Operationalizing Research Questions with EHR Data

Authors: Nicole Weiskopf, Ayo Babatunde, Steven Bedrick, Mary Roberts Davis

Operationalizing research questions and hypotheses using electronic health record (EHR) data remains a major methodological bottleneck in clinical and epidemiologic research. A core challenge is not simply identifying the “correct” EHR variable, but rather acknowledging that multiple defensible operationalizations often exist for the same concept. This talk will describe a conceptual approach grounded in theoretical substruction to guide the translation of research questions to EHR variables. This approach may be interwoven with the use of informatics methods and resources to yield practical artifacts that can be leveraged to translate between different data models and standards. Specific examples and common challenges using the OMOP common data model will be included.