Long waits for autism evaluations are common, especially in rural areas where families must travel to specialty centres. Researchers at the University of Missouri School of Medicine, led by Kristin Sohl, partnered with Cognoa to test CanvasDx, an FDA-approved medical device. CanvasDx uses AI algorithms together with patient data to predict a positive or negative autism diagnosis; when the information is unclear it reports an "indeterminate" result.
The project used the ECHO Autism community, which trains primary care clinicians across Missouri and beyond. The aim was to see whether CanvasDx could support primary care clinicians in places without nearby autism specialty services and increase local access to evaluation.
In the study the average travel distance to specialty care was 97 miles, and keeping care local meant families received a diagnosis 5–7 months earlier. In a group of 80 children the device produced determinate results for 52% and did not give any false positives or false negatives; it never contradicted a clinician's diagnosis.
Difficult words
- evaluation — process to judge or measure a medical conditionevaluations
- specialty — specific area of medical care or practice
- partner — work together with another person or grouppartnered
- algorithm — a set of rules a computer followsalgorithms
- indeterminate — not clearly positive or negative result
- clinician — healthcare professional who treats patientsclinicians, clinician's
- false positive — test result says condition exists but doesn'tfalse positives
Tip: hover, focus or tap highlighted words in the article to see quick definitions while you read or listen.
Discussion questions
- Do you think primary care clinicians should use AI tools like CanvasDx? Why or why not?
- How would receiving a diagnosis 5–7 months earlier change care for a family in a rural area?
- What other challenges might families still face even if local autism evaluations are available?
Related articles
Reducing unsafe responses in large language models
Researchers studied how large language models (LLMs) handle safety and tested training methods to reduce unsafe outputs while keeping performance. They identified key challenges and a technique that preserves safety during fine-tuning.
Study finds many “clean” products for textured hair still risky
A new study examined 150 hair products labeled “Target Clean” from a Target store in South Los Angeles. Researchers compared ingredient lists with the EWG database and found many products still pose moderate health hazards.
How social media in China shapes eating disorders
Online communities on platforms like Xiaohongshu use codes and posts that normalise strict eating and extreme thinness. Research shows more Chinese teenagers now show signs of eating disorders and experts say removal of posts is not enough.
How civil society adapts to AI and surveillance
In April 2026 Global Voices and IRIS published ten case studies from the Global Majority showing how civil society groups respond to AI and algorithmic platforms. Responses include co‑opting, countering and innovating, with local and cross‑border strategies.