Extrapolated Recoupment: Response Clock and Challenge Points
What is actually challengeable when a payer extrapolates a sample across years of claims, which clock starts on the notice date, and the itemisation to demand first.
On this page: Direct answer
Direct answer
Extrapolated recoupment response: what operators need to know
What is actually challengeable when a payer extrapolates a sample across years of claims, which clock starts on the notice date, and the itemisation to demand first. The response clock runs from the notice date. A notice that reaches you late has already spent part of your window, so record both dates and raise the gap immediately.
Extrapolation is what turns a records request into an existential number. A reviewer reads a sample, finds an error rate, and applies that rate to every claim paid under your NPI for the look-back period. A few dozen charts become a demand measured in hundreds of thousands of dollars, and it arrives with a response window measured in days.
The instinct is to argue the charts. That is part of it, but the leverage is distributed differently than most operators expect, and in one important respect the obvious target is off-limits. This guide separates what is genuinely challengeable from what is not, puts the clock on the notice date rather than the day the envelope reached your desk, and gives you the itemisation request that should go out before you concede anything.
Key takeaways
The short version
- The response clock runs from the notice date. A notice that reaches you late has already spent part of your window, so record both dates and raise the gap immediately.
- Demand the itemisation first: which claims were sampled, which were found in error, and why. You cannot rebut a percentage.
- In Medicare, the decision that the error rate was high enough to extrapolate is shielded from administrative and judicial review. Do not build your case on it.
- What is challengeable in Medicare is the sampling methodology and the individual claim determinations — and overturning claims lowers the error rate the extrapolation rests on.
- Commercial and Medicaid recoupments are governed by your contract and state law, not by the Medicare extrapolation statute. Many states cap look-back absent substantiated fraud.
- Incident-to and supervision documentation is a recurring focus in behavioral-health reviews. Audit your own before someone else does.
Take the template with you
Free to copy · no email required
Itemisation and sampling methodology request
The request to send on day one: the sample list, the per-claim findings and criteria, the sample frame, the statistical method, the reviewer's credentials, and the basis for the look-back.
Saves as extrapolation-itemisation-request.txt
Audit response and sample-scoring tracker
Two things in one sheet: the response clock with both dates, and a row per sampled claim so you can score your own error rate before negotiating against someone else's.
case_ref,payer_or_auditor,line_of_business,notice_date,receipt_date,delivery_gap_days,stated_deadline,extension_requested_date,extension_granted_to,charts_requested,date_range_from,date_range_to,lookback_years,state_lookback_limit_checked,fraud_allegation_made,extrapolation_asserted,demand_amount,itemisation_requested,itemisation_received,recoupment_by_offset,offset_start_date,stayed_pending_appeal,appeal_deadline,counsel_engaged,regulator_complaint_filed,owner,next_action,next_action_due AUD-0001,[payer] via [auditor],fully insured commercial,2026-10-02,2026-10-16,14,2026-10-12,2026-10-17,,140,2023-01-01,2026-06-30,3,pending counsel,no,yes,,2026-10-17,,unknown,,unknown,,yes,no,[name],chase itemisation; score sample internally,2026-10-24 claim_ref,date_of_service,code_billed,rendering_clinician,payer_finding,deficiency_cited,criteria_cited,our_assessment,supervision_documented,time_documented,treatment_plan_current,clinician_enrolled_on_dos,defensible,appeal_this_claim,notes,,,,,,,,,,,,,
Saves as extrapolated-recoupment-tracker.csv
1. The clock, and why it is usually already running
Audit response windows are short — frequently in the range of ten to thirty days depending on the program and the payer — and they are counted from the date on the notice. That is not the date you received it. A request routed through a billing platform, a third-party auditor's fax queue, or a practice's general mailbox can take weeks to reach the person who has to answer it, and the window does not pause for any of that.
The practical consequence is severe: a provider can be found non-compliant for failing to produce records in a window that had already closed before they knew a request existed. So the first two actions are clerical rather than clinical. Record the notice date and the receipt date as separate fields, and if there is a gap, say so in writing on day one and ask for an extension measured from actual receipt. A request made before the deadline and documented is worth far more than an explanation offered afterwards.
| Field | Why it matters later |
|---|---|
| Date printed on the notice | This is what the payer's clock runs from |
| Date it reached the responsible person | Establishes a delivery gap and supports an extension request |
| Channel it arrived by | Fax, portal, platform, or post — each has a different failure mode worth documenting |
| Number of charts and the date range | Tells you immediately whether a look-back limit may apply |
| Stated deadline | Calendar it, then set an internal deadline well short of it |
| Reviewing entity | The payer, or a third-party auditor acting for it — the escalation path differs |
| Whether extrapolation is mentioned | Changes the stakes and the response strategy from the first day |
2. Demand the itemisation before you concede anything
A finding expressed as a percentage is not something you can answer. If the determination says you were ninety-seven per cent non-compliant, there is no argument available until you know which charts were in the sample, which of them were found deficient, and what the specific deficiency was in each. Reviewers do not always provide this unprompted, and the absence of it is itself worth raising.
Ask for the sample frame as well as the findings. The frame is the universe of claims the sample was drawn from, and whether it was drawn properly is one of the few things you can genuinely contest. A request that covers both arrives as one letter and costs you nothing extra.
- The complete list of claims in the sample, with claim numbers and dates of service.
- For each claim, the determination and the specific deficiency identified.
- The criteria, policy, or manual provision applied to each deficiency.
- The sample frame: the universe of claims from which the sample was drawn, and its size.
- The sampling methodology, including the sampling unit, stratification, and how the sample was selected.
- The statistical method used to extrapolate, including the point estimate and any confidence interval applied.
- The reviewer's name, credentials, and whether a licensed clinician reviewed clinical determinations.
- The basis for the look-back period applied.
3. What is genuinely challengeable — and the one thing that is not
For Medicare, section 935(a) of the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 added section 1893(f)(3) to the Social Security Act, which limits the use of extrapolation to cases where there is a sustained or high level of payment error, or where a documented educational intervention has failed to correct the error. That reads like an invitation to argue the threshold was not met.
It mostly is not. The same provision bars administrative and judicial review of the determination that the threshold was satisfied, and federal appellate courts have applied that bar, including in the D.C. Circuit and the Third Circuit. Contractors have also been permitted to make the finding themselves. An appeal strategy whose centre of gravity is "the error rate was not high enough to extrapolate" is aimed at the one target the statute protects.
The leverage is elsewhere, and it is real. The sampling methodology is reviewable. The individual claim determinations are reviewable — and this is the lever that matters most, because the extrapolation is calculated from the error rate the sample produced. Overturn enough individual claims and the rate falls, which pulls the extrapolated figure down with it even though the extrapolation itself was never directly attacked.
| Target | Medicare | Commercial / Medicaid managed care |
|---|---|---|
| The decision to extrapolate at all | Review barred by statute — do not build on it | Governed by contract and state law; often genuinely contestable |
| Sampling methodology and frame | Reviewable; a defective frame or selection is a real argument | Contract and state law; demand the methodology |
| Individual claim determinations | Reviewable, and the highest-leverage route — each reversal lowers the rate | Same; the arithmetic works identically |
| Look-back period | Set by Medicare rules and reopening provisions | State statutes frequently cap it absent substantiated fraud, waste, or abuse |
| Whether a qualified reviewer made clinical findings | Worth establishing and documenting | Worth establishing; some states require clinician review |
| Procedural defects in the request | Document delivery gaps and missing itemisation | Same, plus any contractual notice terms |
4. The look-back limit is the first thing to check outside Medicare
Many states limit how far back an insurer may reach to recover a paid claim — commonly a defined period such as one or two years — with an exception where fraud, waste, or abuse is alleged and substantiated. A three-year look-back asserted without any such allegation may simply exceed what state law allows, and that is a threshold argument that disposes of a large share of the demand without touching a single chart.
This varies substantially by state and by line of business, and self-funded employer plans are generally outside state insurance regulation altogether. So this is a question for counsel and your state's insurance department rather than something to assert from a template. But it should be the first question asked, because the answer can change the size of the problem before any clinical work begins.
- Identify the line of business first: self-funded ERISA, fully insured commercial, Medicaid managed care, Medicare Advantage, or Medicare fee-for-service.
- For state-regulated business, find your state's recoupment look-back statute and whether an exception was invoked here.
- Check your participation agreement for notice, look-back, and offset terms — contracts sometimes grant more than state law requires.
- Note whether any allegation of fraud, waste, or abuse was actually made and substantiated, or merely implied by the look-back chosen.
- Ask whether recoupment will be taken by offset against current claims, and on what schedule — cash flow is often the real emergency.
- Consider a complaint to the state insurance regulator in parallel with the appeal where the facts support one.
5. What reviewers look for in behavioral health specifically
Behavioral-health reviews concentrate on a predictable set of documentation questions, and they are the same ones worth auditing internally before anyone else does. Supervision and incident-to arrangements recur most often, because they combine a provider-eligibility question with a documentation question, and because the rules differ across payers and states in ways that are easy to get wrong at scale.
| Area | What the reviewer is testing | What should be in the record |
|---|---|---|
| Supervision and incident-to | Whether the billing arrangement was permitted and documented | Supervising clinician identity, presence requirement met, signature and date, the arrangement's basis |
| Time-based psychotherapy codes | Whether the documented time supports the code billed | Start and stop times or total time, consistent with the code's definition |
| Medical necessity for the level of care | Whether the record supports the intensity billed | Dated assessments, measures, risk, and the criteria the payer applies |
| Treatment plan currency | Whether a plan existed, was individualised, and was updated | Goals, interventions, review dates, signatures |
| Service delivery evidence | Whether the service occurred as billed | Notes matching dates, duration, modality, and participants |
| Credential and enrolment status | Whether the rendering clinician was eligible on the date of service | Licensure, enrolment effective dates, panel status |
| Group versus individual | Whether the modality billed matches what happened | Participant counts and modality recorded contemporaneously |
6. The first ten days
- 01
Day 0 — log the notice as data
Notice date, receipt date, channel, chart count, date range, deadline, reviewing entity, and whether extrapolation is mentioned. Assign one owner.
- 02
Day 0 — identify the line of business
Everything downstream depends on it: which rules apply, which look-back limit, which appeal path, which regulator.
- 03
Day 1 — send the itemisation and methodology request
One letter covering the sample list, the per-claim findings, the criteria applied, the sample frame, the methodology, and the basis for the look-back.
- 04
Day 1 — request an extension if there was a delivery gap
Cite the notice date and the receipt date. Ask for the window to run from actual receipt.
- 05
Days 2–7 — assemble and review the charts yourself first
Score them against the areas above before you send anything. You need your own view of the error rate, because that number is what the demand is built from.
- 06
Day 7 — involve counsel if extrapolation is in play
Where a demand is extrapolated across years, the look-back and sampling questions are legal ones, and the amounts usually justify the cost.
- 07
Before the deadline — respond in full, noting what is outstanding
Send the records with an index, note the itemisation request that remains unanswered, and preserve every argument explicitly rather than implicitly.
Common questions
Answers before you build.
Can we challenge the decision to use extrapolation at all?+
In Medicare, essentially no. The statute that limits extrapolation to a sustained or high level of payment error also bars administrative and judicial review of the determination that the threshold was met, and federal courts have enforced that bar. Outside Medicare it is a different question entirely, governed by your contract and state law, and there it is often worth contesting.
If we cannot attack the extrapolation, what actually reduces the number?+
Overturning individual claim determinations, because the extrapolated figure is computed from the sample's error rate. Each reversal lowers that rate and the extrapolation scales down with it. The sampling methodology is the other reviewable target: a defective frame or selection undermines the estimate itself.
The notice reached us after the deadline had passed. Is that fatal?+
Not necessarily, but it has to be documented immediately and in writing rather than explained later. Record the notice date and the receipt date, state the gap, and request that the window run from actual receipt. Whether relief follows depends on the program, the contract, and sometimes the regulator — but an undocumented gap is worth nothing.
Does an appeal stop the plan taking the money back?+
It depends on the program and the contract. Some stay recoupment pending appeal; others offset against current claims while the appeal runs, which is a cash-flow emergency separate from the merits. Ask the question explicitly and early, and plan for the answer rather than assuming the favourable one.
Should we just negotiate a settlement?+
Often it is the right commercial answer, but not before the itemisation arrives and you have scored the sample yourself. Settling against a percentage you have never seen the basis for means negotiating without knowing whether the underlying findings would survive. Get the detail, form your own view of the error rate, then decide.
Practical closeout
Use this operator checklist.
- The response clock runs from the notice date. A notice that reaches you late has already spent part of your window, so record both dates and raise the gap immediately.
- Demand the itemisation first: which claims were sampled, which were found in error, and why. You cannot rebut a percentage.
- In Medicare, the decision that the error rate was high enough to extrapolate is shielded from administrative and judicial review. Do not build your case on it.
- What is challengeable in Medicare is the sampling methodology and the individual claim determinations — and overturning claims lowers the error rate the extrapolation rests on.
- Commercial and Medicaid recoupments are governed by your contract and state law, not by the Medicare extrapolation statute. Many states cap look-back absent substantiated fraud.
- Incident-to and supervision documentation is a recurring focus in behavioral-health reviews. Audit your own before someone else does.
Continue through the cluster
Verified customer case studies are added only with customer permission and supporting evidence; none is implied by these operational examples.
Codes on this topic
Look up the code on the remittance
Sources & methodology
Trace the operational claims.
Marsa Health Editorial reviewed the primary and research sources below on October 19, 2026. We translate them into workflow controls, distinguish proposals from final rules, and flag where plan, program, state, contract, or clinical requirements vary.
- 01Change Request 3756: Limitation on use of extrapolation (Program Integrity Manual) Centers for Medicare & Medicaid ServicesThe transmittal implementing section 935(a) of the MMA, which added section 1893(f)(3) of the Social Security Act: extrapolation requires a sustained or high level of payment error, or a documented educational intervention that failed.Accessed or rechecked October 19, 2026
- 02Medicare fee-for-service compliance programs Centers for Medicare & Medicaid ServicesThe hub for the review programs that generate records requests and overpayment findings, and which contractor runs each one.Accessed or rechecked October 19, 2026
- 03Targeted Probe and Educate Centers for Medicare & Medicaid ServicesHow the probe-and-educate rounds work, including the education step whose documented failure is one of the two statutory gateways to extrapolation.Accessed or rechecked October 19, 2026
- 0429 CFR 2560.503-1: Claims procedure Electronic Code of Federal RegulationsParagraph (h)(3)(i) gives group health claimants 180 days to appeal; (h)(2)(iii) requires reasonable access to and free copies of all documents relevant to the claim; (m)(8) defines what counts as relevant.Accessed or rechecked October 19, 2026
- 05Appealing health plan decisions Centers for Medicare & Medicaid ServicesCMS guidance on internal claims and appeals rights, notice contents, and the relationship between internal appeals and external review.Accessed or rechecked October 19, 2026
- 06Medicare and Mental Health Coverage Centers for Medicare & Medicaid ServicesThe MLN booklet covering which behavioral-health services Medicare pays for, the settings they are furnished in, and the supervision and provider-type rules attached to them.Accessed or rechecked October 19, 2026
- 07CMS Interoperability and Prior Authorization Final Rule CMS-0057-F Centers for Medicare & Medicaid ServicesCurrent implementation dates, decision timeframes, denial-reason requirements, metrics, and API provisions for impacted payers.Accessed or rechecked October 19, 2026
- 08Electronic Prior Authorization Centers for Medicare & Medicaid ServicesCurrent CMS provider-readiness guidance for 2027 electronic prior authorization, EHR questions, FHIR testing, and workflow preparation.Accessed or rechecked October 19, 2026
Organizational author. Editorial review covers source accuracy, search intent, workflow boundaries, and human-oversight requirements. This material is educational and does not provide clinical, legal, coding, or coverage advice.
No named clinical or legal expert reviewer is attributed to this version. Marsa Health does not invent reviewer credentials.
Read our editorial methodRevision history
What changed and when
October 19, 2026
Initial publication, source review, and operational editing.