Three diagnostic systems for autism: DSM-III, DSM-III-R, and ICD-10.

Demello et al. (1992) compared three autism checklists side-by-side. They looked at DSM-III, DSM-III-R, and ICD-10 rules.

The team asked expert clinicians to judge the same group of kids. Then they counted how often each checklist matched the experts.

DSM-III-R called almost every child autistic. ICD-10 and the older DSM-III were pickier and lined up better with the clinicians.

In short, DSM-III-R over-tags kids. If you want tighter labels, use ICD-10 or plain DSM-III criteria.

Aznar et al. (2005) saw the same problem in OCD and Tourette’s. They built a new label, ‘Tourettic OCD,’ to fix the overlap. Both papers say DSM can be too wide and push for sharper boxes.

Lanovaz et al. (2017) give a fix for single-case data: gather at least three points in A and five in B. That move cuts false alarms, just like picking ICD-10 cuts false autism hits.

Horner et al. (2022) tell single-case researchers to tighten design and reporting. Their call for stricter rules mirrors R et al.’s warning that loose DSM-III-R rules flood the autism pool.

Your intake team may still keep DSM-III-R habits. Flip to ICD-10 check-boxes or add a second reviewer. You will spare kids from unneeded labels and free up slots for those who truly need ABA. One quick swap, cleaner caseload.