Loading and Proving a Contracted Fee Schedule

An underpayment is invisible from inside the transaction. The claim was accepted, an amount was allowed, an adjustment was taken, the account closed — nothing in that sequence reports that the allowed amount was lower than the contract required. The only thing that can reveal it is a number from outside: what the contract said the amount should have been.

Validating Contractual Adjustments covers running that comparison at the moment of posting, and it states the dependency plainly: no loaded contract, no validation. This article is that dependency — producing the number, and proving it is right before anything is decided on it.

Compensation terms vary by agreement and this site does not describe what any contract contains. What can be described is the structural categories they fall into, because each one presents a different loading problem.

A rate stated per code

The agreement or its exhibit lists amounts against codes. Easiest to load and the most work to maintain, because every code the practice bills has to be present, and one that is not is silently unpriced — the case that produces a variance report full of blanks rather than findings.

A rate expressed against a published schedule

The agreement prices by reference to an external, published fee schedule. The loading problem here is not arithmetic; it is version, and it is the subject of the next section.

A rate produced by a methodology

The amount is derived from something other than a per-code price — a bundled or episode-based arrangement, a per-day arrangement, a percentage of the charge. The loading problem is that the expected amount is not a lookup at all: it is a calculation over a set of lines, so the comparison has to happen at the same level the payment does.

This is the most common source of a disagreement in which nobody is doing anything wrong. Pricing against a public schedule feels like pricing against a fixed reference, and it is not — the reference is itself a computation with three inputs.

Under 42 U.S.C. § 1395w-4(b)(1) (opens in a new tab), the physician fee schedule amount for a service is the product of three things: the relative value for the service, the conversion factor for the year, and the geographic adjustment factor for the fee schedule area. Relative values themselves have three components, and the geographic factor varies by locality. So the published amount for one code is a different number in two localities, and a different number in two years.

• Which year? A schedule that updates annually means a contract priced against it either moves with it or is fixed to a stated year. Which of those is true is a contract term, and it is one of the more consequential ones.
• Which locality? The geographic adjustment is by fee schedule area. A practice with sites in two areas can be correctly paid two different amounts for the same code under one agreement.
• Which schedule? Physician services, clinical laboratory, durable medical equipment and ambulance each have their own published schedule. A single reference to “the fee schedule” does not necessarily resolve to one of them for every line a practice bills.

A load that answers “what should this have paid?” for a claim from eighteen months ago is a different artifact from a list of current rates. The difference is a handful of fields, and each one exists because of a specific way the simpler version fails.

1. 01

   The code, and any modifier that changes the price

   Some modifiers change what is payable rather than only what is reported — a component split, an assistant, a bilateral service. A schedule keyed on code alone will report a variance every time one of those appears, and the variance will be the schedule's fault.
2. 02

   An effective date range, not a current value

   Rates change, and a claim is priced by the terms in force on its date of service. A loaded schedule that holds only today's rate cannot validate anything but today's claims — and a variance review, by its nature, looks at older ones.
3. 03

   The place or locality, where the contract varies by it

   If the agreement prices against a geographically adjusted schedule, or sets different terms by site, then location is part of the key. Loading one rate for a multi-site practice quietly averages away the differences.
4. 04

   The product or plan, where the agreement covers more than one

   One signature can cover several product lines on different terms. A schedule that does not distinguish them produces variances that are really the wrong row being compared.
5. 05

   A pointer back to the term it came from

   Which clause or exhibit this row was derived from. Without it, a disputed rate becomes an argument about somebody's data entry rather than about the contract, and nobody can reconstruct the reasoning a year later.

A loaded schedule is a claim about a contract, and like any claim it can be wrong. Trusting an unvalidated load is worse than having none, because it converts a known gap into confident, wrong findings — and a variance report that produces false positives is abandoned within a month.

1. Recompute a sample by hand from the contract. Take a handful of the codes the practice bills most, work out the expected amount directly from the compensation terms, and compare that to what the load says. This catches the systematic errors — wrong version, wrong locality, wrong methodology — which are the ones that matter.
2. Compare the load against amounts actually allowed on clean claims. Not to set the rate from what was paid, which would be circular, but to see where the two disagree. Agreement across a broad sample is evidence the load is right; a cluster of disagreement is a lead.
3. Test the boundary dates. Price a claim from just before and just after a rate change and confirm each resolves to the right row. Effective-date errors are invisible in a spot check and systematic in a review.

A schedule is loaded once and wrong from then on unless something keeps it current. Two kinds of change do it: the contract's own amendment machinery, and — where the agreement prices against a published schedule that updates on its own cycle — a change nobody sent notice of at all, because nobody had to.

Both are worth a standing check rather than a discovery. The amendment side has its own rules about what notice is owed and when a change may take effect — when a payer contract changes covers them, and the response window that is the real deadline. The published-schedule side has no notice at all: the schedule updates, and a contract that follows it has moved. A practice that reloads only when it is told to will be a version behind for as long as it takes someone to notice a pattern in the variances.