Back to 2025 Abstracts
Attitudes Towards Ambient Artificial Intelligence Scribes in an Academic Urology Practice
Sharath Reddy, MD, Aleksandra M. Golos, BS, Mursal Gardezi, MD, Thomas V. Martin, MD, Marianne Casilla-Lennon, MD.
Yale School of Medicine Department of Urology, New Haven, CT, USA.
Background: Medical scribes improve productivity and reduce clinician burnout. Recently, ambient artificial intelligence (AI) scribes have emerged. We assessed clinician experiences with an AI scribe trialed in an academic urology department.
Methods: Four urology attendings and three advanced practice providers (APPs) used an ambient AI scribe in clinic. An initial survey administered at one month and a follow-up survey at ten months collected practice characteristics, assessments of user experience (Table 1), and suggestions for improvement. Results were analyzed using descriptive statistics and independent two-sample t-tests.
Results: All participants completed both surveys. Three (75%) attendings and two (67%) APPs were still using the AI scribe at follow-up; increased efficiency was the most commonly cited reason. Reasons for discontinuation included lack of pre-charting functionality and minimal impact on efficiency. Of the two attendings with prior scribe experience, one was still using the AI scribe at follow-up and the other returned to a human scribe. Assessments of the scribe were more favorable at follow-up, especially in the “Workload” domain, though differences were not statistically significant (Table 1). APPs initially had more favorable assessments compared to attendings (though differences were not significant), but they were comparable at follow-up (Table 2). Suggestions for improvement included integration of dot phrases, labs, and imaging; and better documentation of physical exams and risk-benefit discussions.
Conclusions: Urology attendings and APPs reported positive experiences with an ambient AI scribe, citing efficiency as a key benefit. Improvements in functionality and documentation quality may enhance adoption and satisfaction.
Assessment Domains, Statements, and Ratings among Clinicians Using the Ambient AI Scribe. | | | | | | | | | |
| Domain | | Assessment Statements | | Clinicians Using AI Scribe at One Month (N=7) | | Clinicians Using AI Scribe at Ten Months (N=5) | | P-Value |
| Navigation | | 1. The ambient AI scribe mobile app is easy to use.
2. The ambient AI scribe Epic app is seamlessly integrated into my clinic workflow. | | 3.2 | | 4.3 | | 0.188 |
| Functionality | | 1. The ambient AI scribe provides accurate notes for my clinic visits.
2. The ambient AI scribe structures my clinic notes clearly.
3. The ambient AI scribe is able to accurately summarize clinical information pertinent to urologic patients.
4. Notes generated by the ambient AI scribe are comparable to those generated by a scribe (N/A if no prior scribe). | | 3.3 | | 4.0 | | 0.300 |
| Workload | | 1. Using the ambient AI scribe has reduced my clinic workload.
2. Using the ambient AI scribe has reduced the number of hours I spend outside of work finishing documentation.
3. It takes less than 10 minutes to edit my ambient AI scribe-generated notes. | | 3.1 | | 4.7 | | 0.052 |
| Clinical Impact | | 1. Using the ambient AI scribe during clinical encounters allows me to spend more time focusing on the patient.
2. Overall, the ambient AI scribe has had a positive impact on my clinic experience. | | 3.4 | | 4.4 | | 0.278 |
| Responses were on a 1-5 Likert scale (1 = “Strongly disagree” ⋯ 5 = “Strongly agree”). |
Summarized Average Domain Ratings by Clinical Role, Initial and Follow-Up Surveys. | | | | | | | | | | | | | |
| | | Initial Survey (N=7) | | Follow-Up Survey (N=7) |
| | | Attending (n=4) | | APP (n=3) | | P-Value | | Attending (n=4) | | APP (n=3) | | P-Value |
| Navigation | | 2.4 | | 4.3 | | 0.07 | | 3.6 | | 4.0 | | 0.73 |
| Functionality | | 2.5 | | 4.3 | | 0.08 | | 3.6 | | 3.8 | | 0.84 |
| Workload | | 2.3 | | 4.2 | | 0.09 | | 4.1 | | 3.8 | | 0.80 |
| Clinical Impact | | 2.8 | | 4.3 | | 0.19 | | 3.8 | | 4.0 | | 0.85 |
Back to 2025 Abstracts
