Rugged Terrain: The Challenges of Using EHR Across Systems

The purpose of this presentation is to discuss the rough road one must travel when trying to use electronic health record (EHR) data to do research across multi-hospital systems. This “rough road” is actually the merger of two uniquely challenging paths: (1) the path toward operationalizing EHR data to do significant research and (2) the path toward getting different, competitive hospital systems to collaborate. Starting down these trails means tackling some pretty brutal terrain—still, the CCNR initiative has begun tackling this terrain and all of its attendant challenges.

The EHR offers advantages over a paper-based health record system. While paper-based records contribute to fragmentation of care (patient encounters are documented on site-specific charts accessible only at the local level), the EHR accommodates collection of structured, coded, electronically available data that longitudinally records the story of a patient’s health experiences beyond the local level. The use of data elements for purposes other than clinical documentation and billing are growing in response to meaningful use of electronic health information mandated by national policy. The increasing use of EHR and their “meaningful use” have potential to improve the quality, safety, and cost of health care through immediate access and reuse of clinical data. So in conclusion—anecdotally and intuitively—EHR will improve health care and thereby patient health.

But moving beyond the anecdotal and the intuitive to the tangible and factual requires research. The CCNR has been mandated by its clinical partners to (a) use the EHR for all research projects and (b) develop a data-extraction/data-sharing mechanism to inform nurse leaders’ decision-making. Natural language processing and IT-hardware consultants have started working with the CCNR on (a) getting structured, semi-structured, and unstructured data out of the EHR, (b) making those data from different EHR systems mutually intelligible, and (c) moving/storing those massive datasets in a secure way (challenging path #1). Numerous healthcare systems have initiated (a) monthly meetings of representatives from all external clinical stakeholders, (b) establishment of consensus goals, and (c) frequent electronic contact. These measures have fostered a greater sense of inter-system trust (challenging path #2). Path #1 leads to extraction of various forms of data from different proprietary EHR systems and then translating those data into a uniform CCNR nomenclature. Path #2 leads to facilitation of collaborative clinical research, quality improvement, benchmarking, and real-time, data-driven decisions that improve healthcare outcomes. Again, the CCNR has begun hiking up both of these difficult paths—and we are making progress.

Data challenges still to be discussed include the accuracy, completeness, transformation, recoverability, provenance, and granularity. The concept of semantic harmonization and implications related to the conduct of research also must be addressed in greater detail.

The realities of doing research across different facilities using electronic health records, even when working from EHRs from the same vendor are extremely challenging. The CCNR has already made headway in this tough landscape. The CCNR stakeholders are energized to negotiate the challenges yet to come.