Electronic medical record integration with a database for adult congenital heart disease: Early experience and progress in automating multicenter data collection. (1st October 2015)

• Record Type: Journal Article
• Authors: Broberg, Craig S.; Mitchell, Julie; Rehel, Silven; Grant, Andrew; Gianola, Ann; Beninato, Peter; Winter, Christiane; Verstappen, Amy; Valente, Anne Marie; Weiss, Joseph; Zaidi, Ali; Earing, Michael G.; Cook, Stephen; Daniels, Curt; Webb, Gary; Khairy, Paul; Marelli, Ariane; Gurvitz, Michelle Z.; Sahn, David J.
• Is Part Of: International journal of cardiology. Volume 196(2015)
• Page Start: 178
• Page End: 182
• Publication Date: 2015-10-01
• Subjects: Electronic health records -- Congenital heart disease -- Healthcare information systems -- Multicenter data collection; Cardiology -- Periodicals; Electronic journals; 616.12
• DOI: 10.1016/j.ijcard.2015.05.140
• Languages: English
• ISSNs: 0167-5273

Abstract

Abstract: Background: The adoption of electronic health records (EHR) has created an opportunity for multicenter data collection, yet the feasibility and reliability of this methodology is unknown. The aim of this study was to integrate EHR data into a homogeneous central repository specifically addressing the field of adult congenital heart disease (ACHD). Methods: Target data variables were proposed and prioritized by consensus of investigators at five target ACHD programs. Database analysts determined which variables were available within their institutions' EHR and stratified their accessibility, and results were compared between centers. Data for patients seen in a single calendar year were extracted to a uniform database and subsequently consolidated. Results: From 415 proposed target variables, only 28 were available in discrete formats at all centers. For variables of highest priority, 16/28 (57%) were available at all four sites, but only 11% for those of high priority. Integration was neither simple nor straightforward. Coding schemes in use for congenital heart diagnoses varied and would require additional user input for accurate mapping. There was considerable variability in procedure reporting formats and medication schemes, often with center-specific modifications. Despite the challenges, the final acquisition included limited data on 2161 patients, and allowed for population analysis of race/ethnicity, defect complexity, and body morphometrics. Conclusion: Large-scale multicenter automated data acquisition from EHRs is feasible yet challenging. Obstacles stem from variability in data formats, coding schemes, and adoption of non-standard lists within each EHR. The success of large-scale multicenter ACHD research will require institution-specific data integration efforts.