Does broadening diagnostic criteria explain the rise in autism prevalence?

The short version: wider criteria and better recognition explain part of the rise, but the retrieved studies cannot tell you how much.

The best evidence here is a UK population study covering 1998 to 2018. It found a 787% rise in new autism diagnoses, with the steepest increases among adults and females [2]. That pattern fits a widening net: people who would never have been counted before are now being diagnosed. The authors themselves say policy and diagnostic changes likely contributed [2]. An older UK district study reached the same conclusion, calling a doubling of recorded cases "increased diagnosis of previously missed cases" rather than a true rise in how often autism occurs [3].

But notice what these studies cannot do. They count diagnoses, not autism itself. The UK authors admit they cannot give true prevalence because of missing and private diagnoses [2]. A commentary on older adults argues existing data come only from the most disabled, already-diagnosed people, so the real size of the autistic population is unknown [4]. None of these records tests whether broader criteria fully account for the increase. That comparison, applying old and new criteria to the same population, is not in this retrieval.

My call: broadening and better recognition clearly explain part of the rise, but no study here proves they explain all of it, and the true prevalence question remains open. Confidence: moderate on the partial explanation, low on the full accounting.