The 2021 E/M guidelines shifted coding from counting exam bullets to medical decision-making complexity. Most practices never updated their coding habits. The result: providers default to 99213 when their documentation supports 99214. The difference is $30-$40 per visit. A single provider seeing 20 patients/day who undercodes by one level on just 40% of visits loses $57,600-$76,800 per year. This isn't upcoding. It's billing for the work you're already doing and documenting.
If you manage billing for a family practice, internal medicine group, or geriatric practice, pull your E/M distribution report right now. Look at the ratio of 99213 to 99214 visits. If more than 50% of your established patient visits are billed as 99213, you're almost certainly undercoding.
The national average for internal medicine is approximately 45% 99214, 35% 99213, and 15% 99215. For geriatrics, 99214 should dominate because elderly patients with multiple chronic conditions inherently involve moderate-to-high complexity medical decision-making. If your geriatric practice bills 60% at 99213, the coding doesn't match the patient population.
This isn't about gaming the system. It's about billing accurately for the clinical work your providers are already performing and documenting.
Before 2021, E/M coding was based on three components: history, exam, and medical decision-making (MDM). You needed to document a specific number of history elements and exam bullets to qualify for each level. This led to "bullet counting" where providers structured notes around documentation requirements rather than clinical relevance.
The 2021 guidelines simplified this. For established patient visits (99211-99215), E/M level is now determined by either medical decision-making complexity or total time spent. The history and exam requirements were eliminated as mandatory elements for level selection.
This change should have resulted in practices billing more 99214 and 99215 visits because the documentation threshold shifted from counting exam bullets to capturing the complexity of what the provider actually did. A provider managing 4 chronic conditions with medication adjustments and care coordination is performing 99214-level work regardless of whether they documented 8 exam bullets.
What actually happened: most practices kept coding the same way they did before 2021. Old habits persisted. The default stayed at 99213.
| Element | 99213 (Low Complexity) | 99214 (Moderate Complexity) |
|---|---|---|
| Number of problems addressed | 2 or more self-limited problems | 1+ chronic illness with mild exacerbation, or 2+ stable chronic conditions, or 1 undiagnosed new problem with uncertain prognosis |
| Data reviewed | Minimal data (ordering a lab, reviewing one source) | Moderate data (ordering and reviewing tests, obtaining records from external source, independent interpretation of results, discussion with external provider) |
| Risk of complications | Low risk (OTC drugs, minor surgery with no risk factors) | Moderate risk (prescription drug management, decisions about minor surgery with risk factors, decisions about elective major surgery) |
| Time-based alternative | 20-29 minutes total time | 30-39 minutes total time |
| Medicare reimbursement | ~$97-$110 | ~$133-$150 |
The key question for every established patient visit: did the provider manage a chronic condition, prescribe a medication, or review test results? If yes, the visit almost certainly meets 99214 criteria under the current guidelines. A diabetic patient with a medication adjustment is 99214. A hypertensive patient whose BP medication is being titrated is 99214. A geriatric patient with 3 stable chronic conditions being managed simultaneously is 99214.
99213 is appropriate for genuinely simple visits: a follow-up for a resolved acute problem, a quick recheck with no medication changes, a patient with a single self-limited condition. If the provider spent time on medication management, chronic disease monitoring, or clinical decision-making beyond the obvious, it's 99214.
The #1 reason providers default to 99213. They've heard stories of audits, downcoding, and recoupment, and they'd rather leave money on the table than risk a payer audit. The irony: undercoding doesn't protect you from audits. Payers audit in both directions. A practice that bills 80% at 99213 when their patient population suggests 50% can trigger an audit for aberrant billing patterns just as easily as a practice billing 80% at 99215.
The reality: Auditors look at whether the documentation supports the level billed, not whether you billed high or low. If your note documents moderate-complexity MDM and you billed 99213, the auditor would actually upcode it to 99214. Undercoding isn't conservative billing. It's inaccurate billing that happens to cost you money instead of the payer.
Providers who learned to code before 2021 were trained on the bullet-counting method. "I only documented 6 exam elements, so it's a 99213." That logic no longer applies. E/M level is now driven by MDM complexity, and most established patient visits involve moderate-complexity MDM (99214) when you assess the number of problems addressed, data reviewed, and risk of management.
Some EHR systems have coding suggestion algorithms that default to 99213 unless specific documentation triggers are met. These algorithms may not have been updated for the 2021 guidelines, or they may weight history and exam elements that are no longer required for level selection. If your EHR suggests 99213 and the provider doesn't override it, the code goes out at the lower level.
If your billing team was trained on the old guidelines and hasn't been retrained on MDM-based coding, they may be downcoding claims that providers submitted at the correct level. A biller who looks at a note and thinks "there aren't enough exam bullets for a 99214" is applying outdated criteria. Exam bullets no longer determine E/M level for established patients.
The reimbursement gap between 99213 and 99214 varies by payer but averages $30-$40 per visit.
Provider: Internal medicine physician, 20 established patients/day, 240 working days/year
Current coding distribution: 65% at 99213, 30% at 99214, 5% at 99215
Corrected distribution (based on MDM guidelines): 30% at 99213, 55% at 99214, 15% at 99215
Visits shifted from 99213 to 99214: 35% of 4,800 annual visits = 1,680 visits
Revenue per shifted visit: $35 average difference
Annual revenue recovered: 1,680 x $35 = $58,800 per provider per year
For a 3-provider practice, that's $176,400 per year in revenue that was earned, documented, and never billed. Not new revenue. Not upcoding. Just accurate coding for work that's already being performed.
The compound effect: Undercoding doesn't just cost you the visit revenue. It suppresses your clean claim rate analytics because your per-visit revenue looks lower than it should, making it harder to identify when payers are actually underpaying. It also depresses your practice valuation if you ever sell, because revenue-based multiples are applied to your reported collections, not your potential collections.
Step 1: Pull your E/M distribution report. Most PM systems can generate this. You want: the count of each E/M code (99211-99215) billed over the last 6 months, broken down by provider.
Step 2: Compare to specialty benchmarks.
| Specialty | Expected 99213 % | Expected 99214 % | Expected 99215 % |
|---|---|---|---|
| Family Practice | 30-40% | 40-50% | 8-12% |
| Internal Medicine | 25-35% | 45-55% | 12-18% |
| Geriatrics | 15-25% | 45-55% | 18-25% |
| Mental Health | 20-30% | 40-50% | 15-25% |
| Cardiology | 20-30% | 45-55% | 15-22% |
If your 99213 percentage is 15+ points above the expected range for your specialty, you're undercoding.
Step 3: Pull 10 random 99213 charts and re-evaluate. Take 10 visits that were billed as 99213 in the last month. Read the notes. For each one, ask: did the provider manage a chronic condition? Prescribe or adjust a medication? Review and interpret test results? Coordinate care with another provider? If the answer to any of these is yes, the visit likely supports 99214 under MDM-based coding.
Step 4: Calculate the revenue impact. Take the number of 99213 visits that should have been 99214 (from your 10-chart sample, extrapolate the percentage to your full volume). Multiply by the reimbursement difference ($30-$40). That's your annualized undercoding loss.
The single most effective intervention. A 30-minute training session showing providers how the 2021 guidelines define each MDM level, with real examples from their own charts, typically shifts coding distribution within 2 weeks. The key message: "If you're managing chronic conditions and prescribing medications, you're doing 99214-level work. Code it that way."
If your EHR has a coding assistant, verify it's using 2021 MDM criteria. Some systems still default to the old history/exam method. Contact your EHR vendor for the update or configure the suggestion algorithm to weight MDM elements over exam bullets.
Train your billing team (or your outsourced billing partner) to review E/M levels before submission. When a biller sees a note with 3 chronic conditions managed, 2 medications adjusted, and lab results reviewed, and the provider submitted it as 99213, the biller should flag it for upcoding review. This isn't changing the provider's code without authorization. It's a quality check that catches undercoding before revenue is lost.
Under the 2021 guidelines, providers can select E/M level based on total time spent on the encounter (including pre-visit chart review and post-visit documentation) instead of MDM. If total time was 30-39 minutes, it's 99214 regardless of MDM complexity. Many providers underestimate their total visit time because they only count face-to-face minutes. Chart review, care coordination, and documentation time all count under the current guidelines.
Time-based thresholds (established patients): 99212 = 10-19 min. 99213 = 20-29 min. 99214 = 30-39 min. 99215 = 40-54 min. If a provider spends 15 minutes face-to-face, 5 minutes reviewing the chart before the visit, and 12 minutes on documentation after, that's 32 minutes total. Time-based coding = 99214.
After retraining, pull the E/M distribution report monthly for each provider. The shift from 99213-heavy to a balanced distribution should be visible within 30 days. If it's not, the training didn't stick and needs reinforcement. If one provider shifts and another doesn't, the holdout may need individual coaching.
Practices worry that shifting toward more 99214 billing will trigger audits. The data says otherwise. Medicare's own utilization data shows that the national average for 99214 is approximately 45% of established patient visits. If your practice is billing 99214 at 45-55%, you're within the normal distribution. You're only at audit risk if you're an outlier, and a practice billing 70% at 99213 is more of an outlier than one billing 50% at 99214.
The compliance protection is documentation. Every 99214 visit should have a note that clearly documents at least two of the three MDM elements at the moderate level: number/complexity of problems addressed, amount/complexity of data reviewed, and risk of complications/morbidity. If the documentation supports the code, the code is correct. Period.
If you're unsure whether your documentation supports the shift, start with a 20-chart audit. Have a certified coder review 20 visits currently billed as 99213 and independently assess the appropriate level. If 60-70% come back as 99214, you have the data to justify the change and the documentation to support it.
We'll analyze your E/M coding distribution against specialty benchmarks and calculate the exact dollar amount you're leaving on the table. Most practices find $40K-$80K per provider per year in correctable undercoding.
Get a Free Coding Analysis →