Abstract 18583: A Method for Improving the Identification of Heart Failure Patients for Quantitative Clinical Performance Measures using Electronic Health Records
Background: Electronic Health Records (EHR) are recommended for providing relevant data measures for quantifying and improving health care quality in pts with Heart Failure (HF). However, little is known on the feasibility and reliability of corroborating data fields automatically extracted from EHR to detect the HF population. We sought to improve the definition of pts diagnosed with HF using clinical and revenue functions of EHR.
Methods: A standardized approach was developed using Epic's Clarity application to extract and integrate expanded administrative raw data from two sources: “Problem List Diagnosis Files” (PLDF) and “Encounter Diagnosis Administrative Files” (EDAF). HF diagnosis was extracted from the combined data, with each source considered equally valid, by using the ICD-9-CM Diagnosis Code set and methodology recommended by the 2010 Heart Failure Working Group.
Results were compared against the “gold standard”, comprised through manual extraction of HF diagnosis by three clinicians, which was further validated by an experienced EHR research data extractor in a sample of 1000 in- and outpatients.
Results: 68 HF pts were identified by "gold standard manual extraction", while the administrative files (EDAF) detected only 50 pts; use of both sources detected HF 64 pts. Automated data extraction showed high specificity (0.99) for both methods, sensitivity increased from 0.72 with traditional administrative claims to 0.93 by corroborating EDAF with the PLDF.
Conclusions: Corroborating two automated extracted data fields: PLDF and EDAF as equally valid sources was a superior strategy to ensure that patients were not incorrectly excluded from quality measures. The detection of HF using ICD9 codes appeared feasible. Correct definition of the patient population with HF is a critical prerequisite for accurate assessment of all quantitative clinical performance measures.
- © 2012 by American Heart Association, Inc.