TL;DR
Get everyday helpers delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
A study of an ambient AI scribe at Beth Israel Lahey Health found clinicians spent less time actively working on notes after enrollment, while note length increased. Same-day and after-hours note closure and clinical productivity did not improve; researchers said heavier users saw a larger reduction in note-active time.
An ambient AI scribe was associated with less time clinicians spent actively working on notes, but did not improve same-day note closure, after-hours note closure or clinical productivity in a study of a 14-hospital health system. The findings, published in Annals of Internal Medicine, are based on clinicians’ use of Heidi at Beth Israel Lahey Health during 2025.
Across users, median note-active time fell from 21 minutes to 18.6 minutes after clinicians enrolled in Heidi. Median note length rose from 4,854 to 5,187 characters. The study describes an association following enrollment; it does not establish that the tool alone caused the changes.
The researchers, led by Jennifer P. Stevens, MD, of Beth Israel Deaconess Medical Center, analyzed outpatient clinicians who delivered direct patient care and averaged at least 10 encounters per month. Of 4,993 eligible clinicians in 2025, 1,140 enrolled in Heidi between January and November. After exclusions, 902 clinicians were included; 67% used the tool at least once, while 33% did not use it after enrolling.
Usage level mattered for the note-time measure. Among clinicians classified as high users during their first three months, note-active time decreased by 2.50 minutes, compared with a 0.29-minute decrease among enrolled nonusers and a 0.78-minute decrease among low users. The high-user reduction was significantly larger than both comparison groups; the low-user change did not differ significantly from that of enrolled nonusers.
Time Saved Did Not Lift Productivity
The results speak to a key question for health systems adopting AI documentation tools: whether reduced time spent on one task translates into more completed work or less work outside clinic hours. In this study, the answer was not apparent in the measured outcomes. Note-active time declined, but clinicians did not close notes sooner on the same day or after hours, and the researchers found no improvement in clinical productivity.
That distinction matters to clinicians and organizations weighing the cost and workflow demands of AI scribes. A shorter period actively working on a note may ease one part of documentation work, but it does not by itself show that the overall administrative burden has fallen. The accompanying editorial also cautioned that editing and review may offset initial time savings, and that the study did not measure time spent editing notes inside Heidi before they were transferred to the electronic health record.
The increase in note length is another open issue, not proof of better or worse documentation. Longer records could reflect more detail, but the study does not establish what that extra text contained or whether it changed care. Those questions matter because clinicians and patients rely on notes as a record of visits and a source of information for later care.
Products Worth Considering
Amazon Kindle Scribe Colorsoft 64GB (1st Generation)
Apple iPad Air 13-inch (M4), 128GB, Wi-Fi 7, Space Gray
Jackery Explorer 2000 v2 Portable Power Station with Two 200W Solar Panels
As an affiliate, we earn on qualifying purchases.
How Heidi Was Introduced
Beth Israel Lahey Health launched Heidi in January 2025, rolling it out in phases. Participation was voluntary, and the system was implemented separately from the health system’s electronic health record so clinicians could use it within their workflows. Eligible staff included physicians, non-physician practitioners, trainees and allied health professionals.
The researchers compared outcomes before and after enrollment among clinicians who signed up, and also examined differences by use level, using enrolled nonusers as a comparison group. The study comes amid continuing concern about clinician burnout: its authors cited a figure of 43.2% of clinicians reporting burnout symptoms in 2024. They noted that electronic health record work can compete with patient-facing time and extend beyond work hours. The study tested documentation and productivity outcomes at one health system; it does not establish that the same results will apply across other organizations or AI scribe products.
“By reducing documentation burden, AI scribes offer a promising strategy to address a contributor to EHR-related clinician burnout.”
— Jennifer P. Stevens and colleagues
Editing Time and Note Quality
The study did not capture the time clinicians spent reviewing and editing AI-generated notes within Heidi before transferring them to the EHR. That leaves the net time saved uncertain: some documentation work may have shifted to a stage the analysis did not measure. The findings also do not explain why notes grew longer or whether their accuracy, completeness or usefulness changed.
The reported results concern one health system and a particular implementation. The source material does not establish how the tool affected patient outcomes, access to care or health equity, nor whether AI-generated notes contained inaccuracies or omissions in this study. Ostherr raised those possibilities in her editorial, but they were not confirmed as outcomes by the reported analysis.
Further Tests of Scribe Effects
The study authors called for future research on productivity, documentation accuracy and health equity. A fuller assessment would also need to account for clinicians’ review and editing time and examine whether longer notes change the record of patient visits or later care. The reported study does not specify a date for follow-up research or a next milestone.
For health systems considering ambient scribes, the findings offer a measured result rather than a broad verdict: heavier users had a larger reduction in note-active time, while the measured productivity outcomes remained unchanged. Whether those time reductions persist, offset editing work or produce wider benefits remains to be established.
Key Questions
What did the study find about AI scribes?
At one 14-hospital system, use of Heidi was associated with lower median note-active time and longer notes after enrollment. The study found no improvement in same-day or after-hours note closure or clinical productivity.
How much did note-active time change?
Across users, median note-active time fell from 21 minutes to 18.6 minutes. High users had a 2.50-minute reduction, larger than the changes reported for enrolled nonusers and low users.
Did the study prove that AI scribes caused the changes?
No. The study compared outcomes before and after enrollment and reported associations. The findings do not by themselves establish that Heidi caused the changes.
Why are the results not a clear measure of total time saved?
The analysis did not measure time spent editing notes in Heidi before they were transferred to the EHR. The accompanying editorial said that review and editing could offset initial time savings, leaving the net effect uncertain.
What remains unknown about the longer notes?
The study reported an increase in median note length but did not establish why notes grew or whether the added text affected accuracy, care quality or patient outcomes.
Source: medpagetoday.com
