You're Probably Donating RVUs.

On your last shift, a chart may have earned a 99285 and received a 99284. You considered the dissection, pulled the outside records, read the EKG yourself, and weighed admission. The note said: "chest pain, troponin negative, discharge."

You cannot know how often that happens from a headline or a chart checker's flags. Compare the codes reported for a sample of encounters with the levels your billing team finds supported by the complete records.

The method below turns that review into a wRVU estimate. It uses public Medicare work RVU values and states every assumption. Use your own verified level differences, shifts, and group rate; the result is an estimate for that sample and compensation model.

Honest framing first

The work RVU (wRVU) values here are the Medicare figures for the ED evaluation and management (E/M) codes, unchanged since the 2023 guidelines took effect and published on the CMS fee schedule. Your group's compensation per wRVU is a contract term, not a Medicare number, so the dollar conversions are illustrations to replace with your own rate. Nothing here estimates how many charts you personally downcode.

For the scoring rules, use the high-complexity MDM guide. The 99284-versus-99285 comparison walks through a borderline chart. This article starts once you have a reviewed difference to count.

The unit economics of one level

The ED E/M codes carry these Medicare work RVUs: 99283 is 1.60, 99284 is 2.74, and 99285 is 4.00. Critical care (99291, the first 30 to 74 minutes) is 4.50.

A 99284 documented as a 99283 loses 1.14 wRVU. A 99285 documented as a 99284 loses 1.26. Critical care documented as a 99285 loses 0.50, and documented as a 99284 loses 1.76.

Recent Medicare fee schedules have published conversion factors around $32 to $33 per RVU. Many groups use a different dollars-per-wRVU rate, so use yours. At an illustrative $32 compensation rate, the 99284-to-99285 gap is about $40 per chart; at a group rate of $50 per wRVU, it is $63. These are compensation illustrations, not estimates of total Medicare reimbursement.

Scenario math you can plug into

The formula:

Lost wRVUs per year = downcodes per shift x shifts per year x wRVU gap per downcode

Here is one worked scenario: one downcode per shift, 180 shifts a year, and a 1.26 gap (every downcode is a 99285 scored as a 99284).

1 x 180 x 1.26 = 226.8 wRVUs per year

At an illustrative $32 compensation rate per wRVU, that is about $7,260. At a group rate of $50, it is about $11,340. Change an input and the result changes in proportion. Two downcodes per shift double it. A mix of 4-to-3 and 5-to-4 downcodes lowers the average gap to somewhere between 1.14 and 1.26. Fewer shifts lowers it. You can find each input except the first one.

Build the audit before projecting a year

Review 20 or 30 recent coded encounters with your billing team using the AMA MDM table. For each, record:

• The E/M code actually reported.
• The level supported by the complete record after coding review.
• The reason for any difference, including work performed but incompletely documented.
• The applicable wRVU difference and your compensation rate.

Keep confirmed differences separate from possible omissions that need clarification. A missing sentence or an AI flag is not proof that a higher code was earned. Estimate differences per shift from the reviewed sample, label the sample size and case mix, and use that rate in the formula. A small sample can identify a pattern; it does not establish a stable annual loss.

Where the losses cluster

Start with reviewed E/M level differences. Track critical care, procedures, and interpretations separately so the estimate does not count the same work twice.

Critical care is another. A sepsis, GI bleed, or respiratory failure case may meet the 30-minute threshold while the claim still shows no 99291 because the time statement and supporting detail were never written. The ACEP critical care FAQ explains what counts.

Procedures are another source. A laceration repair may code as simple when the note does not say "layered closure." A reduction may lack the pre- and post-reduction neurovascular exam. An incision and drainage may omit loculations. The procedure was done, but the note supports less of it.

Interpretations are a fourth. The EKG read at the bedside, the X-ray reviewed before radiology, and the point-of-care ultrasound without documented views or image retention each need a line in the note.

The compounding effects

A downcoded chart affects more than that chart's wRVUs.

Group metrics are built from the same claims. wRVUs per hour reflect documentation as well as the clinical work. Two clinicians with similar patients and care can post different numbers when one writes "decision regarding hospitalization" and the other writes only "discharge home."

Staffing decisions and compensation reviews use those metrics. A department that looks lower acuity on paper can be staffed and paid like a lower-acuity department. The effect can reach from an individual note to a contract negotiation.

The time trade

The practical objection is time. If capturing the level takes five more minutes per chart, that is two hours across a 24-patient shift.

The supporting sentences are short: the differential named, the external record cited, the interpretation written out, and the disposition written as a decision. On an open chart, they may take a minute or two. Against a $40 to $63 gap per chart, that can be a reasonable trade, but it is still time you may not have at 2 AM with a full board.

That is where a drafting tool can help. When the MDM already includes the differential, risk, data reviewed, and disposition reasoning, the remaining work is to read and correct the draft. The clinician still decides what belongs in the chart, but the documentation burden is lower.

Keep the ethical line clear

Upcoding means claiming work that did not happen: a differential you did not consider, records you did not open, or a deliberation you did not have.

The appropriate fix is to write down the medicine you already practiced so the note supports the level the visit earned. The ACEP E/M FAQ explains that the level follows the problems addressed, data analyzed, and risk managed. If you evaluated a threat to life, the chart can show it. If you did not, it should not.

Document the medicine you already practiced

The work happens at the bedside. Whether the claim reflects it depends on whether the note records it while the shift is still moving.

palmER was built by emergency medicine physicians for that part of the shift. Paste the available chart information into the MDM Assistant and it drafts the differential, data reviewed, and disposition reasoning. Chart CheckER reads the finished note before you sign and flags the unclaimed element, missing time statement, or procedure detail that could lower the code. You review the draft and keep the work you already did visible in the chart.

Start your free 30-day trial of palmER and use your next 20 coded encounters to measure the difference. No credit card required.