Will electronic health record data become the standard resource for clinical research?

Electronic health record (EHR) systems are increasingly common in most of the developed world (1). Clinic encounters, in the age of EHRs, generate hundreds of data elements used in high impact, innovative research. With many health care systems having 10–20 years of EHR utilization, longitudinal analysis can be conducted using standardized measures of medical diagnoses, laboratory results, medications, vital signs, provider type and demographic information (2,3). The size of existing data and promise for growth and linkage of databases across the United States offers an ‘unprecedented’ (4) opportunity for retrospective cohort designs, and facilitating identification of eligible subjects and recruitment to randomized trials (5). These databases can be used to exponentially increase the rate of discovery when contrasted with chart abstraction or primary data collection. In the United States, the Patient-Centered Outcomes Research Institute’s PCORNet (6) links millions of patient records from health care systems across the United States and the Health Care Systems Research Network (http://www.hcsrn.org/en/About/) has a long history of using EHR data to conduct research with standard variable definitions from large health care organizations across the USA. More specific to primary care, large data bases capturing numerous ambulatory family medicine and general internal medicine clinics are available through the DARTNet Institute (http://www.dartnet.info/). These networks allow researchers to collaborate and test hypothesis in multiple cohorts or, as appropriate, in large merged samples derived from entire health care systems to primary care practices as possible with the DARTNet Institute (http://www.dartnet.info/).