Estimated reading time: 7 minutes · Last updated:
Memorial Hermann's emergency physicians who used Ambience Healthcare's ambient documentation software produced roughly $37,500 more in yearly productivity value than colleagues who went without it. That AI scribe value rests on a gap in work RVUs per clinical hour, stretched across 160 ten-hour shifts. It is a gross number, with subscription, implementation and other costs left out. KLAS Research published it in an ROI validation report that Ambience announced on 3 November. Users saw 1.7 more patients per 10-hour shift. The figure cannot show a hospital finance chief whether the tool pays for itself. The method behind it, though, is one a buyer can rerun on its own billing records.
Key takeaways
- Gross, not net: Memorial Hermann's $37,500 per emergency physician measures extra productivity value against non-users before any Ambience software or implementation cost is subtracted.
- Per-shift yardstick: The gain came from 160 ten-hour shifts, so per shift it works out to about $234, or close to $23 for each clinical hour. Any all-in vendor quote has to come in under that bar.
- Not a causal estimate: Ambience's Justin Beeker said the billing comparison did not adjust individually for patient acuity, staffing or physician tenure, so it shows relative performance rather than proof the tool caused the gain.
- Pilot became rollout: After a 15-month pilot with about 100 emergency physicians, Memorial Hermann widened the platform to more than 560 clinicians in emergency medicine and ambulatory care.
- Providence counts time: Providence CEO Erik Wexler said his physicians report saving two to four hours a day, while a JAMA Network Open study of 1,547 Providence clinicians found smaller, modest per-person gains.
Table of contents
- Key takeaways
- How Was the $37,500 Figure Calculated?
- What $37,500 Means per Shift and per Hour
- Why the Comparison Falls Short of Proving Cause
- How Providence Measures the Same Kind of Tool in Hours
- At-Risk Fees Turn Measurement Into Contract Terms
- Will the $37,500 hold up for other buyers?
- What to be careful about
- Frequently asked questions
How Was the $37,500 Figure Calculated?
Justin Beeker, Ambience's head of value attainment, walked reporter Nathan Eddy through the method, as first reported by Healthcare Finance News. The starting point was work RVUs generated per clinical hour. A work relative value unit is a measure of the physician effort behind a service, and it feeds into both physician pay and reimbursement.
Analysts took the hourly difference between Ambience users and non-users and scaled it to a year of 160 ten-hour shifts. Beeker called that shift count conservative. The underlying records were monthly summaries for each provider, supplied by the outside revenue cycle firm that bills for Memorial Hermann's emergency departments. Both groups worked in the same department during the same months.
KLAS built its report on long interviews with Memorial Hermann staff about the results and how they were measured. The health system combined clinician surveys, operational data and the independent billing figures. Users were set against two groups: physicians with no documentation support and physicians paired with human scribes. That second group gives a benchmark beyond simply doing nothing.
What $37,500 Means per Shift and per Hour
The annual total hides the number a finance team actually needs. Divide $37,500 by 160 shifts and the gross gain is roughly $234 a shift. Across about 1,600 clinical hours, that works out to about $23 per clinical hour.
The hourly rate travels better than the annual one. An emergency group whose doctors work more than 160 shifts a year would see a bigger yearly figure from the same hourly gap. A group leaning on part-time physicians would see a smaller one. Comparing a vendor quote against the hourly rate stops Memorial Hermann's staffing pattern from leaking into another system's budget.
Two forces drive the gain. One is volume, the 1.7 extra patients per shift. The other is coding: users recorded a 1.3% rise in work RVUs per encounter, meaning each documented visit was credited with slightly more physician effort.
Why the Comparison Falls Short of Proving Cause
Beeker was explicit about the limits. The billing comparison made no individual adjustment for how sick patients were, how departments were staffed or how long physicians had practised. He said the results describe how users performed against non-users and should not be treated as a causal estimate. The report also left documentation accuracy and clinical quality unmeasured.
That matters most for the 1.3% coding rise. Ambience presents it as more accurate capture of physician work, which is plausible when a fuller note records how complex a visit was. A sicker caseload among users could produce the same bump, and an unadjusted comparison cannot separate the two. Any productivity gain built on higher RVUs per encounter therefore calls for coding compliance review.
The design does carry safeguards. The revenue cycle partner applied one coding compliance standard to both groups, and clinicians still review and sign off on their own notes. Running both groups side by side in the same period strips out seasonal and site swings that a before-and-after study would absorb. Physician users gave the platform a Net Promoter Score of 78, based on 40 survey responses.
How Providence Measures the Same Kind of Tool in Hours
Providence offers a second benchmark in a different unit. Erik Wexler, president and CEO of the 51-hospital system, spoke on 16 November at a panel hosted by the Federal Reserve Board and the Federal Reserve Bank of San Francisco, where he said physicians report that ambient documentation gives them back between two and four hours every day.
Published research points lower. In May, JAMA Network Open published a study that tracked 1,547 Providence clinicians using ambient AI; it recorded less time on note-writing and less after-hours charting. Each clinician gained only a little, but the study concluded that across a big system those gains could amount to real relief. Memorial Hermann's survey responses lean the same way, with physicians reporting that more of their documentation got done during the visit.
The two systems are counting different things, and that creates a trap. Providence leads with time, partly self-reported; Memorial Hermann leads with dollars drawn from billing. If the hour a doctor saves is the hour that lets them see 1.7 more patients, a business case that books both the relief and the revenue has counted it twice. A crowded emergency department may care most about throughput, while a group fighting burnout may value hours returned.
At-Risk Fees Turn Measurement Into Contract Terms
Beeker added that Ambience ties part of its fees to outcomes measured on the customer's own data. Once money depends on the result, the choice of comparison group, the billing source and any acuity adjustment stop being research details. They decide what the hospital pays for.
Christopher Langan, MD, CEO of Memorial Hermann Health System Physician Organization, said the project began with physician workload and a pledge to measure whether the technology delivered. Any health system preparing a pilot of its own can turn that pledge into three questions. Once subscription, implementation and internal staff time are all counted, what does each physician cost, and is that figure above or below the roughly $234-a-shift benchmark? Which comparison group and billing feed will at-risk fees be judged against, and will the analysis correct for acuity, staffing and tenure? And who audits documentation accuracy and coding compliance, given that a productivity comparison measures neither?
| Measure | Memorial Hermann | Providence |
|---|---|---|
| Headline metric | $37,500 gross yearly value per ER physician | Two to four hours saved a day, per CEO Erik Wexler |
| Basis | Billing data, users vs non-users | Physician reports; JAMA Network Open study |
| Scale | About 100 ER physicians in pilot; 560+ clinicians now | 1,547 clinicians in the study; 51 hospitals |
| Main driver | 1.7 more patients per shift; 1.3% more work RVUs per encounter | Less note-writing and after-hours charting |
| Known limit | No acuity, staffing or tenure adjustment | Gains per clinician modest in the study |
Will the $37,500 hold up for other buyers?
The case for
- The method uses ordinary provider-level billing summaries, so other systems can repeat it on their own data.
- Comparing users and non-users in the same departments and months removes seasonal and site effects.
- Memorial Hermann expanded from about 100 pilot physicians to more than 560 clinicians after 15 months.
- Including a human-scribe comparison group gives a benchmark beyond no assistance at all.
The case against
- Without adjustment for acuity, staffing or tenure, part of the gap may reflect caseload rather than the software.
- The figure is gross, so net return depends on contract costs that vary by customer.
- Documentation accuracy and clinical quality were not assessed.
- Results come from emergency medicine only and may not carry over to ambulatory clinics.
What to be careful about
- A higher work RVU count per encounter can draw coding compliance scrutiny if it reflects documentation inflation rather than real complexity.
- Counting both time saved and extra visits in one business case can double-book the same physician hour.
- At-risk fee contracts measured on unadjusted comparisons can reward gains driven by case mix rather than the tool.
- Annualizing on 160 shifts can overstate or understate value for groups with different staffing patterns.
The bottom line
Memorial Hermann's $37,500 is an honest number with clear edges. It shows Ambience users out-producing non-users in the same departments and months, by 1.7 patients a shift and 1.3% more work RVUs per encounter. It does not show the software caused that gap, and it says nothing about cost. For a buyer, the useful figure is about $23 per clinical hour, set against an all-in quote. The harder work sits in the contract: when fees ride on outcomes, the comparison group, the billing source and the acuity adjustment determine what counts as success. The next real test is whether ambulatory clinics produce similar results.
What to watch
- Watch for a billing-data comparison from Memorial Hermann's ambulatory clinics, which would test whether the $37,500 survives different visit patterns; no date has been set.
- Watch whether Memorial Hermann or Ambience publishes an analysis adjusted for acuity, staffing and tenure; none has been announced.
- Watch how buyers write comparison groups and acuity adjustment into at-risk fee contracts with Ambience as more systems sign; no timeline has been given.
Frequently asked questions
Did Memorial Hermann compare AI scribes with human scribes?
Yes. Ambience users were compared with physicians who had no documentation help and with physicians working alongside human scribes. The published $37,500 figure is the gap against non-users, and separate results for the human-scribe comparison have not been released.
What did Memorial Hermann physicians think of Ambience?
Physician users gave the platform a Net Promoter Score of 78, drawn from 40 survey responses. They also said more of their documentation was finished during the patient encounter rather than afterwards.
How much time do AI scribes save doctors at Providence?
Physicians tell Providence CEO Erik Wexler they get back two to four hours daily. When a JAMA Network Open study measured 1,547 Providence clinicians, each one's savings came out modest, though the study found that the total could be meaningful across a large system.
Related reading