Design and implementation of an automated patient-care dashboard to provide individualized patient care data and quality metrics to emergency medicine residents

Abstract

Background

The emergency department (ED) is a high-stakes training environment for emergency medicine (EM) residents and residents' ability to reflect and self-evaluate patient care is of critical importance. Patient care dashboards have been shown to increase adherence to quality guidelines and improve patient outcomes. The objectives of this study were: (1) to create a comprehensive list of evidence-based, psychologically safe patient care and quality metrics to include in a patient care dashboard for EM residents; (2) to design an EM patient care residency dashboard in a secure, cloud-based environment integrated with the electronic health record (EHR); and (3) to pilot the usability and acceptability of the dashboard among EM residents.

Methods

We created a list of potential EM resident patient care metrics using ACGME Emergency Medicine Defined Key Index Procedure Minimums, leading EM quality indicators, and current EM dashboard literature. We surveyed PGY-1 to -4 EM residents at a single residency program for their recommendations about inclusion, exclusion, and the psychological safety of each metric. We then developed a dashboard utilizing Power BI software integrated with Epic EHR. After development, we conducted a 2-month pilot evaluation for usability and acceptability among EM residents utilizing a mixed-methods approach.

Results

We identified 41 metrics within five domains (productivity metrics, patient safety and leading quality indicators, key procedures, complex/high-acuity cases, and uncertain diagnosis) to consider for inclusion in the dashboard. Residents (n = 32/68; 47% survey completion rate) recommended inclusion of 33 metrics; among these, three were identified as moderate–high psychological risk (ED length of stay, patients per hour, death within 24 h) whereas the rest were considered low psychological risk. Based on these survey results, we created an EM resident patient dashboard using Microsoft Power BI. Over a 2-month pilot period with 16 residents, user data showed a change between each resident's prior patient care review practices and review practices when using a dashboard; specifically, there were notable variations in frequency of use, time spent per review session, number of patients reviewed per session, and data categories reviewed. Eleven of 16 residents completed the technology usability and acceptability survey, with general acceptability and few concerns on usability.

Conclusions

Our dashboard provides individualized patient care data to EM residents related to productivity, patient safety and quality, key procedures, complex/high-acuity cases, and uncertain diagnoses. A pilot group of EM residents found the dashboard acceptable and useable. Continued research is needed to explore ideal implementation and integration of patient care dashboards in residency training.