Ambient AI for Provider Well-Being

An EHR-embedded Pragmatic Stepped-wedge Individual Randomized Clinical Trial of Ambient Artificial Intelligence to Improve Provider Well-Being

The goal of this clinical trial is to learn whether using Ambient Artificial Intelligence for provider documentation will enhance provider well-being and improve documentation quality.

Participants will complete their documentation using the Ambient AI software.

Study Overview

• Status: Completed
• Conditions: Healthcare Provider Well-being
• Intervention / Treatment: Other: Artificial Intelligence

Detailed Description

Provider documentation of patient visits is a time-consuming activity that extends throughout the workday and often continues outside office hours. This ongoing clerical burden negatively impacts provider well-being, contributing to burnout and job dissatisfaction.

To address these challenges, the researchers propose to test the effectiveness of the Ambient AI tool in improving provider fulfillment and note documentation efficiency. The overarching goal is to leverage AI technology to enhance provider well-being and documentation quality.

• Study Type: Interventional
• Enrollment (Actual): 66
• Phase: Not Applicable

Contacts and Locations

Study Locations

• United States
  • Wisconsin
    • Madison, Wisconsin, United States, 53705
      • University of Wisconsin

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: Yes

Description

Inclusion Criteria:

• Healthcare Provider in the UW Health Outpatient Clinic Setting
• Willingness to engage and use ambient technology
• Owns an Apple mobile device, as the software is accessible only on this platform
• Adult (greater than 18 years)
• English or Spanish speaking
• Completed the training and in-servicing required to use the tool
• Providing outpatient care to no less than 20 encounters on a weekly average

Exclusion Criteria:

• Planned leave in 6 weeks following randomization
• Not registered onto Epic's mobile Haiku system for access
• Enrolled in a virtual scribe program and not willing to leave the program

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Health Services Research
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: None (Open Label)

Number of Arms

3

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: Ambient Listening Group 1; Participants randomized to this arm will start using Ambient AI at week 7.	Other: Artificial Intelligence; Ambient AI software intervention is implemented into the providers workflow. The software incorporates Automated Speech Recognition technology with Large Language Models to generate clinical documentation in real-time.
Experimental: Ambient Listening Group 2; Participants randomized to this arm will start using Ambient AI at week 13.	Other: Artificial Intelligence; Ambient AI software intervention is implemented into the providers workflow. The software incorporates Automated Speech Recognition technology with Large Language Models to generate clinical documentation in real-time.
Experimental: Ambient Listening Group 3; Participants randomized to this arm will start using Ambient AI at week 19.	Other: Artificial Intelligence; Ambient AI software intervention is implemented into the providers workflow. The software incorporates Automated Speech Recognition technology with Large Language Models to generate clinical documentation in real-time.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Change in provider fulfillment index	Using the Professional Well-Being Academic Consortium Survey Measures, the fulfillment index is a 6-item questionnaire. Each question is scored on a 0-4 Likert scale, with total possible scores of 0-24. Higher scores indicate greater fulfillment	Baseline to weeks 6, 12, 18, and 24
Change in provider burnout	Using the Professional Well-Being Academic Consortium Survey Measures, the burnout index includes two subcomponents of work exhaustion and interpersonal disengagement. It is a 10-item questionnaire, scored together as a composite. Each question is scored on a 0-4 Likert scale, with a total possible range of 0-40. Higher scores indicate greater burnout, indicating lower well-being	Baseline to weeks 6, 12, 18, and 24

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Change in time spent on documentation outside work hours	Providers will track amount of time spent on EHR (electronic health records) outside of scheduled patient hours per 8 hours of patient scheduled time	Baseline to 24 weeks
Change in task load	Hours spent on documentation per 8 hours of scheduled patient time	Baseline to 24 weeks
Change in meaningfulness of work	Meaningfulness of work is a 4-item questionnaire, where each question is scored on a 0-4 Likert scale. Total possible scores range from 0-16, with higher scores indicating lower sense of meaningfulness of work.	Baseline to 24 weeks
Change in meaningful relationships	Using the Professional Well-Being Academic Consortium Survey Measures, the negative impact of work on relationships index is a 4-item questionnaire. Each question is scored on a 0-4 Likert scale, with total possible scores ranging from 0-16. Higher scores indicate less meaningful relationships.	Baseline to 24 weeks

Collaborators and Investigators

• Sponsor: University of Wisconsin, Madison
• Investigators: Principal Investigator: Majid Afshar, MD, MS, University of Wisconsin, Madison; Principal Investigator: Joel Gordon, MD, University of Wisconsin, Madison

Study record dates

Study Major Dates

• Study Start (Actual): August 15, 2024
• Primary Completion (Actual): March 27, 2025
• Study Completion (Actual): March 27, 2025

Study Registration Dates

• First Submitted: July 18, 2024
• First Submitted That Met QC Criteria: July 18, 2024
• First Posted (Actual): July 24, 2024

Study Record Updates

• Last Update Posted (Actual): May 30, 2025
• Last Update Submitted That Met QC Criteria: May 28, 2025
• Last Verified: May 1, 2025

More Information

Terms related to this study

• Keywords: Artificial Intelligence; Physician well-being; Healthcare provider
• Other Study ID Numbers: 2024-1028; A534285 (Other Identifier: UW Madison); Protocol Version 7/7/2024 (Other Identifier: UW- Madison)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: YES
• IPD Plan Description: As part of the UW Learning Health System, the data queries, data dictionary, and data procedure will be mirrored over from the UW Health GitHub instance to SMPH GitLab instance for secondary research use. All patient data except timestamps will be deidentified (limited dataset) and managed and stored on SMPH servers and devices and provided through Clinical Research Data Service (CRDS) team in ICTR's Center for Health Informatics Institute and the UW SMPH Honest Broker. Provider data will be stored with identifiers for the follow-up focus groups and interviews.
• IPD Sharing Time Frame: Duration of storage will be at least 10 years.
• IPD Sharing Access Criteria: Access to data and recordings and security measures will require IRB approval and request through CRDS.
• IPD Sharing Supporting Information Type: STUDY_PROTOCOL

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No
• product manufactured in and exported from the U.S.: No