Overlap between autism and specific language impairment: comparison of Autism Diagnostic Interview and Autism Diagnostic Observation Schedule scores.

The team gave the ADI-R interview and ADOS observation to children already diagnosed with specific language impairment. They wanted to see how many of these kids also crossed autism spectrum cut-offs.

No new tests were created. They simply applied the standard ADOS and ADI-R scoring rules to a group that only had language problems on paper.

Four out of ten children with SLI met the spectrum cut-off on at least one of the two tools. Yet their language scores did not predict how high they scored on autism items.

In plain words, many kids who only "talk late" can look autistic on our best gold-standard tests.

Gotham et al. (2007) updated the ADOS algorithm just one year earlier. Their tighter rules were the same ones that flagged the SLI kids, so the overlap is happening under the most current scoring.

Hattier et al. (2011) later showed that even in a real clinic, ADOS cut-offs alone give false positives. Their data extend this study’s warning: always blend test scores with clinician judgment.

Heslop et al. (2007) add another wrinkle. They proved that picking an ADOS module that is too hard can raise scores. Many SLI children have weak language, so module mismatch could partly explain the high autism numbers seen here.

If you test a child who only has language delays, be ready for ADOS or ADI-R scores that flirt with the spectrum range. Do not rush to re-label; gather developmental history, check peer play, and watch beyond the cut-off. Match the ADOS module to the child’s real language level, and use team discussion before you call it autism.