Switching emergency physicians to RVU-based pay changed nothing measurable

Design

observational study with two-way fixed effects and two staggered difference-in-differences models, 2019-2024

Population

167 general emergency departments, 2191 site-quarters

Primary outcome

RVUs per patient, patients per hour, left-without-being-seen rate

Effect

no significant effect on any primary outcome; advanced practice provider hours rose relative to physician hours in some models

Whether payment model drives clinician behaviour is usually argued from theory rather than data. This study used six years of operational data from 167 general emergency departments, 2191 site-quarters in total, comparing departments that moved from hourly to relative value unit-based physician compensation against those that stayed hourly, using two-way fixed effects and two staggered difference-in-differences specifications, adjusted for acuity, boarding hours, volume and payor mix.

There was no significant effect on relative value units per patient, patients per hour, or left-without-being-seen rates. Nor on length of stay for discharged patients, admission rates, 72-hour returns with admission, CT usage, the proportion of visits billed as critical care, net promoter score or attrition. The one signal was that advanced practice provider hours rose relative to physician hours in some models - a staffing response rather than a behavioural one, and inconsistent across specifications.

The result cuts against both sides of a familiar argument. Those who expect productivity pay to lift throughput did not get it; those who expect it to drive upcoding, over-imaging and unsafe discharge did not get that either. In a department where boarding and volume dominate every operational metric, what an individual physician is paid per patient may simply not be the binding constraint.

• Do not expect a compensation change alone to move throughput or left-without-being-seen rates.
• The safety fears about productivity-based pay - upcoding, CT overuse, unsafe early discharge - were not borne out here.
• Watch the skill-mix effect: advanced practice provider hours rose relative to physician hours in some models.
• If throughput is the problem, look at boarding and inflow, which dominated the covariates.
• This is US billing-model data; the RVU construct has no direct equivalent in Indian practice, though the underlying question about incentive-driven behaviour does.

Why it matters

It undercuts both the productivity case for RVU-based pay and the patient-safety case against it, with operational data rather than argument.

Don't overread it

This is observational operational data, not a randomised comparison, and sites that chose to switch may differ from those that did not in ways the covariates do not capture.

The statistics, in plain English

Difference-in-differences compares the change at departments that switched with the change at those that did not, so anything affecting emergency medicine generally over those six years drops out. Staggered models handle sites switching at different times, which is where naive before-and-after analyses go wrong. A null result across this many sites and quarters is reasonably strong evidence of no large effect, though a small one could still hide.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.