Colby Redfield1, Abdulhakim Tlimat1,2, Yoni Halpern3, David W Schoenfeld1, Edward Ullman1, David A Sontag4,5, Larry A Nathanson1,2, Steven Horng1,2,6. 1. Department of Emergency Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA. 2. Division of Clinical Informatics, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA. 3. Department of Computer Science, New York University, New York, New York, USA. 4. Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA. 5. Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA. 6. Center for Healthcare Delivery Science, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.
Abstract
OBJECTIVE: Linking emergency medical services (EMS) electronic patient care reports (ePCRs) to emergency department (ED) records can provide clinicians access to vital information that can alter management. It can also create rich databases for research and quality improvement. Unfortunately, previous attempts at ePCR and ED record linkage have had limited success. In this study, we use supervised machine learning to derive and validate an automated record linkage algorithm between EMS ePCRs and ED records. MATERIALS AND METHODS: All consecutive ePCRs from a single EMS provider between June 2013 and June 2015 were included. A primary reviewer matched ePCRs to a list of ED patients to create a gold standard. Age, gender, last name, first name, social security number, and date of birth were extracted. Data were randomly split into 80% training and 20% test datasets. We derived missing indicators, identical indicators, edit distances, and percent differences. A multivariate logistic regression model was trained using 5-fold cross-validation, using label k-fold, L2 regularization, and class reweighting. RESULTS: A total of 14 032 ePCRs were included in the study. Interrater reliability between the primary and secondary reviewer had a kappa of 0.9. The algorithm had a sensitivity of 99.4%, a positive predictive value of 99.9%, and an area under the receiver-operating characteristic curve of 0.99 in both the training and test datasets. Date-of-birth match had the highest odds ratio of 16.9, followed by last name match (10.6). Social security number match had an odds ratio of 3.8. CONCLUSIONS: We were able to successfully derive and validate a record linkage algorithm from a single EMS ePCR provider to our hospital EMR.
OBJECTIVE: Linking emergency medical services (EMS) electronic patient care reports (ePCRs) to emergency department (ED) records can provide clinicians access to vital information that can alter management. It can also create rich databases for research and quality improvement. Unfortunately, previous attempts at ePCR and ED record linkage have had limited success. In this study, we use supervised machine learning to derive and validate an automated record linkage algorithm between EMS ePCRs and ED records. MATERIALS AND METHODS: All consecutive ePCRs from a single EMS provider between June 2013 and June 2015 were included. A primary reviewer matched ePCRs to a list of ED patients to create a gold standard. Age, gender, last name, first name, social security number, and date of birth were extracted. Data were randomly split into 80% training and 20% test datasets. We derived missing indicators, identical indicators, edit distances, and percent differences. A multivariate logistic regression model was trained using 5-fold cross-validation, using label k-fold, L2 regularization, and class reweighting. RESULTS: A total of 14 032 ePCRs were included in the study. Interrater reliability between the primary and secondary reviewer had a kappa of 0.9. The algorithm had a sensitivity of 99.4%, a positive predictive value of 99.9%, and an area under the receiver-operating characteristic curve of 0.99 in both the training and test datasets. Date-of-birth match had the highest odds ratio of 16.9, followed by last name match (10.6). Social security number match had an odds ratio of 3.8. CONCLUSIONS: We were able to successfully derive and validate a record linkage algorithm from a single EMS ePCR provider to our hospital EMR.
Authors: Jason McMullan; Ryan Gerecht; Jordan Bonomo; Rachel Robb; Bryan McNally; John Donnelly; Henry E Wang Journal: Resuscitation Date: 2014-02-18 Impact factor: 5.262
Authors: Craig Newgard; Susan Malveau; Kristan Staudenmayer; N Ewen Wang; Renee Y Hsia; N Clay Mann; James F Holmes; Nathan Kuppermann; Jason S Haukoos; Eileen M Bulger; Mengtao Dai; Lawrence J Cook Journal: Acad Emerg Med Date: 2012-04 Impact factor: 3.451
Authors: Emil L Fosbøl; Christopher B Granger; Eric D Peterson; Li Lin; Barbara L Lytle; Frances S Shofer; Chad Lohmeier; Greg D Mears; J Lee Garvey; Claire C Corbett; James G Jollis; Seth W Glickman Journal: Am Heart J Date: 2013-01-22 Impact factor: 4.749
Authors: Jonathan Fix; Amy I Ising; Scott K Proescholdbell; Dennis M Falls; Catherine S Wolff; Antonio R Fernandez; Anna E Waller Journal: Public Health Rep Date: 2021 Nov-Dec Impact factor: 2.792