Studying Provider and Team Behaviors via EHR Audit Logs
Using EHR Audit Logs to Capture Contextual Contributors to Diagnostic Excellence
Sponsor: Gordon and Betty Moore Foundation
The digitization of health records and the metadata captured by these systems create a novel and exciting opportunity to rapidly advance our understanding of the diagnostic process and diagnostic excellence. Specifically, the ability to truly understand how frontline clinicians interact with clinical information when making diagnostic decisions, and how these interactions impact diagnostic quality, may be greatly improved by analysis of a largely-unused, highly-granular level of electronic health record (EHR) data (referred to as “audit log” data) that are captured primarily for administrative compliance purposes today. This grant to the University of California, San Francisco Office of Sponsored Research (UCSF) will leverage the use of audit log data to capture contextual dimensions of diagnostic assessment (including clinical decision-making, clinical team structure and clinical processes) from three health systems (UCSF, Stanford and Kaiser Permanente Northern California). By identifying key contextual factors that contribute to diagnostic performance, this work may ultimately improve the efficiency of diagnostic processes and reduce errors.
Image Viewing among Cardiologists and the Resulting Impact on Patient Outcomes
Traditionally, a radiologist interprets the image and generates a text-based imaging report with their interpretation of the findings. However, human interpretation and synthesis of primary imaging data into text reports is a process with inherent imprecision and inaccuracy. As a result, a cardiologist may choose to rely on the imaging report, the primary image, or both when making diagnostic and treatment decisions, and they also may choose to consult with a radiologist when doing so. We hypothesize that primary review of imaging (ideally with an imaging expert) improves care – and that by reviewing reports alone, information is missed that can help patient care. Our goal is to study the effect on inpatient outcomes of cardiologists viewing imaging reports compared to viewing images themselves (with or without an imaging expert – i.e., radiologist).
The University of California Behavioral Economics and Access Log Research Collaborative (UCBEAR)
The University of California Behavioral Economics and Access Log Research Collaborative (UCBEAR) is a joint venture between CLIIR and faculty in the Haas School of Business and the Department of Economics at UC Berkeley. The goal of this collaborative is to develop a rich dataset describing EHR users at UCSF, their day to day actions and the outcomes of their actions. These data will help researchers create a body of work that addresses questions about the impact of clinical training on note writing efficiency, the extent of cognitive biases in diagnostic processes and other explorations at the intersection of behavioral economics, medicine and clinical informatics.