Misconception Repair Lab
Diagnosing childhood autism when the history shows a regression, not a from-birth pattern
-
1. The tempting shortcut
"The child now shows autism-like social and communication difficulties, so this is childhood autism."
-
2. Stop and think
Before you read on: what is wrong with this reasoning? Try to answer in your own words, then reveal the explanation below.
-
3. Reveal — why it fails
Childhood autism specifically requires onset evident before age 3, with no preceding period of normal development. A documented period of apparently normal development followed by a definite loss of previously acquired skills points instead toward Rett syndrome or childhood disintegrative disorder, both of which are defined by that regression pattern. Classifying a genuine regression as childhood autism erases a clinically important distinction — the two groups differ in course, associated features (such as Rett syndrome's deceleration of head growth), and what the regression itself may signal.
-
4. Reveal — the better reasoning rule
- Does the vignette establish a period of normal development, and if so, for how long, before any change was noticed?
- Is there a documented loss of previously acquired skills (language, hand use, social engagement), rather than a pattern present since infancy?
- Regression after a normal period should always be checked for before defaulting to childhood autism, even when the current presentation looks similar.
Corrective example: In F84's regression boundary case, a child who developed typically for the first 18 months and then lost previously acquired words and purposeful hand use, with decelerating head growth, should be worked through as a regression pattern (pointing toward Rett syndrome) rather than classified as childhood autism.
5. Try the transfer case
Apply the better rule above to a fresh case.