The conversation about NP reimbursement tends to focus on the structural — the 85% Medicare rate, payer contract terms, the policy landscape. These matter. But in our experience working with NP-owned practices across the country, a significant portion of the revenue gap isn't structural at all. It's recoverable. And it starts with how complexity is being documented and coded.

Undercoding — billing a lower-complexity visit code than the clinical encounter actually supports — is endemic in independent primary care. It is not primarily a knowledge problem. Most experienced NPs understand E&M coding. It is a workflow problem: documentation templates that don't prompt for the right components, coding habits that default to familiar codes, and a billing review process that catches denials but not silent underpayment.

What the Complexity Data Shows

Across E&M visits in Duet's network over the past twelve months, the distribution of visit complexity scores tells a specific story.

Just over half of all E&M visits — 50.9% — fall at complexity score 4, the level that typically maps to a 99214. Another 20.4% reach score 5, the highest complexity tier. That means roughly 71% of E&M visits in our network involve moderate-to-high complexity by the internal scoring model.

The CPT code distribution doesn't fully reflect that. The 99214 accounts for 38.5% of visits. The 99215 — the appropriate code for the highest-complexity tier — accounts for 7.8%.

There is a gap between what the clinical complexity of these visits supports and what is being billed. Some of that gap is legitimate: not every score-5 visit maps cleanly to a 99215, and documentation quality varies. But not all of it is legitimate, and the portion that isn't represents recoverable revenue that is already being earned in the exam room.

The Add-On Code Picture

The average number of add-on codes billed per E&M visit across our network is 0.23 — meaning the large majority of visits capture no add-ons at all.

This is where some of the clearest money is being left on the table.

At complexity score 5 — the highest tier, where clinical complexity is most evident — practices average 0.69 add-ons per visit and capture G2211 on 56.5% of eligible Medicare visits. At score 3, add-ons drop to 0.17 per visit and G2211 capture falls to 35.2%.

The behavioral assessment add-on shows the same pattern: captured on 22.8% of score-5 visits, 8.1% of score-4 visits, and 4.3% of score-3 visits. SDOH screening, smoking cessation, and other preventive add-ons show capture rates under 1% across most complexity tiers — despite the fact that these screenings are routinely performed in primary care visits.

These codes exist because CMS and commercial payers have explicitly created reimbursement pathways for work that primary care providers are already doing. The screenings are happening. The assessments are being conducted. The documentation is not consistently capturing them in billable terms.

Why This Happens

The root cause is rarely negligence or ignorance. It is that most EHR documentation templates are not structured to prompt for add-on code capture systematically. A clinician completing a visit note is thinking about the patient, not about whether the social history section will support an SDOH screening code. Without a template that surfaces these prompts at the right moment in the documentation workflow, add-on codes get missed — not occasionally, but consistently, across every visit.

The same dynamic applies to Medical Decision Making components for the primary E&M code. The three elements that determine MDM level — number and complexity of problems, amount and complexity of data reviewed, and risk level of the management plan — need to be explicitly present in the documentation to support a higher-complexity code. When templates are built around narrative rather than structured MDM components, visits that justify a 99215 get documented and billed as 99214s.

The fix is not retraining clinicians on coding rules they already know. It is rebuilding the documentation workflow so that accurate complexity capture is the path of least resistance, not an additional cognitive task at the end of a full clinical day.

What a Coding Review Actually Finds

When we conduct coding reviews for practices joining Duet's network, the most common findings are:

Systematic undercoding on established patient visits. Practices where 99213 is the modal code for established patients, despite a panel that includes significant chronic disease burden, are almost always undercoding. The complexity is there. The documentation habit isn't.

Near-zero add-on capture despite routine performance. Practices that conduct behavioral assessments, SDOH screenings, and smoking cessation counseling as standard protocol — but capture these as add-on codes rarely or never — are performing unreimbursed work at scale.

G2211 applied inconsistently rather than systematically. The network average G2211 capture rate is 41.8%. Practices at the high end of the range have made G2211 documentation a standard part of their Medicare visit workflow. Practices at the low end are applying it ad hoc, which means it gets captured when someone remembers and missed when they don't.

Time-based coding underutilized. When total visit time — including pre-visit chart review and post-visit documentation — supports a higher E&M level than MDM alone would, time-based coding can justify the higher code. Most practices default to MDM-based coding for every visit without evaluating whether time-based coding would better represent the work performed.

The Revenue Implication

The average allowed difference between a 99214 and a 99215 in our network is approximately $49 per visit. For a practice billing 3,000 E&M visits annually, shifting 10% of those visits from 99214 to accurately coded 99215 — where the clinical complexity already supports it — represents roughly $14,700 in additional annual revenue. Add consistent G2211 capture on eligible Medicare visits at roughly $16 per applicable claim, and the recoverable revenue compounds further.

These are not optimistic projections. They are what we observe when practices implement structured documentation workflows and systematic coding review. The care is already being delivered. The question is whether the documentation reflects it.

Talk to a Duet advisor about what a coding review would find in your practice →