Healthcare / Health system
McLeod Health: Drafting clinical notes from outpatient visit conversations
- Company
- McLeod Health
- Country
- United States
- Adoption stage
- In operation
- Source published
- In use from
- Date basis
- The date the source was published. It can differ from the date adoption started.
- How the source was checked
- Read the full source text
The work problem
According to a paper by McLeod Health authors in JMIR Medical Informatics, McLeod Health is a nonprofit regional health system with 7 hospitals and more than 1,200 providers serving 18 counties in North and South Carolina in the United States. The paper explained that such nonacademic health systems face financial and staffing pressures together with physician satisfaction problems and difficulties in recruitment and retention. It noted that ambient AI documentation tools, which create notes from visit conversations, could reduce these burdens, but that vendor selection is easily undermined by cognitive biases, unvalidated marketing claims and limited real-world testing.
Technology and data
According to the paper, in spring 2024 McLeod Health compared 4 vendors through tests that simulated real clinical visits. It created 15 outpatient visit scripts with organizational leaders playing patients, and physicians, revenue cycle experts and nonclinical reviewers scored the notes each tool produced from the visit conversations captured in real time, for accuracy, billing quality and readability, without knowing which vendor made them. The top 2 vendors demonstrated workflows integrated with the Epic electronic health record, and the final vendor was chosen after physicians' usability feedback and reviews by the clinical informatics and IT security teams. A 90-day pilot then ran across 5 outpatient specialties from October 2024, and system-wide implementation began in January 2025. Health system leadership directed that time saved in the pilot be used to reduce after-hours documentation rather than to increase patient volume.
Results
According to the paper, among the 23 providers analyzed in the pilot, after-hours documentation time (pajama time) fell by 35.4% (P=.054) and time spent in notes fell by 28.3%, saving 3.6 hours per provider per month. Established patient volume rose by 8.5% and the distribution of visit complexity codes shifted, and the authors estimated a revenue gain of 2,629 US dollars per provider per month. In the patient satisfaction survey, several items improved significantly, with the item on the provider listening up 8.4 points and the item on receiving enough treatment information up 10.4 points. Since system-wide implementation began in January 2025, 250 providers have joined in phases with an 81% adoption rate, and they have produced nearly 150,000 notes with the tool.
Limits and open questions
The analysis covers 23 providers from the top 30% for documentation efficiency in a single health system, and the authors said they could not rule out seasonal effects because there was no control group. The revenue gain is an estimate, and the body of the paper does not name the vendor used.
Sources
- Selecting, Scaling, and Measuring the Value of Ambient AI in a Nonacademic Health System: Multiphase Pilot Studypmc.ncbi.nlm.nih.gov, Accessed
Compiled from public sources. These are not results from ATF Works customers.