SHM Converge Scientific Abstract ePoster Gallery-Innovating to Improve Hospitalist Feedback on Clinical Decision Support Alerts

Pdf Summary

This project evaluated new ways to collect real-time hospitalist feedback on clinical decision support (CDS) alerts. The motivation was that electronic health records and alert fatigue can contribute to clinician burnout, reduce satisfaction, and negatively affect patient safety. Traditional feedback methods such as surveys or focus groups often lack specificity and are difficult to act on quickly.<br /><br />In collaboration with Clinical Informatics, CDS, and Hospital Medicine teams, the health system created a five-question custom survey to gather immediate feedback on high-priority alerts. The survey was embedded into seven selected alerts in late July 2024. It assessed alert utility, usability, clinical relevance, and alignment with the “Five Rights of CDS” framework. The system later shifted to a vendor-provided feedback tool that was rolled out across approximately 400 alerts in October 2024.<br /><br />The custom survey collected 72 responses between July and October 2024. After the vendor tool launched, it generated 1,956 responses by March 28, 2025, averaging about 20 to 70 responses per day. Alert ratings generally ranged from 1.4 to 2.2 out of 5, and the most common score was 1. However, 43% of responses did not include a numeric rating, making analysis more difficult. Free-text comments still provided useful positive and negative insights, but responses with both a rating and comments were the most actionable.<br /><br />A key strength of the feedback system was closed-loop communication: the Clinical Informatics team acknowledged feedback from users, and a spike in unrated responses helped identify and fix a new alert quickly after deployment.<br /><br />Overall, embedding feedback links directly into alerts greatly increased the volume and usefulness of clinician input. The authors conclude that meaningful feedback, timely analysis, and action on results are essential to improve CDS quality, reduce alert fatigue, and support future AI-driven clinical tools.

Asset Subtitle

Leigh Anne Goodman

Meta Tag

Author List Corneliu Antonescu, Leigh Anne Goodman, Nikhil Sood, Vikeen Patel

Category Innovations

Concept Clinical Decision Support

Concept Alarm Fatigue

Concept EHR Alert

Concept CDS Improvement

Concept Closed-Loop Communication

Distinguished Non-Finalist

Presenter Organization University of Arizona College of Medicine - Phoenix

Presenting Author Leigh Anne Goodman

Track Technology in Hospital Medicine

Keywords

clinical decision support

hospitalist feedback

alert fatigue

electronic health records

clinician burnout

custom survey

vendor feedback tool

alert usability

free-text comments

closed-loop communication

Clinical Decision Support

Alarm Fatigue

EHR Alert

CDS Improvement

Closed-Loop Communication