Crossover Evaluation of Two Ambient AI Scribe Tools in the Emergency Department

Abstract This study aimed to compare two ambient AI documentation tools, Dragon Ambient eXperience (DAX) and Abridge in the emergency department (ED), assessing perceived effects on work burden, usability, documentation quality, satisfaction, and overall preference. We conducted a single-site, prospective crossover study in an ED over 6 weeks, from April to June 2025. Out of 20 faculty enrolled, 18 completed both phases. Participants used both ambient AI scribe tools in alternating 3-week phases. Pre-tool, tool-specific, and post-tool surveys captured four domains: burden, usability, quality, and satisfaction. Adoption was the proportion of notes containing any ambient output. Paired Wilcoxon tests and linear mixed-effects models were used to compare tools, adjusting for order and adoption. DAX was associated with greater reduction in overall perceived work burden compared with Abridge (median: 1.5 vs. 2; p = 0.025). Usability was high and comparable between the tools (SUS medians: 73.5 vs. 73.5, p = 0.94; UMUX-Lite medians: 86 vs. 82.5, p = 0.079). Scores from a modified version of the Physician Documentation Quality Instrument (PDQI-9) favored DAX (median: 39 vs. 36.5; p = 0.011). DAX received higher satisfaction ratings (median likelihood-to-recommend: 9 vs. 7.5; p = 0.015), but adjusted models suggested these differences reflected order effects more than inherent tool differences. Post-pilot preferences showed no overall preference after accounting for order, with first-tool exposure significantly shaping ratings. In this 6-week crossover study in the ED, both ambient AI scribes were highly usable, and perceived to reduce documentation burden while preserving note quality. Findings support the feasibility and perceived value of ambient AI scribes in the ED and motivate larger, longer-duration, multi-site evaluations with objective outcomes.