The CAR-T Reimbursement Gap

Historical analysis. This report evaluates the FY2026 payment environment. CMS has since finalized FY2027 IPPS rates effective October 1, 2026. HealthPulse's preliminary FY2027 analysis identifies an illustrative national operating-plus-capital base payment of $333,445 for MS-DRG 018 and a $49,346 IPPS fixed-loss amount. These figures are not hospital-specific payment estimates.

Published March 2026; one executive-summary sentence qualified September 28, 2026 (it had said the hospital loses money on every case, which the cited sources do not establish for every case), and the summary line above it likewise (it had said the financial model at most oncology programs is structurally broken). Content type: benchmark research. Sources: CMS, OIG, KFF, peer-reviewed journals, ASTCT, MedPAC and manufacturer pricing disclosures.

Executive summary

Three gaps that do not appear in standard revenue cycle reports. The therapy works. At the published cost medians, the FY2026 DRG payment does not cover the drug alone. The DRG pays $314,231. The drug costs approximately $530,000. Total care per case runs $608,100 at the median, and 8.5% of cases exceed $1 million. Outlier payments close part of the gap on 66% of cases, and three structural features of the outlier mechanism guarantee they never close it entirely. In Medicare Advantage the problem compounds: denial rates for high-cost therapies run systemically higher than traditional Medicare, and 65 health systems have exited MA plans across 2024 and 2025 combined. The third gap is the asymmetry between what payers model before they negotiate and what most programs know when they sit down.

Section 1. What CMS pays against what CAR-T costs

MS-DRG 018 covers every inpatient CAR-T case under the Medicare Inpatient Prospective Payment System. The FY2026 base rate, effective October 1, 2025, is $314,231, a 16.8% increase over the FY2025 rate of $269,139 and the largest single-year jump since the DRG was created. The median wholesale acquisition cost across all seven FDA-approved CAR-T products is approximately $530,000, before the infusion suite, the ICU or the 100-day follow-up.

Source: Avalere Health, Sept 9, 2025; CMS FY2026 IPPS Final Rule, Federal Register, Aug 4, 2025.

The clinical trial adjustment factor deserves specific attention. CMS cut it from 0.33 to 0.16 for FY2026, a 43% reduction, and the effective clinical trial payment drops to $50,277. For a drug costing approximately $500,000 that is an operating crisis for programs treating trial-enrolled patients, well past a reimbursement challenge.

Drug acquisition cost across all seven approved products ranges from $462,000 for Tecartus to $593,533 for Kymriah, with a median around $530,000. Every approved product exceeds the FY2026 base rate before a single bed-day is counted. Total cost of care, which adds conditioning chemotherapy, inpatient management, CRS and neurotoxicity management and follow-up, runs to a published median of $608,100 (Di et al., JNCI Cancer Spectrum, Aug 2024), with 8.5% of cases exceeding $1 million.

Why outlier payments do not close the gap

MS-DRG 018 has the highest outlier qualification rate of any DRG in the IPPS system, at 66% of cases by FY2023. Three structural features limit how much it can recover.

• The 20% haircut. Medicare pays 80 cents on the dollar for all excess costs above the fixed-loss threshold. The remaining 20% is unrecoverable by design.
• Charge compression through the cost-to-charge ratio. Drug charges are converted to costs using the hospital's pharmacy CCR, typically 0.25 to 0.35, which systematically understates actual drug acquisition costs for high-WAC products.
• The threshold floor. The first $40,397 of excess costs above the DRG base rate receives no outlier reimbursement at all.

Published per-case loss estimates, even after outlier payments, range from $63,000 to $304,000 at academic medical centers (Penn Center for Cancer Care Innovation / JCO, 2020). A separate survey of 92 FACT/REMS-certified centers found drug acquisition cost alone exceeded the full DRG payment by a median of $186,500 (Leech, Cohen, Jagasia et al., The Oncologist, 2020), with financial viability reported as the top barrier to CAR-T program administration.

Section 2. The Medicare Advantage denial problem

Medicare Advantage enrolled 33.6 million beneficiaries in 2024, 54% of all eligible Medicare recipients. For oncology programs that means over half of patients walking through the door are on MA plans, where prior-authorization denial dynamics for high-cost therapies differ structurally from traditional Medicare.

A 2022 HHS Office of Inspector General review (OEI-09-18-00260) of the 15 largest MA organizations found that 13% of prior-authorization denials met Medicare's own coverage rules, and every one of those denials was reversed on appeal. OIG's 2026 MA Compliance Program Guidance cites that report and flags AI-driven prior authorization as an emerging enforcement risk.

Source: KFF, Jan 28, 2026; KFF, Aug 8, 2024.

The 80.7% appeal-overturn rate is the number that matters most. The vast majority of denied MA prior authorizations are ultimately approved when challenged. Only 11.5% of denials are ever appealed, so approximately 3.6 million denials per year go unchallenged against an expected overturn rate above 80%. UnitedHealthcare carried the highest denial rate among major MA insurers at 12.8% in 2024. Humana was lowest at 5.8%.

MedPAC's March 2025 Report to Congress documented a 20% cost premium for MA enrollees over equivalent fee-for-service beneficiaries, approximately $84 billion annually, attributed to coding intensity and favorable selection. The most direct evidence of the resulting strain is behavioral: 32 health systems exited MA plans in 2024 and 33 more in 2025, including Mayo Clinic, Mass General Brigham and Mount Sinai, citing prior-authorization burden and reimbursement inadequacy.

CMS does not publicly report MA prior-authorization data broken out by service type, so a CAR-T-specific MA denial rate does not exist in published form. What the research does show: a University of Nebraska Medical Center analysis found private and MA-insured CAR-T patients experienced a median 37-day delay to apheresis against 17 days for publicly insured patients, a 20-day difference (P=.0005) that is clinically significant for a population that has already failed two or more prior lines of therapy.

Section 3. The contract intelligence gap

The per-case gap and MA denial dynamics are structural features of the reimbursement environment. They exist whether or not a program negotiates aggressively. The contract intelligence gap is different: it is solvable, and only with better information. Payers model CAR-T costs before any negotiation happens. Most cancer center programs launch without negotiated single case agreements, without Revenue Code 891 billing protocols built into charge capture, and without a payer-by-payer analysis of where the outlier threshold sits relative to expected costs.

A single-center University of Nebraska Medical Center study (n=65) found that 68% of CAR-T patients required an SCA, and that SCA negotiation added a median of 24 days to the path to treatment (P<.0001). Brain-to-vein time was 17 days without an SCA against 33 days where one was required. A separate ASH 2023 analysis found 66% of private and MA-insured patients required an SCA. No multi-center benchmark exists for what programs actually negotiate against standard contract rates, because commercial rates remain proprietary.

Revenue Code 0891 is the designated code for reporting CAR-T drug costs on institutional claims, paired with product-specific HCPCS Q-codes. Revenue Code 0892 is for gene therapy. CPT codes 38225 through 38228 replaced the prior Category III codes effective January 1, 2025, and a December 2025 CMS billing instruction (MM14247) clarified three MS-DRG 018 payment-adjustment scenarios that billing teams need to identify before a claim leaves the building. Inpatient CAR-T drug costs are bundled into the MS-DRG 018 payment. Outpatient administration is paid separately at ASP plus 6%, subject to the 72-hour rule, a distinction that matters for programs exploring outpatient administration.

All seven FDA-approved CAR-T products require prior authorization from every major commercial payer. Anthem/Elevance's site-of-service requirement, Center of Medical Excellence designation with community-only offices explicitly excluded, is the most restrictive among major payers. No major commercial payer has publicly documented outcomes-based agreement structures for CAR-T as of this report. Where they exist they are negotiated confidentially between manufacturers and individual payers.

Research question

Does FY2026 Medicare reimbursement for CAR-T therapy cover the documented cost of delivering it, how do Medicare Advantage claims-administration patterns affect realized payment, and what accounts for the gap between what payers model in negotiation and what most programs bring to the table?

Evidence base

CMS FY2026 IPPS Final Rule (Federal Register, Aug 4, 2025); Avalere Health (Sept 2025); Pulice & Schulman, HMPI (Feb 2026); Di et al., JNCI Cancer Spectrum (Aug 2024); Ghobadi et al., Advances in Therapy (Jun 2024); Cliff et al., ASCO Educational Book (Jul 2023); Penn Center for Cancer Care Innovation / JCO (2020); Leech, Cohen, Jagasia et al., The Oncologist (2020); HHS OIG Report OEI-09-18-00260 (Apr 2022); KFF (Jan 2026, Aug 2024); MedPAC March 2025 Report to Congress; Heartland Institute (Dec 2025); UNMC / Blood: Advances in Immunotherapy (Dec 2025); ASH 2023 Annual Meeting; ASTCT CAR-T Coding and Billing Guide (Dec 2025); CMS MM14247 (Dec 2025).

Methodology

Comparison of FY2026 CMS payment rates against published drug-acquisition and total-cost-of-care data; analysis of the outlier payment mechanism's structural limits against CMS outlier qualification and payment statistics; cross-reference of OIG and KFF Medicare Advantage denial, appeal and overturn data; review of published single-case-agreement prevalence and billing-protocol requirements.

Limitations

Figures are drawn from published secondary sources and public CMS, OIG and KFF data, not from claims-level analysis of any specific hospital. Drug acquisition cost is a published aggregate across seven approved products and does not reflect any single product's negotiated price. SCA prevalence data, the 68% and the 66%, comes from single-center and single-analysis sources (UNMC, n=65; ASH 2023) rather than a multi-center benchmark. A CAR-T-specific Medicare Advantage denial rate does not exist in published form, so MA findings here are drawn from oncology-level and system-level data rather than CAR-T-specific claims data.

Classification

What the evidence establishes

A documented, publicly sourced gap between FY2026 Medicare payment and CAR-T delivery costs; three specific structural limits on the outlier payment mechanism; a documented pattern of high Medicare Advantage denial-overturn rates against low appeal rates; and documented SCA prevalence and negotiation-delay data at the centers studied.

What it suggests

That most CAR-T programs face a structural reimbursement shortfall rather than an operational one; that MA plans' administrative denial patterns compound that shortfall for the majority of patients now enrolled in MA; and that contract intelligence, meaning an understanding of payer assumptions before negotiation, is a solvable gap distinct from the two structural ones.

What it does not prove

That any specific hospital or program is experiencing this gap, or its magnitude, for a specific payer mix, contract set or patient population. The SCA prevalence figures should not be read as a nationally representative benchmark.

Additional data required for a more specific conclusion

A program's actual payer contracts, claims and remittance data, denial and appeal history, and existing SCA terms where they exist.