Try Stream free for 30 days. $149/month. Annual billing at $99/month. 30-day free trial, no credit card required. Try it on your next visit →
How It Works
Product
All Features Huddle Stream Inlet
Specialties
Primary Care Direct Primary Care Pediatrics SNF & Long-Term Care About Blog Pricing FAQ Log In Start free trial →
Revenue Tool

What is conservative coding costing you?

Undercoding is rarely carelessness. It is what happens when a visit was genuinely complex, the documentation could not show it in a defensible shape, and the safe code was the lower one. This estimates the fee-for-service revenue that pattern leaves behind in a year — using published Medicare rates and your own numbers.

Rates: Medicare Physician Fee Schedule, January 2026 · national, non-facility

Physicians, NPs and PAs who bill under their own visits.
MGMA puts primary care around 18–25 patients a day, or roughly 3,500–4,500 a year.
3,500
Your assumption, not ours. The default is deliberately conservative — see the note below on why we don't use the published research here.
8%
As a share of all visits. G2211 is a Medicare code, so start from your Medicare mix and subtract whatever you already bill — if you claim it routinely, set this to zero.
35%
Estimated annual revenue not captured
$0
—
E/M level, one step down $0
—

The step from 99213 ($95.19) to 99214 ($135.61) is $40.42. Commercial plans generally pay more than Medicare, so this is a floor rather than a midpoint.

G2211 not billed $0
—

G2211 is the visit-complexity add-on for a continuing care relationship — $16.40 on top of the base visit. In its CY2026 rule CMS noted that actual reporting came in below its own utilization estimate, which means many practices absorbed the budget-neutrality reduction that funded this code and then did not bill it. From January 2026 it also applies to home and residence visits.

This is not an argument for upcoding

It is an argument for documentation that can defend the level you already earned. Billing a level you did not perform is fraud, and every clinician is right to be cautious about it. The gap worth closing is between the complexity of the care delivered and the complexity the record can demonstrate — nothing beyond that.

What this deliberately leaves out

The estimate is built to be conservative, so several real sources of uncaptured revenue are not counted:
  • Separately-identifiable problems addressed during a preventive visit, billable with modifier 25 — often another $95–$136 each
  • The larger step from 99214 to 99215 ($46.47), counted here at the smaller level-3-to-4 rate
  • Commercial reimbursement above Medicare rates
  • Chronic care management, transitional care and other time-based codes
  • Denials and rework caused by documentation that could not support the claim

Send me the breakdown

Leave your email and we'll send you The Defensible Visit — our guide to documenting medical decision-making so it survives review. You'll get the link straight away. No call unless you ask for one.

Your inputs are sent along so the breakdown matches what you see here. We don't share your details.

Here it is

The Defensible Visit — how to document medical decision-making so it survives review. The 2-of-3 MDM rule, what actually counts as data, and a one-page checklist.

Read the guide

We've got your numbers too, and we'll be in touch if there's anything useful to add. No call unless you ask for one.

Sources

  • Payment amounts: Medicare Physician Fee Schedule, CY2026, national non-facility. 99213 $95.19 · 99214 $135.61 · 99215 $182.08 · G2211 $16.40. CMS Physician Fee Schedule · CY2026 final rule
  • Visit volume: MGMA provider productivity benchmarks
  • Modifier 25 and preventive visits: AAFP
  • On the undercoding rate: published prevalence studies report figures far higher than our default — 44% of pediatric visits in one review, 33% in a family-physician survey. We do not build on them, because almost all predate the 2021 E/M overhaul, after which level selection changed to medical decision-making or time and reported coding shifted upward materially. Those studies do not describe how anyone codes in 2026. The rate above is your assumption, and the default is set low on purpose.