Vineland Adaptive Behavior Scale in Multicenter International Clinical Trials: Challenges and Solutions for a Successful Implementation.

Background:

The assessment of adaptive functioning is critical in neurodevelopmental disorders (e.g., Autism Spectrum Disorder (ASD), Down syndrome (DS)). The Vineland Adaptive Behavior Scale, Second Edition (Vineland-II^TM) is the gold standard measure for assessing adaptive functioning. The Vineland-II^TM is a complex clinical assessment that poses several challenges to the successful implementation in multicenter, international clinical trials. A rigorous rater selection and a standardized rater training program, linguistically validated translations including expert clinical review to assess cross-cultural appropriateness of item content, and specialized data management, are needed for the successful implementation of the Vineland-II^TM in these studies. Because hand scoring is prone to error, use of the scoring software program is required. Procurement and merging of the Vineland-II^TM output from multiple sites requires accurate processing of the files from the scoring software in a standardized manner to a central resource. The management of these data requires both sophisticated data management processes and expert clinical review to ensure data accuracy.

Objectives:  Summarize challenges and solutions for the successful implementation of the Vineland-II^TMin multicenter, international clinical trials.

Methods:  Data for this research are derived from Cogstate’s implementation of rater training, data management, and monitoring of the Vineland-II^TM in three pharmaceutical industry sponsored trials (one USA trial in ASD, one USA trial in Down syndrome, and one international trial in Down syndrome (USA, France, and Spain). Data from 596 Vineland- II^TMadministrations from 140 raters across 47 sites were available for processing and review.

Results: To our knowledge, these data reflect one of the largest compilations of Vineland-II^TMdata from industry sponsored clinical trials. A few sites implemented the software and data transfers without difficulty. Many sites had considerable difficulty using the software and transferring data accurately. Common problems included incorrect entry of date of birth, missing scores, incorrect application of basal and ceiling rules, and mislabeling of files for processing. Detailed metrics regarding the frequency of these issues and required site follow-up will be presented in the full poster.

Conclusions: By combining expert clinical review of Vineland-II data and specialized data management expertise, we implemented a process that resulted in the successful transfer and processing of nearly 600 Vineland- II^TM administrations. Several important conclusions follow from this work. First, sites and CRAs require considerable training on the implementation of the Vineland-II^TM scoring software to ensure accurate and timely data transfers. Second, clinical reviews need to be conducted on an ongoing basis during the study (e.g., identification of patterns of scoring for each patient population, review of outliers in data defined by 5^th and 95^th percentiles, and consistency of scores compared to other clinical assessments). Third, a clear process must be implemented to manage the scale-specific queries generated during the data reviews. Fourth, centralized data entry into the scoring program could be utilized instead of relying on site-level entry. Finally, our clinical quality assurance indicates that systematic efforts must be made to review data outliers, changes in raters across visits, unusual patterns of change scores, as well as identical scores across visits.