Tags

Type your tag names separated by a space and hit enter

Derivation and validation of a machine learning record linkage algorithm between emergency medical services and the emergency department.

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.

Authors+Show Affiliations

Department of Emergency Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.Department of Emergency Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA. Division of Clinical Informatics, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.Department of Computer Science, New York University, New York, New York, USA.Department of Emergency Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.Department of Emergency Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA. Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA.Department of Emergency Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA. Division of Clinical Informatics, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.Department of Emergency Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA. Division of Clinical Informatics, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA. Center for Healthcare Delivery Science, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.

Pub Type(s)

Journal Article

Language

eng

PubMed ID

31605488

Citation

Redfield, Colby, et al. "Derivation and Validation of a Machine Learning Record Linkage Algorithm Between Emergency Medical Services and the Emergency Department." Journal of the American Medical Informatics Association : JAMIA, 2019.
Redfield C, Tlimat A, Halpern Y, et al. Derivation and validation of a machine learning record linkage algorithm between emergency medical services and the emergency department. J Am Med Inform Assoc. 2019.
Redfield, C., Tlimat, A., Halpern, Y., Schoenfeld, D. W., Ullman, E., Sontag, D. A., ... Horng, S. (2019). Derivation and validation of a machine learning record linkage algorithm between emergency medical services and the emergency department. Journal of the American Medical Informatics Association : JAMIA, doi:10.1093/jamia/ocz176.
Redfield C, et al. Derivation and Validation of a Machine Learning Record Linkage Algorithm Between Emergency Medical Services and the Emergency Department. J Am Med Inform Assoc. 2019 Oct 12; PubMed PMID: 31605488.
* Article titles in AMA citation format should be in sentence-case
TY - JOUR T1 - Derivation and validation of a machine learning record linkage algorithm between emergency medical services and the emergency department. AU - Redfield,Colby, AU - Tlimat,Abdulhakim, AU - Halpern,Yoni, AU - Schoenfeld,David W, AU - Ullman,Edward, AU - Sontag,David A, AU - Nathanson,Larry A, AU - Horng,Steven, Y1 - 2019/10/12/ PY - 2019/05/04/received PY - 2019/08/15/revised PY - 2019/09/10/accepted PY - 2019/10/13/entrez KW - clinical informatics KW - electronic patient care records KW - emergency medical services KW - machine learning KW - patient matching KW - prehospital care KW - record linkage JF - Journal of the American Medical Informatics Association : JAMIA JO - J Am Med Inform Assoc N2 - 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. SN - 1527-974X UR - https://www.unboundmedicine.com/medline/citation/31605488/Derivation_and_validation_of_a_machine_learning_record_linkage_algorithm_between_emergency_medical_services_and_the_emergency_department L2 - https://academic.oup.com/jamia/article-lookup/doi/10.1093/jamia/ocz176 DB - PRIME DP - Unbound Medicine ER -