Email Record: Yield and Bias In Defining a Cohort Study Baseline from Electronic Health Record Data