Comparison of Diagnostic Classifications Using the Original and Revised ADOS Algorithms in Individuals with Fragile X Syndrome

Background: The Autism Diagnostic Observation Schedule (ADOS) is a widely-used, gold standard tool for the diagnosis of autism and autism spectrum disorders.  It consists of various items which are administered directly with individuals in a 30-45 minute session.  These items are then used to code domains of Language and Communication, Reciprocal Social Interaction, Imagination, and Stereotyped Behaviors and Restricted Interests.  Items from the Language and Communication, and Reciprocal Social Interaction domains are used to code the original ADOS algorithm, which provides overall diagnostic categories of autism or autism spectrum.  Gotham et al (2007) reported on a revised version of the algorithm which was developed for modules 1 through 3, and also incorporates items from the Stereotyped Behaviors and Restricted Interests domain for the overall scoring algorithm.

Objectives: We were interested in seeing whether the ADOS classifications of individuals with fragile X syndrome would change when using the new algorithm, what types of changes would be seen, and with what frequency.

Methods: We used an existing sample of 304 individuals with fragile X syndrome who had been seen previously for research studies utilizing the ADOS.  We re-scored the ADOS modules 1, 2 and 3 using the new version of the algorithm reported by Gotham et al. (2007), and compared the results of the classification with the revised algorithm to those of the original algorithm.

Results: We re-scored ADOS administrations using the revised algorithms for 304 individuals (246 males, 58 females).  The subjects ranged in age from 1 to 50 years (mean=10, ± 7).  Of the 304 cases, 89 were module 1 (29%), 126 were module 2 (41%), and 89 were module 3 (29%).
Overall, 93 of 304 cases (30%) showed some type of change in their overall ADOS classification.  Specifically, changes seen were non-ASD to autism spectrum (n=14), non-ASD to autism (n=9), autism spectrum to autism (n=58), autism spectrum to non-ASD (n=7), and autism to autism spectrum (n=5).

Conclusions: A significant number of the 304 cases analyzed showed some type of change in their ADOS classification, and changes were seen in both ‘directions’.  The most frequent change shown was from autism-spectrum to autism (58/93 cases, 62%), which may suggest that the addition of the repetitive behavior items to the new algorithm has a significant influence for the classification of autism in the fragile X population.  These findings further support the need for using multiple tools for the diagnosis of autism (see Risi et al. 2006 and Harris et al. 2008)