72 months old with ASD in the primary care setting. Output of the device to aid diagnosis is available within days as compared with the current average time to diagnosis of three years. This would enable earlier ASD specific interventions and increase the chances of optimal outcomes. Additionally, these results provide preliminary evidence of the potential for this diagnosis aid to address some of the disparities that exist in time to diagnosis for children who are non-White, female, and with lower socio-economic status.

Presenting Author(s)

Sharief Taraman, MD, DABPN, DABPM, FAAP
Physician
Children's Health Orange County, UC Irvine SOM, Chapman University, AAP, Cognoa
Orange County, California
Disclosure: Cognoa, Inc (Employee, Stockholder/Ownership Interest (excluding diversified mutual funds))

Co-Author(s)

jm
jonathan T. megerian, MD , PhD, FAAP
Medical Director Thompson Autism Center at CHOC, Associate Professor of Pediatrics , UCI School of M
Childrens Hospital of Orange County, University of California at Irvine
Ladera Ranch, California
Disclosure: No Financial Relationships Disclosed
SD
Sangeeta Dey, Psy.D.
Pediatric Neuropsychologist
Stanford Children's Health
San Jose, California
Disclosure: No Financial Relationships Disclosed
RM
Raun D. Melmed, MD
Director
Melmed Center
Scottsdale, Arizona
Disclosure: Cognoa (Products/Services: Consultant fee re use of Cognoa) (Consultant)
Daniel Coury, MD
Professor of Pediatrics and Psychiatry, The Ohio State University
Nationwide Children's Hospital
Columbus, Ohio
Disclosure: Cognoa (Products/Services: Consultant on use of autism diagnostic aids in primary care settings.) (Consultant)
ML
Marc Lerner, MD
Clinical Professor
University of California, Irvine
Santa Ana, California
Disclosure: Cognoa (Contracted Research (Principal Investigators must provide information, even if received by the institution), Specialist Clinician (Study Central Panelist))
Christopher J. Nicholls, PhD
Adjunct Faculty
Arizona State University -
Scottsdale, Arizona
Disclosure: No Financial Relationships Disclosed
Kristin A. Sohl, MD
Professor, Pediatrics
University of Missouri
University of Missouri School of Medicine
Columbia, Missouri
Disclosure: Autism Navigator (Products/Services: Consultant : fees for program feedback and development) (Consultant)Cognoa (Products/Services: Consultant: fees about diagnostic reports) (Consultant)Quadrant Biosciences (Products/Services: Medical advisory board - related to science and care of autism spectrum disorder) (Honorarium)
RR
Rambod Rouhbakhsh, MD, MBA, FAAFP
Program Director
Forrest General Hospital Family Medicine Residency
Hattiesburg, Mississippi
Disclosure: No Financial Relationships Disclosed
AN
Anandhi Narasimhan, M.D.
Psychiatrist
Masada Homes
LOS ANGELES, California
Disclosure: No Financial Relationships Disclosed
SG
Sailaja Golla, MD
Pediatric neurologist
Children’s hospital of Orange County
Orange, California
Disclosure: No Financial Relationships Disclosed
SS
Safiullah Shareef, M.D.
Neurologist
Texas Child Neurology
Plano, Texas
Disclosure: Texas Original (Products/Services: Speakers Bureau - July 2021 - on the Medical Cannabis Compassionate Use Program) (Speaker's Bureau)
Jennifer Shannon, MD
Regional Medical Director
Cognoa
Bellevue, Washington
Disclosure: Cognoa (Products/Services: Regional Medical Director Employee) (Employee)
CK
Colleen A. Kraft, MD, MBA
Professor of Pediatrics, Keck School of Medicine, University of Southern California
Children's Hospital Los Angeles
Los Angeles, California
Disclosure: Cognoa (Products/Services: Senior Medical Director for Clinical Adoption, working to transform pediatric practice to improve care for kids with developmental disabilities) (Employee)DotCom Therapy (Products/Services: Board of Directors, general direction on clinical strategy and operations) (Ownership or partnership, Stockholder/Ownership Interest (excluding diversified mutual funds))Happiest Baby (Products/Services: Consulting on clinical partnerships) (Consultant)
HA
Halim Abbas, n/a
Chief AI Officer
Cognoa
San Jose, California
Disclosure: Cognoa Inc (Products/Services: Employee - Chief AI Officer) (Employee)
DG
Diana Gal-Szabo, PhD
Consultant
Cognoa Inc.
Phoenix, Arizona
Disclosure: Cognoa Inc. (Products/Services: Consultant providing medical affairs support to Cognoa Inc.) (Consultant)