Diagnosis of Autism Spectrum Disorders in a Global Clinical Drug Trial: Challenges and Solutions

Background: Clinical trials face critical challenges to ensuring rater competency and diagnostic accuracy. Challenges are magnified when diagnosing Autism Spectrum Disorders (ASD) because accurate diagnosis requires an astute understanding of the population and the ability to synthesize information from multiple sources. The phenotypic overlap between ASD and other psychiatric disorders is well documented. The Autism Diagnostic Interview – Revised (ADI-R) and Autism Diagnosis Observation Schedule (ADOS) have become the gold standard measures for classification of ASD with increased sensitivity and specificity. There are, however, challenges to their use in international clinical drug trials:  

• Extensive training and demonstration of skill is required for use in research
• Ongoing training programs vary across the globe
• Cultural, ethnic and language-specific diversity
• No global centralized repository of trained clinicians

Objectives: 1) To train clinicians to use the ADI-R and ADOS in a global clinical drug trial; 2) Minimize variability in administration and scoring of these measures; 3) Meet required timelines and enrollment expectations; 4) Ensure that raters are able to accurately classify individuals with ASD; and 5) Design training for clinicians that, although time intensive, is manageable for a clinical trial setting.

Methods: Raters from 10 countries (Americas, Europe and Asia) participated in an ADI-R and ADOS training program for a clinical trial in pediatric ASD. Rater experience, education and previous training documentation were evaluated against stringent criteria and prequalified raters were determined to fall into one of the following tiers:  

1. Research reliable, with approved documentation
2. Experienced with the measure
3. Not experienced with the measures but approved to participate in training due to prior education and population experience

No additional training was required for research reliable clinicians (n=19 ADI-R; 17 ADOS). “Experienced” raters (n=21) were required to view and accurately rate the certification video(s) only and were not required to attend a 5 day rater training meeting (RTM). 116 approved raters trained on the ADI-R, ADOS or both measures at one of 7 RTMs. Training included 2.5 days each of ADI-R and ADOS training provided by research reliable trainers from the Autism Trainers Consortium. Following meetings each rater watched and scored the certification video, permitting tiered individualized training for this initially diverse group.

Results: Criteria for approval to rate in the trial included accurate classification by video. Preliminary data showed agreement in diagnostic classification for autism verses non-spectrum was excellent for both the ADI-R and ADOS.

Conclusions: International clinical trials in ASD face the challenge of accurately diagnosing subjects in a standardized way. The ADI-R and ADOS can be effectively and efficiently used by employing a modified training program combined with accurate assessment of previously trained clinicians. Efforts are being made to consolidate information about trained clinicians at all levels for use by researchers as a resource for future trials. The central location would also be a virtual workspace to share information about these measures to ensure a global standard for diagnosis of ASD in clinical drug trials.