Assessment & Research

The very early identification of autism: outcome to age 4 1/2-5.

Eaves et al. (2004) · Journal of autism and developmental disorders 2004
★ The Verdict

Autism diagnosed at age 2 almost always holds, and later IQ shifts hinge on the child’s starting ability, not therapy hours.

✓ Read this if BCBAs who assess or write IFSPs for toddlers and preschoolers.
✗ Skip if Clinicians who only serve school-age or adult clients.

01Research in Context

01

What this study did

The team followed 49 children who got an autism diagnosis at age 2.

They tested them again at age 4½-5 to see if the label stuck and if IQ scores moved.

No control group—just the same kids twice.

02

What they found

Nine out of ten kids still had autism 2-3 years later.

One third gained 20 or more IQ points.

Another third lost 20 or more.

Hours of therapy did not predict who rose or fell.

Kids who started with the lowest skills had the biggest swings.

03

How this fits with other research

Saunders et al. (2005) watched autistic preschoolers play.

They saw messy, disorganized pretend play that differed even from kids with other delays.

Hatton et al. (2004) now show these same toddler-aged kids keep the autism label, so the quirky play is an early red flag you can trust.

Konstantareas et al. (1999) mapped Wechsler sub-test profiles in older, high-functioning autistic children.

They found Comprehension scores lower than Block Design.

Hatton et al. (2004) extend this by showing IQ can jump or drop sharply before age 5, so single early scores are not fixed limits.

Together the papers say: diagnose early, expect uneven cognitive growth, and keep re-testing.

04

Why it matters

You can tell families an age-2 autism diagnosis is solid.

Do not promise therapy dose alone will raise IQ; watch the child’s starting skill set instead.

Plan to re-test near kindergarten—some kids will need a harder curriculum, others more life-skills focus.

Free CEUs

Want CEUs on This Topic?

The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.

Join Free →
→ Action — try this Monday

Pull your youngest clients’ baseline VB-MAPP or IQ data and schedule a re-test before kindergarten transition.

02At a glance

Intervention
not applicable
Design
pre post no control
Sample size
49
Population
autism spectrum disorder
Finding
mixed

03Original abstract

Forty-nine 2 years olds with social and language characteristics suggestive of autism were identified by community professionals and screening tools, then given a diagnostic assessment and reexamined at age 4 1/2. Agreement between autism clinic and screenings was high, with 88% receiving a diagnosis on the autism spectrum. The children were lower functioning relative to the autism population, thus more likely to be identified early. Reliability of diagnoses from 2 1/2 to 4 1/2 was high with 79% staying in the same diagnostic category, but more so for clear autism than for PDDNOS. About a third improved over 20 IQ points and similar number similarly declined. Changes were not related to amount or type of intervention but were related to the children's characteristics. Higher functioning children with milder autism were the most improved.

Journal of autism and developmental disorders, 2004 · doi:10.1023/b:jadd.0000037414.33270.a8