Statistical Methods for Electronic Health Records

This course provides an introduction to the statistical and methodological issues that arise when conducting research using electronic health record (EHR)-derived data and other forms of real- world health data. Students will learn how these data are generated through clinical care and administrative processes, how they are transformed into research-ready datasets (e.g., common data models, linkage to claims or registry data), and how their unique features influence study design, analysis, and interpretation. The course highlights the longitudinal, high-dimensional, and often irregular nature of EHR data, as well as common challenges such as missing data, measurement error, coding variability, and incomplete capture of outcomes and exposures. Students are expected to have taken a two semester course sequence in statistical inference and to have a working knowledge of R.