The Pluripotent Rendering of Clinical Data for Precision Medicine

Health care and biomedical research are awash in data. Traditional data warehouse methodologies do not scale to this challenge; nor do their schema match the variety of analytic use cases. An alternative model, which shreds data into well-formed constituent data elements, conformant with the emerging CIMI-FHIR standards and stored together with the complete, raw, source data using modern and scalable data utilities such as Hadoop and its derivatives, affords the creation of pluripotent data repositories. Such repositories can be leveraged to generate any number of data marts, registries, and analytic data sets, each of which "just in time" binds an appropriate use-case specific data model. We call this notion PiCaRD: Pluripotent Clinical Repository of Data. We believe such nimble biomedical data management strategies are crucial for Precision Medicine discovery and application.