Development and validation of a heart failure with preserved ejection fraction cohort using electronic medical records

• DOI: 10.1186/s12872-018-0866-5
• Publication Date: 2018-06-28T00:00:37Z
• Authors: Yash R. Patel; Jeremy M. Robbins; Katherine E. Kurg...; Tasnim Imran; Ariela R. Orkaby; Robert R. McLean; Yuk-Lam Ho; Kelly Cho; J. Michael Gaziano; Luc Djousse; David R. Gagnon; Jacob Joseph
• Keywords: Angiology

Heart failure (HF) with preserved ejection fraction (HFpEF) comprises nearly half of prevalent HF, yet is challenging to curate in a large database electronic medical records (EMR) since it requires both accurate HF diagnosis and left ventricular (EF) values be consistently ≥50%. We used the national Veterans Affairs EMR cohort HFpEF patients from 2002 2014. EF were extracted clinical documents utilizing natural language processing an iterative approach was refine algorithm for verification HFpEF. The final utilized following inclusion criteria: any International Classification Diseases-9 (ICD-9) code (428.xx); all recorded ≥50%; either B-type natriuretic peptide (BNP) or aminoterminal pro-BNP (NT-proBNP) OR diuretic use within one month HF. Validation performed by 3 independent reviewers doing manual chart review 100 cases controls. established 80,248 (out total 1,155,376 ICD-9 HF). Mean age 72 years; 96% males 12% African-Americans. analysis had sensitivity 88%, specificity 96%, positive predictive value negative 87% identify cases. developed sensitive, highly specific detecting database. This may applicable other databases patients.