Art of analytics: Balancing electronic data and human knowledge for appropriate clinical decision making

Electronic health records serve diverse purposes. While designed principally for documenting individual patient care, electronic health records offer the potential for organizational-wide surveillance, data sharing across care sites, ongoing reporting of complex metrics tied to quality improvement and reimbursement, and practice based research. Each of these extensions naturally support data-driven public health efforts and the national effort in achieving the triple aim of improving patient care, improving population health, and reducing health care costs. The analytics potential of electronic health records is especially promising given the complexities of medical care and the need for timely clinical decision support to physicians, nurses, and the care team. The need for certified electronic health records has become increasingly clear, and federal funding for the purchase and implementation of these systems supports that notion. What is not receiving sufficient attention, however, is the need for a heightened level of data maturity in primary care. Health analytics is generally equated with specialized software, run only by trained individuals with specialized skill sets. Likewise, health analytics is often equated with “big data,” “data warehouses,” and enterprise systems automating the handling and synthesis of health care data. This focus on software overshadows essential considerations of one’s knowledge, skill, and ability to carefully consider and interpret results from electronic health records, validate the results presented, and apply those data to patient care. The West Virginia University School of Public Health Office of Health Services Research has worked for more than a decade with state-wide primary care centers in fostering an enhanced level of data maturity. Efforts begin with attention to the quality of the data stored in electronic health records and extend to the application of those data in selecting appropriate targets for change, reliably measuring change over time, and asking research questions of the data. This work is helping to provide the foundation for West Virginia’s first Practice Based Research Network which has recently been registered with the Agency for Healthcare Research and Quality. Attending to the fundamentals of data management, analysis, interpretation and action complement the intuition and experiential knowledge of the health care team and add value to the reporting and clinical decision support tools afforded through health analytics.