Examiners select one of five modules to administer, which is chosen on the basis of examinee expressive language level and chronological age. Each module engages the examinee in a series of activities involving interactive stimulus materials. The ADOS-2 assesses communication, social interaction, play/imagination, and restricted and repetitive behaviors. The ADOS-2 has updated protocols, revised algorithms, and a Toddler Module for children between 12 and 30 months. Observations are recorded and scored by the examiner. Modules 1 through 4 provide cutoff scores to aid in interpretation. The Toddler Module provides “ranges of concern” rather than cutoff scores. Administration time is 40 to 60 minutes. Accurate administration and coding of the ADOS-2 are highly standardized and valid assessment requires training. Training options include clinical workshops or Training Packages offered by WPS.
Age: 1 year to 100 years
Time to Administer: 40–60 min.
Method of Administration: Using a series of standardized “presses”, clinician engages the examinee in a range of activities using interactive stimulus materials that are designed to elicit specific behaviors that are consistent with ASD
Modules 1-4 yield Overall Total Score based on algorithm items, which permit comparison of cutoff scores associated with “autism”, “autism spectrum”, and “non-spectrum”.
Toddler Module yields Overall Total Score that is associated with “ranges of concern” rather than cutoffs, which are designed to assist in forming clinical impressions
Note 1: The ADOS-2 developers caution against reporting of scores in reports because they are easily misinterpreted in ways that result in invalid use of scores. The test authors encourage focusing on descriptions of the behaviors observed during ADOS-2 administration that lead to the overall ADOS-2 classification.
Note 2: Instrument requires special training to administer.
Subscales: Overall Total; Domain scores for Social Affect, Restricted and Repetitive Behaviors; Comparison Scores (to provide an indicator of severity and aid in comparison across Modules or time)
Autism Related Research
The table below details research specific to the use of the current ADOS-2. However, its predecessor, the ADOS (Lord, Risi, Lambrecht, Cook, Leventhal, DiLavore, & Rutter, 2000) has been used and studied extensively. Selected studies specific to the use of the original ADOS include: (a) those with autism scored significantly higher on both the ADI-R and the ADOS in a diagnostic validity study of the two measures for young children with and without autism (Gray, Tonge, & Sweeney, 2008); (b) development of the ADOS algorithms were derived from and revised based on research to include two domains, which improved predictive value (Gotham, Risi, Pickles, & Lord, 2007); (c) ADI-R and ADOS both contribute to clinician judgment, resulting in more consistent and rigorous application of diagnostic criteria (Risi et al., 2006)—a finding echoed in a later study, wherein LeCouter et al., (2008) found that agreement between the ADI-R and the ADOS was good between instruments, and the study had a complementary effect in aiding diagnosis and confirmed the importance of a multidisciplinary assessment process with access to information from different sources and settings; and (d) while the ADOS and ADI-R lead to approximately 75% agreement with team diagnoses, the Gilliam Autism Rating Scale (GARS) was generally ineffective at discriminating between children with various diagnoses and consistently underestimated the likelihood of autism (Mazefsky & Oswald, 2006).
Dorlack, Myers, & Kodituwakku (2018)
Age Range: n/a: Compared multiple studies that reported data allowing for computation of sensitivity and specificity of at least one ADOS and ADOS-2 Module
Sample Size: Six studies permitted paired comparisons
Evaluation of relative merits of ADOS and ADOS-2 algorithms for Mods 1-3 (sensitivity and specificity comparisons)
Outcome:Dorlack, Myers, & Kodituwakku (2018)
For Mod 1: pooled sensitivity remained unchanged, while pooled specificity measures were more variable, with estimates for ADOS-2 algorithm used with children with no words decreasing and for those with some words increasing.
For Mod 2: pooled sensitivity increased from the ADOS; pooled specificity remained constant for children younger than 5 but decreased for children age 5 and older.
For Mod 3: pooled sensitivity increased and pooled specificity decreased for the ADOS-2.
Conclusion: future research is needed to improve diagnostic validity, as changes from ADOS to ADOS-2 sensitivity and specificity were nuanced depending on the module and characteristics of the child.
Maddox, Brodkin, Calkins, Shea, Mullan, Hostager, Mandell, & Miller (2018)
Age Range: 18 years and up
Sample Size: 75
Diagnostic accuracy of ADOS-2 among adults in community mental health
Outcome:Maddox, Brodkin, Calkins, Shea, Mullan, Hostager, Mandell, & Miller (2018)
The ADOS-2 Module 4 accurately identified all adults with ASD; however, it also had a high rate of false positives among adults with psychosis (30%).
Conclusion: social communication difficulties measured by the ADOS-2 are not specific to ASD, particularly in clinically complex settings.
Fusar et al. (2017)
Age Range: 18 years or older
Sample Size: 113
Sensitivity and specificity of the ADOS-2 and ADI-R in diagnosing adult ASD
Outcome:Fusar et al. (2017)
Findings regarding the ADOS-2 are consistent with previous studies of discriminant validity of ADOS-2 Module 4 in samples of adults with average or above average intelligence, which together cautiously suggest it could be a reliable instrument for first evaluations of adults. Agreement between ADOS-2 and ADI-R scores was fair, but agreement between ADI-R and clinical diagnosis was poor, correctly classifying into the spectrum only 55% of the sample. ADI-R presented good specificity but lacked in sensitivity.
Conclusion: The ADOS-2 Module 4 accurate classifies adults with ASD who do not have comorbid intellectual disability; however, the ADI-R may not capture essential autism features in older people. This may be, in part, related to concerns about retrospective reporting for developmental profile required within the ADI-R. Overall, authors emphasize the importance of clinician training and multifaceted data collection when assessing adults on the spectrum.
Hedley et al. (2016)
Age Range: <3. 5 years
Sample Size: 125
Community-based diagnosis with ADOS-2 Toddler Module and Module 1
Outcome:Hedley et al. (2016)
Younger, nonverbal children had higher standardized calibrated severity scores (CSS) , suggesting overlap with developmental and language level; it may also reflect greater symptom severity, leading to earlier referral. The CBCL externalizing score emerged as an independent predictor of ASD symptom severity.
Conclusions: results tentatively support use of the Toddler Module into community practice, but caution in interpretation of the CSS score in younger, nonverbal children is advised.