The Quiet Drift

An independent orthopedic group with twenty providers runs roughly eighty thousand claims a year. A one percent underpayment rate across those claims, at an average claim value of three hundred dollars, is two hundred forty thousand dollars. A three percent rate is seven hundred and twenty thousand. These are not hypothetical numbers. They are the range the industry has been quietly tracking for more than a decade.

What most practice administrators do not fully internalize is this: the money is not being stolen, and it is not being denied. It is being paid, just not in full, and the gap between what was contractually owed and what the payer actually remitted is small enough on any given claim to avoid scrutiny and large enough in aggregate to fund a second associate.

The anatomy of a drift

Underpayment is a specific phenomenon, and it is useful to define it against what it is not. It is not claim denial, where a payer refuses the claim outright. It is not fraud, where a payer deliberately misrepresents a contract. It is the quiet, mechanical result of the gap between what a contract says a procedure should pay and what the remittance advice actually delivers.

The mechanism is drift. Contracts are negotiated on a multi-year cycle, fee schedules update annually or more often, modifier rules shift with CMS guidance, and payer-side interpretation of bundling rules varies by payer and sometimes by claim. Each of these moving parts is small. Their interaction is not.

• ContractFee schedule drift. A payer unilaterally applies a revised fee schedule without updating the contract language the practice is operating against.
• CodingModifier suppression. A valid modifier is stripped or reinterpreted, downcoding the payment without triggering a denial.
• BundlingSilent bundling. Two legitimately billable procedures are paid as one, on the basis of a bundling edit the payer may or may not be entitled to apply.
• UnitUnit miscounting. On destruction codes, time-based codes, and procedures billed per unit, the payer pays one unit when the contract calls for multiple.
• NetworkOut-of-network misapplication. In-network claims paid at out-of-network rates, usually due to provider identifier mismatch rather than coverage.

None of these shows up on an EOB as a red flag. Each appears as a line item that paid, just for less than expected. Pattern recognition across thousands of claims is what surfaces them. In-house billing teams, juggling eligibility verification, coding, submission, denial follow-up, and patient collections, almost never have the bandwidth for that pattern recognition work.

Where in-house billing hits its limit

This is not a staffing criticism. A single billing specialist managing two thousand claims a month, at an industry-standard recovery workload, can touch a meaningful percentage of denials and a small percentage of zero-paid claims. Catching a forty-dollar shortfall on one line item of a mixed remittance, against a contract in a PDF filing cabinet, is a different category of work. It requires a structured representation of every active contract, a parser that reads every 835 remittance, and a comparison engine that surfaces deltas faster than a human can triage them.

The economics of rework make this worse. MGMA estimates the average cost to appeal a denied claim at around one hundred eighteen dollars. For a sixty-dollar underpayment, in-house appeal is a net-negative activity. The only way the math works is when the detection, documentation, and dispute workflow is substantially automated.

What detection looks like at the CPT-code level

The simplest way to describe the work Vareo does is a three-way comparison. For every line item on every remittance, three values must agree:

The flagged claims are then grouped by payer, by pattern, and by dispute channel. Each category gets routed into the correct workflow: a corrected claim resubmission, a formal appeal with documentation, an authorization reopen, or, when it comes to that, a regulatory escalation. The voice AI layer handles the one category that stubbornly resists written workflow, which is phone-based payer follow-up, the place where recovery work historically goes to die.

The economics of recovery

The case for in-house recovery at a small practice almost never pencils out. The infrastructure cost (a contract management system, a dedicated underpayment analyst, a denial management platform, and the institutional knowledge to run all three) runs well into six figures a year before it returns a dollar.

The case for a vendor recovery partner only pencils out when two conditions hold. First, the fee structure must be pure contingency. Upfront or subscription fees transfer risk back to the practice and misalign incentives. Second, the vendor must own the workflow end to end, including the unglamorous phone work, because partial handoffs leave the practice holding the hardest part of the job.

Vareo is structured around those two conditions. Twenty-five percent of what we recover, zero percent of what we do not. If a full audit returns no actionable underpayments, the practice owes nothing and the engagement ends. There is no middle case.