Diagnostic assessment in a sample of autistic and developmentally impaired adolescents.

Researchers tested two ways to spot autism in teens with severe delays. They gave 20 youths the Autism Diagnostic Interview (ADI) and checked them against DSM-III-R rules.

Each teen also saw a senior clinician who gave a best-estimate diagnosis. The team then compared how well each tool matched the expert view.

The ADI agreed with the clinician 90 % of the time. DSM-III-R checklists agreed only 45 % of the time and called many non-autistic teens autistic.

Low specificity means the old DSM labels were too loose. The ADI cut false positives in half.

Two years later Craddock et al. (1994) tried the same ADI in high-IQ kids. They still saw misses, proving the ADI helps but does not replace clinical judgment in any IQ band.

Schaaf et al. (2015) swept 23 studies and showed the pattern held for 20 years: each DSM rewrite kept losing mild cases. Their numbers back Fombonne (1992)’s warning that criteria drift is real.

Pathak et al. (2019) added a twist: even after a correct autism label, daily-living scores lag far behind IQ in teens. Good diagnosis is only step one; step two is tracking adaptive gaps.

When you assess a teen with clear delays, start with the ADI or current ADOS, not just DSM checklist boxes. Pair the tool with your own eyes, and plan to measure adaptive skills next. This one-two punch cuts mislabeling and points you to the life-skills programs the youth actually needs.