Using the Autism Diagnostic Interview-Revised and the Autism Diagnostic Observation Schedule with young children with developmental delay: evaluating diagnostic validity.

The team compared two gold-standard autism tools in young children with developmental delay. They looked at how well the ADOS observation and the ADI-R parent interview matched expert clinical diagnosis in kids aged 20 to 55 months.

All children already had delays in language, play, or social skills. Clinicians gave both tests, then decided together if each child had autism. The study asked which test lined up best with that final team decision.

ADOS algorithm scores agreed with the expert diagnosis more often than ADI-R scores in this young, delayed group. The ADOS pulled ahead even when children had limited speech.

The ADI-R still gave useful information, but its scores were less likely to match the clinicians' conclusion.

The result backs up Oosterling et al. (2010) and Hong et al. (2021), who also found that revised ADOS algorithms and toddler cut-offs work well in young children. Together these studies build a chain of evidence that the ADOS keeps improving with each update.

Le Couteur et al. (2008), published the same year, seems to disagree. That study says using both ADI-R and ADOS together gives the best picture in preschoolers. The difference is small: M et al. still used both tools, they simply found the ADOS score itself was more accurate than the ADI-R score. Both papers agree you should gather data from parent report and direct observation.

Bennett et al. (2008) raises another caution. In a Hispanic sample, the revised ADOS sometimes over-identified autism in children with mild social problems. M et al. worked with children who already had clear developmental delay, so the algorithm worked cleanly there. Culture and symptom level matter when you pick cut-offs.

If you assess toddlers or preschoolers who are slow to talk or play, lean on the ADOS for the diagnostic number. Still interview parents with the ADI-R for developmental history, but trust the ADOS algorithm when the scores clash. And remember: kids from different backgrounds or with subtle symptoms may need clinician judgment on top of the score.