Recent efforts in medical settings to identify social determinants of health have focused primarily on screening for the purpose of improving care for individual patients and getting standardized data into electronic health records (EHRs). Relatively little attention has been given to processes needed to extract data on social determinants of health out of medical records with adequate validity and efficiency to facilitate analysis across individual encounters to inform population health efforts relevant to the health care sector. In this article we describe the rationale for extracting data on social determinants of health from EHRs, including the potential influence of aggregated data on quality improvement activities and health care payment reform. We then discuss opportunities and challenges to pulling these data from EHRs to enable population-level applications, focusing on the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, as one potential data aggregation resource. Standardizing methods for extracting data on social determinants of health from EHRs will require understanding current challenges and refining existing translation tools.