Brief Report: exploratory analysis of the ADOS revised algorithm: specificity and predictive value with Hispanic children referred for autism spectrum disorders.

Terry’s team looked at the revised ADOS algorithm in 78 Hispanic kids. All were referred for autism. They ran modules 1, 2, or 3 and compared scores to final clinic diagnosis.

The goal: see if the new cut-offs catch real cases and avoid false positives in this group.

Module 1 improved a bit. It found more true cases without adding many false ones.

Modules 2 and 3 slipped. They over-called autism, especially in kids with mild social delays. Specificity dropped to 65 % and 50 %.

Fleury et al. (2018) pooled eight studies and saw the same patchy pattern. Their review supersedes Terry’s single site: gains are tiny and still module-dependent.

de Bildt et al. (2011) echoed the worry in Dutch kids. Module 1 worked; module 2 wobbled. This conceptual replication shows the problem crosses cultures.

Hong et al. (2021) looked younger. Toddler Module cut-offs held strong. Their data extend Terry’s warning: once kids age into modules 2/3, accuracy dips again.

If you test Hispanic preschoolers, treat module 2 or 3 scores as a yellow flag, not a stop sign. Pair the ADOS with parent interview or observe twice before you label. Document mild social symptoms clearly and re-check cut-offs against the 2018 review tables.