Capio Ramsay Santé
Capio Ramsay Santé rolls out Tandem's AI scribe; clinicians estimate 29% less time per note
Capio Ramsay Santé, a healthcare provider in Sweden, uses Tandem Health's AI medical scribe to draft clinical notes from patient visits. In a study commissioned by Tandem, 177 surveyed clinicians estimated that documentation time per note fell from 6.69 to 4.72 minutes.
Key figures
Copied exactly from the sources. Copy any figure with its citation.
6.69
to 4.72 minutes (-29%)
Clinicians' estimated documentation time per note, before versus after the scribe (self-reported; survey of 177 clinicians)1
2.41
to 3.14
Mean rating of 'ability to work without stress related to administrative tasks' on a 5-point scale, before versus after1
177
Fully onboarded clinicians who answered the before-and-after survey2
What it took
The problem
Clinicians spend a large share of their working day writing up patient visits. That cuts time with patients and adds to stress and burnout. Capio Ramsay Santé, an integrated healthcare provider in Sweden, wanted to ease this across primary care, specialist care and hospitals.
What they built
Capio rolled out Tandem Health's AI medical scribe. The tool listens to the consultation, turns it into text and drafts the clinical note. The clinician reviews and edits the draft. It also drafts documents and certificates, and clinicians can ask it in plain language to add or change content.
How they did it
According to Tandem, Capio first ran a pilot at ten clinics and then began a wider rollout. The study covered April 2024 to October 2025 across several specialties. For the survey, the authors included only clinicians who had used the scribe for at least 3 months and had created more than 100 notes. The study was registered in advance on the Open Science Framework, a public research registry.
Results
In the survey of 177 clinicians, estimated documentation time per note fell from 6.69 to 4.72 minutes, and the mean rating for working without administrative stress rose from 2.41 to 3.14 on a 5-point scale. Median editing time was 93 seconds per note and did not fall with continued use. The authors say the results may not apply to clinicians who stopped using the tool or never fully adopted it.
All reported figures
| Measure | Figure | Citation |
|---|---|---|
| Clinicians' estimated documentation time per note, before versus after the scribe (self-reported; survey of 177 clinicians) | 6.69 to 4.72 minutes (-29%)1 | |
| Mean rating of 'ability to work without stress related to administrative tasks' on a 5-point scale, before versus after | 2.41 to 3.141 | |
| Median editing time per note, a logged measure | 93 seconds1 | |
| Fully onboarded clinicians who answered the before-and-after survey | 1772 | |
| Notes analyzed from clinicians' editing records (1,295 clinicians), as in the peer-reviewed version | 236,1532 |
Lessons for your project
Ask clinicians for before and after estimates, but add a logged measure alongside. Here the time saving is self-reported while the editing time is recorded.
Pilot in a few clinics first. Capio started with ten before a wider rollout.
Measure adoption as well as time saved. The evaluation only counted clinicians who had used the tool for at least 3 months, so plan how to treat people who stop.
Track the stress and presence-with-patients effects too. The study rated both, and staff wellbeing is part of the business case.
Caveats
- Tandem Health commissioned the study. Three of its six authors work for Tandem and hold shares or options, one works for Capio and one is the spouse of a Tandem employee. This is not an independent evaluation.
- The documentation-time figures are clinicians' retrospective estimates, not logged timings. The authors say there is no baseline measurement of time and that survey respondents were a selected group of satisfied, sustained users.
- The preprint title gives a different note count (375 thousand), but its text reports 376,596 notes in the original data and 236,153 after filtering, the count in the peer-reviewed title. The time estimates come from 177 survey respondents, so the title figure is not used.
- The peer-reviewed article is read here through the abstract on Tandem's website, because the journal page could not be opened.
- No source gives the go-live date. The study window began on 26 April 2024, so the deployed date is shown as the year 2024.
- The tool's underlying AI model is not disclosed.
Sources
- Impact of an AI medical scribe after 375 000 notes generated across care levels in a European health system
medRxiv (preprint, not peer reviewed) · 2025-12-08 · Research paper · medrxiv.org
- Evaluation of an AI medical scribe after 236,153 notes generated across care levels in a European health system
Tandem Health (abstract of the article in JMIR Medical Informatics) · 2026-03-30 · Research paper · tandemhealth.ai
- Nordic rollout with Capio and European expansion underway
Tandem Health · 2025-03-27 · Press release · tandemhealth.ai
Summarized in our own words from the sources above. Spotted an error? See corrections.