How do the 2023 E/M rules change ED level selection for 99281 to 99285?

What actually changed for 99281 to 99285 on 1 January 2023?

History and exam stopped deciding the level. Since 1 January 2023 the emergency department codes are selected on medical decision making alone, and the descriptors say so: each of 99282 through 99285 now requires "a medically appropriate history and/or examination" plus a named level of medical decision making. The history and the exam are still performed and still documented. They no longer carry the code.

That is the whole change, and it is why a cheat sheet printed in 2022 now costs money in both directions. The old sheet capped a level when the exam was thin. The new rules do not.

What the 99281 to 99285 descriptors now say

99281 is the outlier. It reads as an ED visit "that may not require the presence of a physician or other qualified health care professional," and it carries no medical decision making level at all. The other four map straight onto the medical decision making grid: 99282 is straightforward, 99283 is low, 99284 is moderate, 99285 is high. Five codes, four levels of decision making, one code that sits outside the grid.

Does time count for an ED visit?

No, and this is the most commonly published error on the topic. ED codes have no typical times and no time based selection path, because emergency services are provided on a variable intensity basis. The American College of Emergency Physicians pressed the Joint CPT and RUC Workgroup on the point and won it. Every other office and outpatient E/M family got a total time option in 2021 and 2023. The ED family did not. If a page says time spent supports a 99285, it is wrong, and several pages ranking for these codes say it.

Which emergency room coding guidelines still apply after 2023?

Place of service 23 still belongs on the professional claim, and a mismatch is a live payer edit rather than a coding preference. Modifier 25 still applies when a separately identifiable visit accompanies a procedure, on the same test that decides a same day wellness and problem visit. What no longer applies is anything counting history bullets or exam systems toward a level, including the 1995 and 1997 documentation guidelines that some payer policies still cross reference under a 2026 review date.

How do you score the three MDM elements on an ED chart?

Three elements, and the visit level is scored on the highest two of the three. Meeting one element at high does not produce a 99285 on its own, and this is the rule that decides most disputed charts.

The three are the number and complexity of problems addressed, the amount and complexity of data reviewed and analyzed, and the risk of complications, morbidity or mortality from patient management. Problems and risk are where emergency medicine usually scores. Data is where charts quietly lose a level, because the work was done and never written down in a way a coder can count. That is ordinary coding work against the current grid, not a judgment nobody can check.

Category 1, 2 and 3 data, and what an ED workup actually counts as

Category 1 covers tests, documents, orders and independent historians: each unique test ordered, each unique test result reviewed, each prior external note from each unique source, and an assessment requiring an independent historian. Category 2 is independent interpretation of a test where an interpretation or report is customary and where you are not separately reporting a CPT code for that interpretation. Category 3 is discussion of management or test interpretation with an external physician or other professional involved in the patient's care, and it has a hard requirement: the exchange must be interactive, direct, and not routed through a nonclinical intermediary.

Two consequences follow. A call to the on call cardiologist counts under Category 3 only if the two of you actually spoke, so a message left with a service does not. And an independent EKG read counts under Category 2 only while you are not billing separately for it: the same work cannot be paid twice and counted twice.

Where is the line between 99284 and 99285?

On the problems element and the risk element, not on how sick the patient turned out to be. This is the single most disputed decision in emergency coding and it is the one no ranking page works through.

What "threat to life or bodily function" actually means

High complexity on the problems element needs one acute or chronic illness or injury that poses a threat to life or bodily function, or one or more chronic illnesses with severe exacerbation, progression or side effects of treatment. The phrase that does the work is "poses a threat," present tense, at the time of the evaluation. Guidance is explicit that presenting symptoms likely to represent a highly morbid condition may drive medical decision making even when the final diagnosis is not highly morbid, provided the evaluation and treatment were consistent with the likely nature of the condition.

So chest pain worked up for acute coronary syndrome and discharged with reflux is scored on the threat that was being excluded, not on the reflux. The chart has to show the differential being worked, which is a documentation habit rather than a coding one.

Does a decision to admit make it high risk?

A decision regarding hospitalization or escalation of hospital level of care sits in the high risk column, alongside drug therapy requiring intensive monitoring for toxicity and a decision regarding emergency major surgery. So yes, the admission decision reaches high on the risk element. It does not finish the job, because the level runs on the highest two of three. High risk plus moderate problems plus moderate data is a moderate visit. High risk plus high problems is a 99285.

Can you still bill 99285 when the history and exam are thin?

Yes, and this is the most useful practical consequence of the change. An intubated patient who cannot give a history, a trauma arrival with a two line exam and a resuscitation, an altered patient with an independent historian: under the old rules the missing comprehensive history capped the code. Now the decision making stands on its own.

The record still carries a medically appropriate history and exam, descriptive enough that a coder or auditor can see the complexity addressed. The difference is that they now explain the decision making rather than score it. The same logic runs through the G2211 visit complexity add on, where documentation has to show the thinking rather than fill a template.

What is each ED level worth in 2026, and what does one level cost you?

At the 2026 conversion factor of $33.4009, the professional component of an ED visit runs from about $11 at 99281 to about $171 at 99285, before geographic adjustment.

Those are the national unadjusted amounts: about $11 for 99281, $40 for 99282, $69 for 99283, $118 for 99284 and $171 for 99285. The gaps between adjacent levels are what matter operationally: $29 from 99281 to 99282, $29 from 99282 to 99283, $49 from 99283 to 99284, and $53 from 99284 to 99285.

99283, 99284 and 99285 cpt code payment, and the RVUs behind it

The work relative values run 0.25, 0.93, 1.60, 2.74 and 4.00 across the five levels, with total facility values of 0.33, 1.21, 2.08, 3.54 and 5.13.

Work relative value units climb from 0.25 at 99281 to 4.00 at 99285, and total relative value units from 0.33 to 5.13 across the same range. Note how much steeper the top of the range is than the bottom: the step from 99284 to 99285 is worth more than the entire value of a 99282.

A worked example: nine physicians, 28,000 visits a year

Take a nine physician emergency group billing the professional side of 28,000 ED visits a year, and assume 35 percent of those encounters, 9,800 visits, are documented at high medical decision making. Every one of those reported as 99284 instead of 99285 gives up $53.

At 5 percent of that volume, 490 visits, the group loses $25,970 a year. At 10 percent it loses $51,940. None of it reaches a denial report, because a 99284 pays cleanly and reconciles against a claim nobody questions, which is why it survives year after year. The same $53 runs the other way on a chart that does not support high decision making, and that direction carries audit exposure instead of lost revenue. To run this on your own ED lines rather than a modeled profile, a billing review uses your remits.

Why is the hospital's ED level different from the physician's level on the same visit?

Because two different systems produce them, and both can be right. The physician claim uses the medical decision making grid above. The facility claim uses the hospital's own acuity and intervention point system, counting nursing interventions, monitoring, medications and resources consumed. There is no national facility leveling standard, so two hospitals can legitimately level the same encounter differently.

This is the root cause of most patient billing disputes about ED levels, and no page ranking for these codes explains it. A patient who sees a level 4 physician charge and a level 5 facility charge on the same visit has not been double billed and has not necessarily been upcoded. The split between the two claims works the same way it does on a surgical case, which we cover in the facility and professional fee split.

What goes wrong most often on ED E/M claims?

The expensive failures on ED claims are not level selection errors at all. They are charges never created and denials never worked.

Across the claims we audit, eligibility and coverage errors cause 24% of denials, coding and modifier errors cause 21% of denials, missing or invalid prior authorization causes 17%, duplicate claims 9% and timely filing 6%. Only the second of those is a coding problem. The first is a front end problem, and in an emergency department, where insurance is often captured after the patient is seen, it is the hardest one to fix and the most worth fixing. Checking coverage properly is what eligibility and prior authorization work is for, even in a setting that cannot turn anyone away.

Two more sit underneath the denial numbers. In our billing reviews, 8% of ED and on-call consults performed were never captured as charges across five surgical groups, which is revenue that never becomes a claim. And 19% of denied claims were never reworked or appealed, which is revenue that became a claim and then got abandoned. Our appeals are overturned 68% of the time, so the abandoned ones are not lost causes, they are unworked ones. A named queue that actually gets worked is what denials and AR recovery means in practice.

Payer side downcoding belongs here too. Severity algorithms that downgrade ED levels against the final diagnosis are a live issue for emergency groups, and the answer is the rule above: the level follows the presenting problem worked, not the discharge diagnosis. That appeal is winnable when the chart shows the differential.

Should ED E/M coding sit in house or with a billing partner?

In house wins when one named person owns the medical decision making grid, audits a sample of charts monthly against it, and reads the level distribution by physician rather than in aggregate. That is a small, specific job. The problem is that it is usually nobody's job: 42% of practice managers told us nobody owns denial follow up full time, and 63% could not name their top three denial reasons.

Fully loaded in house billing cost 7.9% of collections for practices under $2M in our reviews, against an outsourced range of 3% to 6%. The spread is not what decides it. Volume and turnover are: ED coding is high volume, rule dense, and punishing when the one person who understands it leaves. Across our client practices, first pass denial rate fell from 14.2% to 6.1% within 90 days, clean claim rate rose from 89.6% to 97.3%, and median days in AR dropped from 54 to 33.

If you are comparing partners, ask one question before price: will they show you level distribution by physician, monthly, against the medical decision making grid. Most will not. The rest of that list is on our medical billing companies page, and the wider revenue cycle view matters more than the ED codes alone.

How does each ED level map to medical decision making?

One table settles most of it: which level, which decision making, which elements, and what the professional side pays in 2026.

Emergency medicine and freestanding emergency departments

This is the home of the question, and the place where the 2023 change is worth the most. Freestanding emergency departments bill the same 99281 to 99285 range under the same rules, with place of service 23 on the professional claim. The operational move is the same in both settings: pull level distribution by physician each month and look at the spread rather than the average. A physician sitting three points below the group on 99285 is usually documenting the differential poorly, not seeing easier patients. That is the work behind emergency room billing.

Urgent care

Urgent care does not use this code family at all. Visits are billed on the office and outpatient range 99202 to 99215, which does carry a total time selection option, and many contracts add S9083 as a global per visit rate that flattens leveling entirely. Coders who move between the two settings routinely carry the time option across into the ED range, where it does not exist. Keeping the two grids physically separate at the coder desk prevents most of it, and urgent care billing runs on its own rules.

Behavioral health in the emergency department

Psychiatric boarding is where ED leveling and reality diverge most sharply. A patient waiting fourteen hours for an inpatient bed generates enormous nursing and monitoring work, all of which lands on the facility claim and none of which lifts the physician level, because time is not a factor. The physician level turns on the risk element, where a decision regarding hospitalization or escalation of care reaches high. Documenting the disposition decision and the safety assessment is what carries the code.

Ambulance and EMS

Transport is a separate claim under the ambulance fee schedule and sits entirely outside the ED physician level. What crosses over is the record: the run sheet is an external note from a unique source and counts under Category 1 data when the emergency physician reviews it, and the crew can serve as an independent historian for a patient who cannot give a history. Both are routinely performed and almost never documented in a way a coder can score.

Primary care and internal medicine

The referring physician who sends a patient to the emergency department, and the one who sees them afterwards, bill the office and outpatient range on their own merits. Neither is inside anything the ED billed. The connection that matters is Category 3: an interactive discussion between the emergency physician and the patient's own physician counts toward the data element for the emergency physician, provided they actually spoke. Primary care practices should expect those calls to be documented on both sides.

Dental emergencies in the emergency department

Non-traumatic dental presentations are a high volume, low yield slice of emergency work and they level low for a reason: a dental abscess without systemic involvement is usually one acute uncomplicated illness with low risk, which is a 99283. The level rises when there is facial cellulitis, airway involvement or systemic infection, which moves the problems element toward a threat to bodily function. Coding these at 99284 by reflex because the patient was in severe pain is a pattern payers audit.