Phenotyping the Transfusion Management Experience of Trainees Using Electronic Health Records.

backgroundClinical experience is foundational to graduate medical education but is difficult to quantify. Self-reported case logs and scheduled rotations provide limited detail on residents' actual experiences. The analysis of electronic health record (EHR) data may provide a more comprehensive view of residents' actual clinical experiences. We aimed to quantify and characterize residents' direct clinical experiences with intraoperative transfusion management as a proof of concept for the utility of EHR data in anesthesiology graduate medical education.

methodsThis retrospective observational study aimed to measure the participation of anesthesiology residency graduates with intraoperative transfusion management at a single institution. Data pertaining to the University of Michigan cases submitted to the Multicenter Perioperative Outcomes Group (MPOG) data registry from January 1, 2012, to December 31, 2022, were analyzed. The primary outcome was participation in intraoperative transfusion management defined by the transfusion of any volume of homologous packed red blood cells (pRBC) documented within the intraoperative anesthesia record. Subgroup analysis characterized resident participation in severe intraoperative hemorrhage resuscitation defined as the transfusion of ≥4 pRBCs. Additional outcomes were participation in non-pRBC blood product transfusion and transfusion of cell-salvaged blood. Surgical case and resident-specific factors were analyzed for association with resident participation in intraoperative pRBC transfusion.

resultsThroughout the study period, 231 residents from 8 graduating cohorts were involved in 222,006 anesthetic cases of which 8529 (4%) included intraoperative pRBC transfusion. Over the entire course of residency training, residents on average participated in 37 (standard deviation [SD] = 8) cases with intraoperative pRBC transfusion with a range of 16 to 60. The mean (SD [range]) participation with severe hemorrhage resuscitation was 10 (3 [2-22]), participation with non-pRBC transfusion was 7 (3 [1-14]), and participation with cell salvaged blood was 13 (4 [3-26]). Five of 19 surgical services (cardiac, transplant, vascular, neurosurgery, orthopedic) accounted for the majority (5668; 66%) of intraoperative transfusion management experiences.

conclusionsVariation exists in the transfusion management experience among trainees at a single academic center. This study supports the feasibility of using EHR data to quantify the clinical experiences of residents beyond the assumptions underlying self-reported case logs and time on specialty rotations. Such analyses may allow for interventions to optimize the frequency, timing, and sequencing of clinical care activities to ensure optimal clinical experience.