Every healthcare claim has a correct price.
We compute it — line by line, with citation to statute and contract.
Built first for private equity diligence on physician practice rollups.
§ 02 — engine
Healthcare reimbursement is specified by federal law, payer contract, and rule corpora — not estimated. For any claim, there is exactly one correct payment, computable from those sources. CAPE — the Crescent Adjudication and Pricing Engine — resolves that payment line by line, with citation to the controlling rule.
the pipeline
In short: CAPE reads a claim, prices it, accounts for how its lines interact, and fully resolves what the payer and patient each owe.
CAPE reads the claim and determines what is payable, under which provider, coverage, modifier set, and bundling posture. This state conditions every downstream computation.
Each payable line is priced from its governing reference: RVU and locality for procedural codes, ASP for drugs, ASC fee schedule for facility, contract tier for commercial payers. The output is a base price per line, before any cross-line interactions.
Most lines do not stand alone. CAPE applies rules that adjust each line based on the rest of the claim: MPPR ranking, bilateral compounding, endoscopy families, add-on relationships, NCCI bundling, MUE limits. The result is the allowed amount each line earns.
Allowed is not paid. CAPE applies the payer’s financial waterfall to the allowed vector: deductible consumption, coinsurance split, sequestration, payer remittance, patient responsibility. The output is the exact amount the payer owes, the exact amount the patient owes, and the citation chain back to the controlling rule.
cape in practice
Medicare Part B·NON-PAR·assignment accepted·deductible remaining $257.00
PAYMENT WATERFALL·
variance
For every line of every claim in the engagement file, CAPE produces this trace.
single-line variance·line 02, 27447-LT, total knee arthroplasty
systematic pattern
Across 8,432 non-PAR claims for CPTs 27130 — 27447, CAPE identifies consistent payment at 90% of MPFS base where statute specifies 95%. Total cluster exposure: $397,847. Mechanism: systematic 5-percentage-point underpayment on the non-PAR multiplier, persistent across the 24-month window.
at engagement scale
n = 219,847 claims·ortho/spine platform·24-month window
Recoverable underpayment surfaced by CAPE across the full claims file.
in deal terms
At a 10x EBITDA multiple, that is $42M of enterprise value clarity that sample-based diligence cannot see. See § M.02 On the obsolescence of sampling methodology in healthcare diligence for our formal argument on why the predominant sampling-based methodology misses this.
Before CAPE, the gap between what a payer pays and what it owes was unmeasurable at claim-line resolution. CAPE makes that gap computable. What is computable can be recovered, modeled forward, or structured into deal terms.
For the full mathematical specifications, citation registry, and additional worked examples, please visit the engine page.
§ 03 — our engagement
We currently work with private equity firms acquiring or operating physician practices, delivering a new kind of reimbursement quality of earnings report on the full claims file of an acquisition target. We tell you, line by line, with citation to statute and contract, what the practice should have been paid, what it actually was paid, and what your platform will pay it post-close. We focus on four pillars: truth, forward run-rate, upside, and downside.
01 / TRUTH
2 — 11%
recoverable variance, typical · % of net professional revenue
What the practice was contractually owed versus what it actually collected. Line by line, every claim, every payer. A payer applying the wrong MPPR factor to an entire procedure family. A contract addendum being ignored on every claim of a certain type. Modifier policies misapplied across a 24-month window. Each surfaces as a named pattern with citation to controlling authority. On a typical ortho/spine target, the recoverable underpayment compounds to $1.2M — $4.8M annually.
Deal impact: Purchase price negotiation · 100-day revenue recovery plan · EBITDA normalization for LBO model.
02 / FORWARD RUN-RATE
10—15%
delta from normal 24-month projection, typical
The base case the practice is on, before integration. CAPE projects next-12-and-24-month earnings under current ownership, incorporating MPFS conversion factor and RVU restructuring, observed commercial rate trends by payer and MAC, provider-level trajectories, and patient mix shift. Straight-line projection from historical EBITDA misses this — on facility-based ortho codes alone, the 2026 MPFS structural changes produce a net −7% headwind that the headline conversion factor increase obscures. See § F.01 MPFS 2026: the headline narrative on ortho/spine reimbursement is wrong.
Deal impact: LBO model baseline · IC base case · sensitivity scenarios.
03 / UPSIDE
$0.8 — $14M
modeled synergy, per engagement
The target’s claims re-run against your platform’s existing payer contracts, line by line. Contract rate migration. Provider coding normalization to platform standards. ASC facility fee migration — the largest single lever in orthopedic rollups, where the facility fee dwarfs the professional fee by 10 to 15x per case. Ancillary capture. Every lever computed deterministically against the new contract registry, not modeled with a portfolio-average rate ratio that diverges from the deterministic answer by 40–60%. See § F.02 Contract migration synergy is computable. Most deal models don’t compute it.
Deal impact: IC memo synergy defense · purchase price justification · bank financing case.
04 / DOWNSIDE
0.4 — 2.1%
audit-window exposure, % of revenue
Every historical overpayment, modifier misuse, and documentation gap deterministically flagged and bounded by statutory recoupment windows. Modifier 25 overuse on E/M services billed alongside minor procedures — currently an active OIG audit target. Medical-necessity documentation that does not support the billed code. Documentation gaps for codes requiring specific supporting elements. Each finding bounded by RAC and MAC recoupment windows by claim type. The 18-month post-close clawback that destroys fund returns becomes a quantified line item in the deal documents.
Deal impact: Escrow and holdback structure · indemnity negotiation · reps and warranties scope.
Engagement turnaround is two weeks from data delivery. The deliverable is a negotiation brief, a full analytical report with citation chain, and a live portal that stays open through close and converts to a monitoring layer post-acquisition. HIPAA-compliant. Security documentation on request. Full engagement detail
§ 04 — our thesis
Healthcare reimbursement has an exact answer. The price of any claim is a deterministic function of roughly five hundred interacting rules across the CFR, the Medicare Physician Fee Schedule, NCCI edits, MUE limits, local LCDs, sequestration adjustments, contract addenda, and patient deductible state. The function exists. The inputs exist. The answer is computable[1]. The industry has spent thirty years approximating it.
Approximation was once rational. Compute was expensive. The rule corpus lived on paper. Claims data lived on microfiche. Sampling a hundred claims, re-adjudicating them manually, and extrapolating to two hundred thousand was not a methodology of first choice — it was the only operation the constraints permitted. Accounting firms built a QoE practice around it. Revenue cycle vendors built ML scoring systems around it. Private credit lenders built underwriting models around it. The entire analytical infrastructure of healthcare reimbursement was constructed on top of an approximation paradigm because calculation was not available.
Those constraints have lifted. The methodology has not.[2]
The cost of approximation in an era that permits calculation is not a cost of precision but a cost of category. Approximation washes out the mechanisms that produce variance — and the mechanisms are the variance. A 50% MPPR factor misapplied to indicator-1 procedures across an entire payer relationship surfaces in the calculated answer as $487,000 in recoverable underpayment, attributed to a specific edit-engine misconfiguration with citation to the controlling authority. In the approximated answer it surfaces as a population mean of two percent variance, indistinguishable from noise, unactionable in negotiation[3].
When multiple arbitrage carried the math, that gap cost very little. Now it does not. With exit multiples compressed by a third, the three to eight percent of net professional revenue that leaks through systematic underpayment, and the eighteen-month audit exposure that destroys returns, have to be found and quantified at line-level resolution, with citation, and structured into deal terms before close. At a ten-times multiple, three to eight percent of revenue is seven to eight figures of enterprise value mispriced on every deal. Across a platform doing ten add-ons over a fund cycle, the cumulative gap exceeds the management fee. Approximation cannot find this. It is the wrong kind of work.
Crescent calculates. Every claim in the file. Every line on every claim. Against the unified rule corpus and the practice’s payer contracts, with citation to the controlling statute and contract clause that made each adjudication correct. The diligence finding is not an estimate of the population. It is the deterministic answer for every claim in the population[4].
Two funds bid on the same deal. One prices it on a hundred sampled claims and a normalization assumption. The other prices it on the full claims file, line by line, with citation. The first fund wins the auction and loses the deal. The second walks away or structures it for what it is. The methodology gap is no longer a marginal advantage. It is the difference between funds that will return capital and funds that will not.
§ 05 — research
Crescent Research publishes on the mathematical foundations of healthcare payment and the methodological errors that have prevented its proper pricing. Methodology notes develop the theory; field notes apply it. New entries roughly monthly.
selected notes
A note on separating contractual truth from payer-behavioral residual. The realized payment R is decomposable into a deterministic contractual quantity C and a probabilistic residual B. The two admit incompatible methodologies.
A note on the structural information loss in extrapolating one hundred claims to one hundred thousand when the underlying function is computable in closed form. The three constraints that justified sampling have lifted; the methodology has not updated.
A line-by-line read on the 2026 changes. The conversion factor increase is more than offset by structural changes that produce 7-8% declines in 2026 payment for facility-based ortho codes.
The largest synergy in physician practice rollups, and the methodology that systematically mismeasures it. Portfolio-average rate ratios diverge from the deterministic answer by 40-60% in either direction.
For the full corpus, please visit our research page.