Autism Detection in Children by Combined Use of Gaze Preference and the M-CHAT-R in a Resource-Scarce Setting.
Kelly Jensen, Sassan Noazin, Leandra Bitterfeld and 21 others
PMID 33591436WHAT IT FOUND
Adding a 15-second gaze test to the M-CHAT-R improved autism screening accuracy in Peruvian children, raising specificity from 63% to 92%.
This reduced false alarms but lowered sensitivity to 64%, meaning some cases would be missed.
Key findings
01Children with ASD spent significantly less time looking at social scenes than typically developing children during the first 15 seconds of the video (30.1% vs 50.1%).
02The combined GP-MCHAT-R model achieved a specificity of 92% with a sensitivity of 64%, compared to the M-CHAT-R alone which had a specificity of 63% and sensitivity of 93%.
03The combined model showed a significant improvement in overall accuracy (AUC 0.89) over the M-CHAT-R alone (AUC 0.78).
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study was small and exploratory, with only 28 children in the ASD group. Cases and controls were not age-matched; the ASD group was significantly older (median 60 months) than the control group (median 42 months). The age range extended to 99 months, far beyond the typical 48-month validation window for the M-CHAT-R. There was no control group of children with other developmental delays, so it is unclear if the gaze differences are specific to ASD or common to any developmental disorder. Traditional inter-rater reliability statistics were not calculated due to resource constraints.
Declared interests
Funding was provided by Innóvate Perú, The GEMS Grants through Johns Hopkins University, the Wendy Klag grant, the Bill and Melinda Gates Foundation, and the Fulbright-Fogarty Fellowship. No commercial conflicts of interest were declared.
The easy way to misread this
Do not interpret the improved specificity as a sign that this tool is ready for clinical use. The sensitivity dropped to 64%, meaning it fails to identify 36% of children with ASD in this sample, which could lead to dangerous delays in intervention.