Regency EF
Developing Data Models and Documentation for Complex Longitudinal Studies: Lessons Learned from NCANDA (304071)
*Kevin M Cummins, University of California, San DiegoJoshua M Freilich, University of California, San Diego
Alyssa Lopez, University of Southern California
Keywords: data model, data documentation
Numerous large-scale national longitudinal studies of adolescent development have made their data available to the global research community. One study, which includes neuroimaging data is the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA). This study has a cohort sequential design, with several stages and contingencies that determines which assessments issued to a participant. In many instances, knowing more than a variable’s codes and variable label is necessary for the appropriate analysis and interpretation. There are over six thousand dimensions in the data with a notable fraction of derived variables. NCANDA’s interactive platform for presenting the study design, data documentation, and data access tools will be described and compared to platforms, workflows, and information provided by comparable studies of adolescent development. NCANDA’s platform was designed to fit into R and STATA workflows. Lessons learned in the development of NCANDA’s platform and populating of the database will be described. Comparisons of the utility and value of the various platforms will be made with the new online NCANDA data portal.