Benefits and Challenges of Linking Electronic Medical Record and Claims Data for Perinatal Research [28H

Abstract

INTRODUCTION: To improve understanding of the postpartum health services received in the 90 days following delivery for mothers of medically fragile infants (MMFI) we linked electronic medical record (EMR) data to commercial insurance claims. METHODS: We conducted a retrospective cohort study of mothers delivering liveborn infants at North Carolina Women's Hospital between 7/1/2014 and 6/30/2016 (n=6849). We linked data from the UNC Health System Epic EMR to member encounters from a commercial insurance provider to measure prevalent conditions, health care utilization, and receipt of recommended services among MMFI compared with mothers of well babies. A list of mothers names with the payor on their delivery account (n=1687), birthdate, delivery date, procedure code for delivery, delivery location, subscriber ID, and billed provider was provided to an honest broker for matching. RESULTS: The honest broker identified members with enrollment and a childbirth-related procedure code during the study period (n=90,490) and applied context-based blocking and fuzzy matching on combinations of first and last name for a 97% match rate (n=1642). Having a finite time-period (a childbirth-related encounter date within 1-2 days of that mother's EMR delivery date) and adding anesthesia procedure codes contributed to a successful linkage. Changes in the mother's medical record number made it challenging to query EMR data on the same woman longitudinally. CONCLUSION: Linking claims to EMR data can elucidate women's health utilization in the early postpartum period by allowing ascertainment of encounters outside of the hospital. Additionally our findings have contributed to improved data quality initiatives in the hospital's EMR system.