Abstract 16040: Improving Heart Failure Surveillance by Corroborating Intra-Institutional Automated Electronic Data Fields Captured During Routine Clinical Care
Background: American Heart Association (AHA) 2020 impact goals rely on valid primary and secondary metrics used for detection and/or surveillance of patients with non-fatal cardiovascular disease, including heart failure (CHF). We assessed the feasibility and reliability of using an automated electronic method for corroborating detection of incident CHF events across existent sources of intra-institutional electronic health records.
Methods: Epic’s Clarity applications (AEEA): “Problem List Diagnosis Files” (PLDF) and “Encounter Diagnosis Administrative Files” (EDAF) were considered to be equally valid sources for detecting patients with CHF. We sought to validate the ability of the proposed method by detecting/following CHF in two different settings: cross-sectionally in the general patient population seen in clinic (S1), and longitudinally in cardiac patients at risk for CHF followed-up at our institution between 2002 and 2012 (S2). Results obtained using AEEA were compared against the “gold standard,” defined as the standardized manual extraction of CHF diagnoses by experienced electronic data extractors using evidence-based guidelines established by the AHA.
Results: We compared the use of AEEA with the gold standard for 1000 patients in the S1 setting and 668 patients in the S2 setting. Use of AEEA for detecting CHF showed high specificity (0.99) and sensitivity (0.93) for patients in the S1 setting, but lower specificity (0.92) and sensitivity (0.89) in S2 patients (Table 1). Information found through the manual search of “Progress Notes” accounted for 40% of the false negative cases in S2 patients.
Conclusions: Use of automated electronic corroborating intra-institutional data is an effective method for identification and surveillance of CHF. Clarity applications targeting unified clinical notes, available with upgrades to Epic 2010, should be further explored as a means of increasing performance of automated wide-scale CHF surveillance.
- © 2013 by American Heart Association, Inc.