ADOS Scoring System and Cutoff Values for Autism
The ADOS-2 uses module-specific algorithms with different cutoff scores depending on the child's age and language level, with the Toddler Module requiring scores of ≥12 (Algorithm 1) or ≥10 (Algorithm 2) for autism spectrum classification, while Modules 1-3 use a two-tier system with lower "autism spectrum" cutoffs (higher sensitivity) and higher "autism" cutoffs (better specificity).
ADOS-2 Toddler Module (Ages 12-30 months)
The Toddler Module uses two separate algorithms based on age and language ability 1:
Algorithm 1 (for toddlers 12-20 months OR 21-30 months with <5 words):
- Cutoff score: ≥12 optimizes sensitivity and specificity
- Area under the curve (AUC): 0.92
Algorithm 2 (for toddlers 21-30 months with ≥5 words):
- Cutoff score: ≥10 optimizes sensitivity and specificity
- Area under the curve (AUC): 0.96
These cutoff scores have been validated in a large replication study of 412 toddlers and align strongly with best estimate clinical diagnosis 1, 2.
ADOS-2 Modules 1-3 (Ages 31 months and older)
Modules 1-3 employ a two-tier classification system with distinct cutoff ranges 3, 4:
Two Classification Levels:
"Autism Spectrum" cutoff (lower threshold):
- Provides high sensitivity (95%+) but lower specificity (63%-73%)
- Results in more false positives (29% in non-ASD group)
- Does not reliably predict diagnostic outcome (56% ASD vs 44% No ASD)
"Autism" cutoff (higher threshold):
- Provides better specificity (76%-86%) with favorable sensitivity (81%-94%)
- More accurate for definitive ASD diagnosis
Scoring Domains:
The revised algorithms assess two primary domains 4:
- Social Affect (SA) domain
- Restricted and Repetitive Behaviors (RRB) domain
- Combined total score determines classification
Critical Clinical Considerations
The autism spectrum classification range is problematic for clinical decision-making because it lacks directionality—nearly half of children scoring in this range do not receive an ASD diagnosis 3. When using ADOS-2 alone, the higher "autism" cutoff provides better diagnostic accuracy than the "autism spectrum" cutoff.
Combined assessment yields optimal results: Using ADOS-2 together with the Autism Diagnostic Interview-Revised (ADI-R) provides balanced sensitivity (77%-80%) and specificity (87%-90%) 5. When evaluated separately, the ADOS-2 (using autism cutoff) demonstrates better diagnostic accuracy than the ADI-R alone 5.
Common Pitfalls to Avoid
- Do not rely solely on the autism spectrum cutoff for diagnostic certainty—it has high false positive rates and poor predictive value for final diagnosis 3
- Recognize that ADOS-2 is not diagnostic in isolation—it must be combined with clinical judgment, developmental history, and ideally ADI-R 6, 7
- Account for age and language level when selecting the appropriate module and algorithm 1
- Be aware that younger children (12-24 months) may be difficult to differentiate from those with developmental delays using ADOS alone 6