Trust but Verify: Discordance in Opioid Reporting Between the Electronic Medical Record and a Statewide Database.

Abstract

The electronic medical record (EMR) is often used as the primary source for patient medication lists and history. We sought to determine the accuracy of the EMR in documenting opioid prescriptions in patients undergoing fracture repair compared to a statewide database. This retrospective study was conducted at an urban level 1 trauma center. Patients > 18 years old were included if they were admitted directly through the emergency room with isolated single orthopedic injuries. Opioid use and prescription data prior to admission and three months following surgery were collected through the EMR and a California statewide database of controlled substance prescriptions. A 2 x 2 McNemar's test was used to identify discordance between the EMR and Controlled Substance Utilization Review and Evaluation System (CURES). A total of 369 patients were included. The EMR reported that 143 patients had an opioid prescription within 30 days prior to admissioncompared to 75 patients reported by CURES (discordance rate [DR]: 34.7%) (p < 0.001). Between postoperative days (POD) 0-30, the EMR reported that 367 patients had an opioid prescriptioncompared to 285 reported by CURES (DR: 22.8%) (p < 0.001). Between POD 30-60, the EMR reported that 142 patients had an opioid prescriptioncompared to 84 reported by CURES (DR: 34.7%) (p < 0.001). Between POD 60-90, the EMR reported that 83 patients had an active opioid prescriptioncompared to 60 patients reported by CURES (DR: 41.0%) (p = 0.10). There is a significant discordance between the databases in documenting opioid use. Physicians should check multiple sources to best assess active opioid prescriptions.