Impact of Ambient AI scribe tools on patient experience in outpatient clinical encounters

ISRCTN	ISRCTN60113237
DOI	https://doi.org/10.1186/ISRCTN60113237
Sponsor	UCLA Health
Funders	University of California, Los Angeles, National Center for Advancing Translational Sciences, National Institute on Aging

Submission date: 11/12/2025
Registration date: 30/01/2026
Last edited: 30/01/2026

Recruitment status: No longer recruiting
Overall study status: Completed
Condition category: Other

Prospectively registered

Protocol

Statistical analysis plan

Results

Individual participant data

Record updated in last year

Plain English summary of protocol

Background and study aims
Ambient AI scribe technologies are designed to improve physician workflow by automating documentation tasks during clinical encounters to reduce cognitive burden and reflect the fastest growing application of generative AI technologies in health care. Our previous study of a 2-month randomized controlled trial focused on the effect of ambient scribes on physician efficiency and burnout. However, there remains a critical evidence gap on the impacts on these technologies on patient-oriented outcomes. Our goal is to use standardized post-encounter patient surveys after the randomized implementation of AI scribes to understand how these technologies shape patients’ perspective of their physician.

Who can participate?
This study includes patients who had outpatient visits with physicians who took part in the parent randomized trial. Only encounters conducted in English are included due to the operational limitations of the initial pilot study.

What does the study involve?
The study builds on a previous pilot RCT that evaluated two vendor AI scribes against a control from 11/04/2024 to 01/03/2025. This study is a secondary analysis using completed CG-CAHPS surveys that patients completed after their visits. These surveys are linked to the provider and date of encounter. No new surveys, clinic visits, or procedures are required. We compare communication scores between visits with physicians who were assigned an AI scribe tool and visits with physicians who were assigned standard documentation without a scribe. All scores are adjusted for each provider’s baseline CG-CAHPS scores 6 months before the study period.

What are the possible benefits and risks of participating?
There are no direct benefits or risks to individual patients as all data come from routinely collected surveys. No patient-level intervention is performed.

Where is the study run from?
The study is conducted at UCLA Health ambulatory clinics in the United States.

When is the study starting and how long is it expected to run for?
The parent randomized trial ran from 04/11/2024 to 03/01/2025, and patient surveys were collected for up to 6 months after each eligible encounter. The end date of survey collection is 03/07/2025.

Who is funding the study?
The study is funded by the UCLA Department of Medicine, with additional support from NIH/NIA grants (R01AG070017-01, K76AG064392-01A1, and K24AG047899) and the NIH/NCATS UCLA CTSI (UL1TR001881).

Who is the main contact?
Dr Aaron Chin, atchin@mednet.ucla.edu

Contact information

Dr Aaron Chin
Public, Scientific

757 Westwood Plz.
Los Angeles
90095
United States of America

Phone	+1 (0)310 825 0775
Email	atchin@mednet.ucla.edu

Dr John Mafi
Principal investigator

1100 Glendon Ave, #908
Los Angeles
90024
United States of America

Phone	+1 (0)310 794 2278
Email	jmafi@mednet.ucla.edu

Study information

Primary study design	Interventional
Allocation	Randomized controlled trial
Masking	Open (masking not used)
Control	Active
Assignment	Parallel
Purpose	Health services research
Scientific title	Evaluation of patient experience in Ambient AI scribe encounters: a retrospective secondary analysis of a randomized controlled trial (AIScribe RCT)
Study objectives	Primary objective: To evaluate the impact of ambient AI scribe use on patient-reported communication quality, measured using the CG-CAHPS communication composite score. Secondary objective: To assess heterogeneity of patient experience across clinical and demographic subgroups (e.g., baseline communication performers, new vs established visits, provider sex, specialty).
Ethics approval(s)	Ethics approval not required
Health condition(s) or problem(s) studied	Quality of care/patient satisfaction
Intervention	This study is a retrospective secondary analysis of a completed randomized controlled trial in which physicians were randomized to one of two ambient AI scribe tools or usual documentation. Providers were randomized using covariate-constrained randomization based on baseline time-in-notes, burnout, and clinic volume. For this secondary analysis, the two scribe arms are combined to evaluate the effect of any ambient scribe use on patient experience. Intervention arms: 1. Nabla AI Scribe: A transcription-based ambient AI scribe capturing physician–patient dialogue and generating draft clinical notes integrated into the EHR. 2. Microsoft DAX Copilot: A transcription-based ambient AI scribe generating encounter summaries for EHR documentation. 3. Control: Usual physician documentation without an AI scribe. Methodology: Patient CG-CAHPS surveys completed within 6 months of eligible encounters were linked to providers and analyzed using provider-level intent-to-treat assignment. The primary outcome is the CG-CAHPS communication composite (0–100). Analyses use adjusted linear regression with provider-clustered robust standard errors. Only English-language encounters were included for survey linkage due to vendor limitations; this restriction applies at the data level and does not affect participant eligibility.
Intervention type	Other
Primary outcome measure(s)	Mean CG-CAHPS Communication Composite Score (0–100) measured using derived mean score from four CG-CAHPS communication items assessing whether the physician: (1) explained things in an easy-to-understand way, (2) listened carefully, (3) showed respect for what the patient had to say, and (4) spent enough time with the patient. Surveys linked to eligible encounters and analyzed at the provider level using intent-to-treat assignment. at Surveys completed within 6 months after encounters occurring during the intervention period (11/04/2024–01/03/2025). Baseline comparison uses surveys from 6 months prior to enrollment (5/4/2024–11/3/2024).
Key secondary outcome measure(s)	Exploratory subgroup analyses measured using the adjusted Mean CG-CAHPS Communication Composite Score, exploratory subgroup analyses will include: baseline communication performance (bottom third), new vs established visit, patient-reported usual provider, physician sex, generalist vs specialist, and high-adopting physicians (top 50th percentile of tool usage). at Surveys completed within 6 months after eligible encounters during the intervention period (11/04/2024–01/03/2025). Three-arm comparison of communication composite scores measured using the adjusted Mean CG-CAHPS Communication Composite Score, exploratory analyses of control vs Nabla vs Microsoft DAX Copilot group. at Surveys completed within 6 months after eligible encounters during the intervention period (11/04/2024–01/03/2025). Top-box CG-CAHPS Communication Composite Score measured using the adjusted Mean CG-CAHPS Communication Composite Score, exploratory analyses of binary outcome (maximal score vs non-maximal) at Surveys completed within 6 months after eligible encounters during the intervention period (11/04/2024–01/03/2025). CG-CAHPS, Overall Provider Rating measured using the overall provider rating (0–10) as part of the CG-CAHPS survey; continuous and binary (top-box) measures at Surveys completed within 6 months after eligible encounters during the intervention period (11/04/2024–01/03/2025). Single-item communication domain scores measured using subset analysis of the four items of the CG-CAHPS communication score: (1) explained things in an easy-to-understand way, (2) listened carefully, (3) showed respect for what the patient had to say, and (4) spent enough time with the patient. at Surveys completed within 6 months after eligible encounters during the intervention period (11/04/2024–01/03/2025).
Completion date	03/07/2025

Eligibility

Participant type(s)
Age group	Mixed
Lower age limit	18 Years
Upper age limit	120 Years
Sex	All
Target sample size at registration	300
Total final enrolment	238
Key inclusion criteria	Ambulatory care physicians within the UCLA Health system who held at least one half-day of clinic per week who participated in the parent randomized trial of ambient AI scribes (NCT06792890)
Key exclusion criteria	1. Trainee providers (residents, fellows, medical students) 2. Allied health professionals (RNs, NPs, PAs) 3. Attendings who work exclusively with trainees 4. Providers who used a human scribe during the study period
Date of first enrolment	04/11/2024
Date of final enrolment	03/01/2025

Locations

Countries of recruitment

United States of America

Study participating centre

UCLA Health Ambulatory Clinics (multiple outpatient sites)

United States of America

Results and Publications

Individual participant data (IPD) Intention to share	No
IPD sharing plan summary	Not expected to be made available
IPD sharing plan

Study outputs

Output type	Details	Date created	Date added	Peer reviewed?	Patient-facing?
Protocol file			15/12/2025	No	No

Additional files

48652_PROTOCOL.pdf: Protocol file

Editorial Notes

15/12/2025: Study's existence confirmed by UCLA IRB.