Building capacity to use disparate electronic medical records and clinical information systems for health outcomes research

Background: Electronic medical records (EMRs) have potential to be a viable source of data for health outcomes research in primary care. However, the use of EMRs for research has been made problematic by: the traditional design of and intent for EMRs to be used for treating patients on an individual basis; the opportunity for entry of nonstandard, free-text data; the lack of interoperability and data comparability across disparate EMRs. In partnership with the West Virginia (WV) Bureau for Public Health, the West Virginia University Office of Health Services Research (OHSR) has developed methodology to export a common de-identified dataset from multiple EMRs and clinical information systems used by WV safety-net clinics and merge these data into a standardized, public domain disease registry for analysis. Methods: OHSR uses a participatory approach in facilitating practice improvement and health information technology integration and use among WV safety-net clinics. Working with each site and corresponding EMR vendor, OHSR developed data export routines for multiple EMRs and uses the Chronic Disease Electronic Management System as a standardized, public domain registry for data aggregation. Once EMR data are transferred to CDEMS, de-identified data are shared with OHSR and analyzed using SAS. Results: Preliminary analyses focus on hypertension detection and control. OHSR developed a decision tree to: identify populations of patients by diagnosis; detect co-morbid conditions; detect whether patients are receiving regular care and documented blood pressure readings; categorize patients with hypertension according to Joint National Committee 7 staging; detect patients with consistently high blood pressure readings but no clinical diagnosis of hypertension as a potential indication of undiagnosed patients. Discussion/Conclusions: Exporting disparate EMR data into a common registry format provides the ability to examine health outcomes across sites and clinical information systems. These analyses, on aggregate and clinic-levels, can aid providers and public health planning.