What are we optimizing for in autism screening? Examination of algorithmic changes in the M-CHAT.
Synnve Schjølberg, Frederick Shic, Fred R Volkmar and 9 others
PMID 34837355WHAT IT FOUND
The M-CHAT-R style scoring produced fewer false positives but missed more children who later received an ASD diagnosis.
A negative screen still does not rule out ASD.
Key findings
01The 20-item scoring identified fewer children with later ASD than the original 23-item criteria: true positives were 93 of 337 versus 105 of 337, and false negatives were 244 versus 232.
02The 20-item scoring reduced false positives among children without later ASD, from 4,048 to 2,757.
03The original criteria missed 232 of 337 children with later ASD, and the 20-item criteria missed 244 of 337.
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 did not administer the actual M-CHAT-R wording, examples, item order, or follow-up interview; it applied M-CHAT-R scoring criteria to the original 23-item M-CHAT responses. No follow-up interview was done for screen positives, so the reduction in false positives reflects only the scoring algorithm change. The dataset did not provide outcomes for developmental disabilities other than ASD, so it cannot show what happened to the false positives. The participants section reports 54,436 children and 332 later ASD diagnoses, while the results section reports 54,463 individuals and 337 ASD diagnoses. Symptoms may not be evident or prototypical at 18 months, so a negative screen at that age may not rule out later ASD.
Declared interests
The authors report no conflicts of interest.
The easy way to misread this
Do not conclude that the M-CHAT-R style scoring is an improvement because it reduced false positives. It also increased false negatives, and a negative screen still does not rule out ASD.