Clinical research studies often leverage various heterogeneous data sources including patient electronic health record, online survey, and genomic data. We introduce a graph-based, data integration and query tool called Carnival. We demonstrate its powerful ability to unify data from these disparate data sources to create datasets for two studies: prevalence and incidence case/control matches for coronary artery disease and controls for Marfan syndrome. We conclude with future directions for Carnival development.
Keywords: biomedical research; cohort studies; information storage and retrieval.