Underpaid Contract Claims, a Six-Code Fee Schedule Load, and a Ninety-Day Appeal Window
A 40-provider multispecialty group holds a commercial contract paying 115% of the Medicare Physician Fee Schedule. A revenue-cycle analyst samples 60 paid claims across three months: 38 paid at 115% of Medicare and 22 paid at 100%. Average gap on the underpaid claims is $31. The group submits about 4,800 claims per month to this payer. The contract gives a 90-day appeal window from the remittance advice date and requires claims within 120 days of service. One clinic's average charge lag is 47 days. The payer portal shows the group's fee schedule loaded at 100% of Medicare for six of twelve taxonomy codes. The analyst proposes sending 1,700 individual appeals.
1. Quantify the opportunity and assess how much confidence the committee should place in the number.
Read the worked answer
Point estimate
- 22/60 = 36.7% of claims underpaid.
- 4,800 claims/month x 12 = 57,600 claims/year.
- 57,600 x 0.367 = about 21,100 underpaid claims.
- 21,100 x $31 = roughly $655,000 per year.
Range
- The 95% confidence interval on 36.7% with n=60 is roughly 26% to 43%, which annualizes to about $460,000 to $760,000. A sample of 60 is thin for a seven-figure number.
But do not extrapolate here - count.
- This is an electronic-data problem, not a sampling problem. If six taxonomy codes are loaded at 100%, the affected population is identifiable, and a 100% comparison of 835 paid amounts against the contract fee schedule on that population is more precise and probably cheaper than defending an extrapolation.
- The analyst's instinct that the exposure is large is probably right, and this sample likely understates it: a random 60 will include taxonomy codes that were loaded correctly, diluting the rate.
Recommendation: report the sample result as a screening finding, then produce a census figure for the six affected codes before anyone puts a number in a formal demand or a financial statement.
2. Before pursuing the underpayment, what must be confirmed about the contract and the root cause?
Read the worked answer
Contract interpretation first - the claim may be wrong, not the payer.
- What exactly does 115% of the Medicare fee schedule mean here? Is the base the MPFS allowable for the locality and the year of service, a fixed year named in the contract, or the amount Medicare actually pays (which is net of the 2% sequestration reduction)? A 2% definitional difference will generate hundreds of thousands of dollars of phantom underpayment.
- Are the affected CPT codes actually on the fee schedule, or are they priced by report, case rate, or carve-out? Unlisted and non-fee-schedule codes may legitimately pay differently.
- Does the contract cap, bundle, or apply multiple-procedure reduction rules that explain the 22 claims?
Root cause second.
- A fee schedule loaded at 100% for six taxonomy codes points to a configuration error on the payer side, not random adjudication. If so, the correct remedy is a mass adjustment, not 1,700 appeals.
- It could also be the group's own credentialing/contracting paperwork - if those six codes were added to the contract later, the payer may argue the rate applies only prospectively from the amendment date. Get the executed amendment.
Do not send appeals until both are answered. An appeal asserting a contract term the payer can rebut with the signature page wastes the window and weakens the later escalation.
3. Assuming the contract reading holds, how should the group sequence the recovery and fix the process?
Read the worked answer
Sequence by deadline and by leverage, not by claim count.
- Work backwards from the remittance dates. Nothing outside the 90-day window is recoverable regardless of merit. Triage: identify which underpaid claims are still live and which are already lost, and tell the committee the split honestly - that number is a governance fact, not an embarrassment.
- Do not file 1,700 individual appeals. Batch by taxonomy code and CPT with a single cover letter citing the contract section, the fee schedule exhibit, and a line-item spreadsheet. Where the payer has a provider-relations or network-management contact, escalate there first with the data - a systemic load error is usually corrected administratively, and the appeal volume drops to the tail the payer refuses.
- Pursue prospective correction in parallel. Every month the load stays wrong adds about $55,000. Prospective correction is worth more than the retroactive fight and is much easier to win.
- If the payer refuses, invoke the contract's dispute-resolution clause rather than continuing to file appeals into a queue.
Process fixes
- Charge lag: 47 days is inside the 120-day filing limit but leaves almost no cushion. A claim needing correction after rejection can miss the limit entirely, and that is unrecoverable revenue that never appears in a denial report. Set a charge-lag target (5-10 days), report it by clinic monthly, and flag lag over 30 days.
- Contract modeling: load every contracted rate into the system, and run a monthly variance report comparing expected to allowed. This problem was found by a manual sample two years too late; a variance report would have caught it in the first remit cycle.
- Ownership: assign the fee-schedule load verification to a named person at contract execution and at every amendment.