Recognizing autism in vulnerable populations: an equity lens and guidance

Early autism diagnosis is especially difficult when developmental history or caregiver input is unavailable (e.g., displaced, institutionalized, or orphaned children). We synthesize observation-first, clinician-led approaches and regulated, technology-enabled aids that minimize reliance on collateral history.

Standardized direct observation (e.g., ADOS-2 modules that do not require a caregiver) can support diagnosis when history is limited, but feasibility hinges on training, cultural adaptation, and service capacity. Many legacy instruments predate DSM-5/DSM-5-TR, creating construct gaps. Recently cleared aids, an eye-tracking system for toddlers, and a software tool combining short videos, caregiver input, and clinician ratings, function as decision support rather than stand-alone diagnostics. Mobile and remote screening paradigms show promise but require independent, cross-cultural validation.

Diagnostic equity necessitates pathways that are effective when medical histories are incomplete. We outline a minimum dataset and a pragmatic workflow that replace caregiver reports with structured observations, school/residential collateral, and carefully integrated objective measures, under clinician synthesis. Priorities include updating legacy instruments, publishing brief observational batteries with known accuracy bounds, and validating tools across languages and contexts.