AI-Assisted Documentation in Emergency Medicine
Date First Published:
July 17, 2026
Last Updated:
July 17, 2026
Report by:
Jessica Ziccarello MD; Tom Peterson MD, Senior EM Resident; EM faculty (Corewell Health/Michigan State University Emergency Medicine Residency Program)
Search checked by:
Jeffrey S. Jones MD, Research Director
Three-Part Question:
Among [emergency department clinicians] does the [use of AI-assisted charting tools] compared to [traditional charting] [reduce documentation time, improve accuracy and/or clinician efficiency]?
Clinical Scenario:
A 30-year-old emergency medicine resident is working an evening shift on a July weekend at a busy Level 1 Trauma Center. She’s already five charts behind, and more patients are filing into the waiting room by the second. As she heads in to see her next patient, she thinks about all the documentation that awaits her once her shift is over. She wonders if using the latest AI-assisted charting tool her residency program just rolled out will help her complete her documentation and get her home on time.
Search Strategy:
Medline 2020-06/26 using PubMed, Cochrane Library (2026), Scopus, and Embase
Search Details:
[("artificial intelligence" OR "AI") AND ("documentation" OR "scribes") AND ("efficiency" OR "accuracy" OR "burnout" OR "time") AND ("emergency")]. Limit to adults and English language
Outcome:
73 Studies were found; four addressed the clinical question.
Relevant Paper(s):
| Study Title | Patient Group | Study type (level of evidence) | Outcomes | Key results | Study Weaknesses |
|---|---|---|---|---|---|
| Ambient Artificial Intelligence Scribe Adoption and Documentation Time in the Emergency Department. Preiksaitis C, Alvarez A, Winkel M, Karamatsu M, Brown I, Sama N, Morris L, Lee JY, Gubbels A, Wahl E, Frye A, Rose C. May 2026 USA | 8,740 encounters, of which 976 (11.2%) used an ambient AI scribe | Single-center, retrospective observational study | Median documentation times and note characteristics between ambient AI scribe assisted and standard encounters | There was a 28% reduction in on-shift documentation time (2:45 min vs. 3:50 min) and 16% reduction in total EHR time (8:39 min vs. 10:21 min). | No randomization; a small group of high-frequency users accounted for the majority of AI encounters (70.5%); selection bias, as AI use clustered in telemedicine encounters, lower-acuity patients, and those not requiring interpreters. |
| Human vs. artificial intelligence in medical charting: a comparative study in the simulated emergency medicine context. Lipinski M, Smith E, Vaillancourt C. March 2026 USA | Emergency Medicine physicians and residents at single hospital | Mixed methods simulated study using standardized patient encounters | Documentation efficiency, note editing requirements, and user preference/satisfaction | Mean dictation time was 2:05 minutes versus 1:15 minutes for AI note editing, representing a 39% time savings (p < 0.0001). Of 240 total note sections across all encounters, 69% required no edits at all and 76% of edits were minor. 63% of participants preferred the AI-generated note over their own voice-dictated note (p < 0.001). | This study was done in a simulated setting with scripted scenarios, therefore limiting generalizability. There was also a very small sample size (16 participants, 3 scenarios each). |
| Ambient Artificial Intelligence Versus Human Scribes in the Emergency Department. Morey J, Jones D, Walker L, Lindor R, Schupbach J, Mullan A, Heaton H. May 2026 USA | 284 ED visits with human scribes and and 426 with AI-assisted charting | Prospective quality improvement (QI) pilot study | Note quality (evaluated using the Physician Documentation Quality Instrument (PDQI-9), scored by two blinded physicians), EHR time metrics, and physician note contribution | PDQI-9 scores were similar for adults, but AI scribes scored lower for pediatric patients (41.36 versus 42.25). More time was spent in the electronic health record notes section per patient when using AI scribes (adult: 4.3 versus 1.8 minutes; pediatric: 3.5 versus 1.6 minutes). Note length was similar but physicians contributed significantly more characters per note when using AI (adult: 60.1% versus 30.8%; pediatric: 62.3% versus 27.1%) | This study was conducted over a short period (2 months) and involved a very small study group (5 early-adopter physicians) in a single academic ED; therefore, it likely does not capture learning-curve effects. |
| Medical Scribe and Ambient Artificial Intelligence Impact on Emergency Physician Documentation Burden and Clinical Productivity. Dutta S, Guan-Ting You J, Dunham L, Cash R, Meeker M, White BA, Joseph JW. June 2026 USA | A total of 198,178 ED encounters over 4 hospitals, in which 8,489 (4.3%) used ambient AI scribes, 15,947 (8.0%) used human scribes, and 173,742 (87.7%) had no scribe | Retrospective cross-sectional observational study | Documentation time and clinical productivity over three groups: ambient AI scribe, human scribe, and no scribe | AI scribes were associated with a 1.6-minute reduction in adjusted median documentation time per note compared to no scribe (95% CI 0.3-2.9 minutes), while human scribes were associated with a 3.3-minute reduction compared to no scribe (95% CI 2.3-4.3 minutes). There was no significant difference in work relative value units (wRVUs) per shift hour among the three groups. | Low AI scribe adoption rate of only 4.3% of encounters. Physicians self-selected into scribe groups. No evaluation of note quality or accuracy. |
Author Commentary:
For emergency department clinicians, current evidence demonstrates that AI assisted notes can provide a modest documentation time savings. However, these early studies are uniformly limited by retrospective or simulation-based designs, small physician samples, single-center settings, and short observation periods, meaning that the true impact of AI scribes and how they affect physician efficiency, note accuracy, and patient safety remains largely unanswered. Overall, AI-assisted charting is promising but has not yet clearly demonstrated reduced documentation time or superior accuracy in real-world ED practice, and still requires close human oversight.
Bottom Line:
While AI-assisted charting in the ED is promising, early data show some reduction in documentation time; however, these time savings have not yet been shown to translate into increased clinical productivity or throughput, and broader evidence comes mostly from non-ED settings.
Level of Evidence:
Level 3: Small numbers of small studies or great heterogeneity or very different population
References:
- Preiksaitis C, Alvarez A, Winkel M, Karamatsu M, Brown I, Sama N, Morris L, Lee JY, Gubbels A, Wahl E, Frye A, Rose C.. Ambient Artificial Intelligence Scribe Adoption and Documentation Time in the Emergency Department.
- Lipinski M, Smith E, Vaillancourt C.. Human vs. artificial intelligence in medical charting: a comparative study in the simulated emergency medicine context.
- Morey J, Jones D, Walker L, Lindor R, Schupbach J, Mullan A, Heaton H.. Ambient Artificial Intelligence Versus Human Scribes in the Emergency Department.
- Dutta S, Guan-Ting You J, Dunham L, Cash R, Meeker M, White BA, Joseph JW.. Medical Scribe and Ambient Artificial Intelligence Impact on Emergency Physician Documentation Burden and Clinical Productivity.