Voluntary CMS LEAD Model
Your benchmark is being set right now. It will not be rebased until 2036.
Measure in 2026. Intervene on 1 January 2027. We will not touch your ordering behaviour until your benchmark is closed — and that is in our contract.
LEAD (Long-term Enhanced ACO Design) is a voluntary CMS Innovation Center model — the buyer's opportunity on the buyer's calendar, not a compliance programme.
The window
As of 11 August 2026, calendar days from that date to the LEAD pre-period end (31 December 2026): 142. Dated facts only — not a live timer.
31 December 2026
ACO REACH concludes 31 December 2026.
1 January 2027
LEAD Performance Year 1 begins 1 January 2027.
CY 2024–2026
The LEAD benchmark is built from claims for 2024, 2025 and 2026.
CMS: Centers for Medicare & Medicaid Services Long-term Enhanced ACO Design (LEAD) Model FAQ
Global / Professional
LEAD offers two risk options: Global, with up to 100% of savings and up to 100% of losses, and Professional, with up to 50% of each.
3% withhold
A 3% quality withhold applies to all LEAD ACOs' benchmarks, earnable back through quality performance. Global-option ACOs are additionally subject to a 1.75–3% benchmark discount.
10% / 30% / 60%
For ACOs new to accountable care the LEAD benchmark weights the three base years 10% / 30% / 60%, so 2026 carries most of the weight. ACOs with prior REACH or Shared Savings Program experience have the three years weighted equally.
Why we will not touch your ordering until 1 January 2027
Measure in 2026. Intervene on 1 January 2027. We will not touch your ordering behaviour until your benchmark is closed — and that is in our contract.
Your benchmark is built from 2024, 2025 and 2026 claims. If you are new to accountable care, 2026 carries 60% of that weight Measured. So a dollar of low-value spend you remove in 2026 mostly reduces your own benchmark, and you carry that lower benchmark until 2036 Measured.
Which is why we measure this year and do not intervene until Performance Year 1 — and why that is a term in the agreement, not a promise in a slide.
LEAD benchmark data appendix — reproducible working
Companion working for docs/PUBLICATION_LEAD_BENCHMARK_2026.md. CMS LEAD payment structure — zero customer data. Figures from lib/tcoc/ledger/lead-benchmark.ts and the Source Register.
Overview
Base years 2024, 2025, 2026. Performance Year 1 begins 2027-01-01. Pre-period ends 2026-12-31. Horizon: 10 years without rebasing. Each share below is labelled Measured and resolves to a Source Register citation.
Counterfactual construction for programme evaluation: /evidence/methodology#counterfactual-construction.
Base-year weights
From benchmarkAbsorptionShare → LEAD_BASE_YEAR_WEIGHTS. Not retyped on this page.
New to accountable care
| Base year | Weight | Provenance |
|---|---|---|
| CY 2024 | 10% | Measured |
| CY 2025 | 30% | Measured |
| CY 2026 | 60% | Measured |
Prior ACO REACH or Shared Savings Program experience
| Base year | Weight | Provenance |
|---|---|---|
| CY 2024 | 33.3% | Measured |
| CY 2025 | 33.3% | Measured |
| CY 2026 | 33.3% | Measured |
Worked absorption example
A permanent $1.00 reduction in CY 2026 claims for a new-to-accountable-care ACO is absorbed into the three-year historical baseline at 60% ($0.60 of the $1.00). That baseline is then fixed for 10 performance years without rebasing. Measured
- Profile
- new-to-accountable-care
- Reduction year
- 2026
- Absorption share
- 0.6
- Absorbed cents per $1.00
- 60
Payment-stack parameters
| Parameter | Value | Applies to | Provenance |
|---|---|---|---|
| Quality withhold | 3% | All LEAD ACOs (earnable back through quality) | Measured |
| Explicit discount (Global) | 1.8%–3% | Global risk option only — taken off the benchmark first | Measured |
| Shared-savings retention | Global 100% · Professional 50% | Two LEAD risk options (savings and losses) | Measured |
| Performance period | 10 years without rebasing (2027-01-01–2036-12-31) | All LEAD ACOs | Measured |
Sources
- LEAD (Long-term Enhanced ACO Design) Model — LEAD (Long-term Enhanced ACO Design) Model, CMS Innovation Center, retrieved 20 July 2026 [cms-lead-model-page]
- LEAD Model Overview — LEAD Model Overview, CMS Innovation Center, retrieved 19 August 2026 [cms-lead-model-overview]
- Centers for Medicare & Medicaid Services Long-term Enhanced ACO Design (LEAD) Model FAQ — Centers for Medicare & Medicaid Services Long-term Enhanced ACO Design (LEAD) Model FAQ, CMS Innovation Center, retrieved 1 July 2026 [cms-lead-tech-faqs]
- LEAD Model Payment Factsheet — LEAD Model Payment Factsheet, CMS Innovation Center, retrieved 12 August 2026 [cms-lead-payment-factsheet]
- ACO REACH Model — ACO REACH Model, CMS Innovation Center, retrieved 20 July 2026 [cms-aco-reach-model]
Full register: /sources.
Why no rebasing changes the maths
Under LEAD the historical baseline stays fixed across the ten-year performance period. A dollar of avoided low-value spend in Performance Year 1 can remain in the margin through 2036 — unlike MSSP, where rebasing at each agreement period competes durable efficiency away. Gross retained, what CMS takes, and net retained are the same §9 engine as /risk.
100,000
6.0 pp (trial midpoint)
12 months
| Arrangement | Gross retained | What CMS takes | Net retained |
|---|---|---|---|
| MSSP BASIC Level ERetention 50% · Withhold 0% · Discount 0% | $2,625,000Modeled | $0Modeled | $2,625,000Modeled |
| MSSP EnhancedRetention 75% · Withhold 0% · Discount 0% | $3,937,500Modeled | $0Modeled | $3,937,500Modeled |
| LEAD ProfessionalRetention 50% · Withhold 3% · Discount 0% Measured | $5,250,000Modeled | $63,000Modeled | $5,218,500Modeled |
| LEAD GlobalRetention 100% · Withhold 3% · Discount 3% Measured | $10,500,000Modeled | $376,110Modeled | $10,123,890Modeled |
The thing that cannot be done later
Once you change how people order, the counterfactual is gone. Nobody can reconstruct what would have happened under the prior pattern of care. Difference-in-differences and interrupted time series need a locked pre-period; a raw before/after after the intervention has started is not a measurement.
The pre-period is the part of this with a calendar bound. Lock it while the LEAD base years are still the present — not a memory.
The offer
Send twenty-four months of professional and facility claims for your attributed population. We return your low-value imaging ledger — attribution mix, risk-adjusted rates, concentration, and organisation-level opportunity. Turnaround: ten business days the first time; three days thereafter. If the number isn't there, we'll tell you, and we'll stop.
One flat file (CSV preferred) — 24 months of professional and facility claim lines for the attributed population. Exact fields in the DATA_REQUEST one-pager. Member ids may be your internal opaque ids; we hash them at the boundary. Member identifiers are hashed at the ingestion boundary. BAA before any real member-level extract. No EHR integration. No commitment to later modules.
Exact fields (claims door DATA_REQUEST)
| Field | Required? | Notes |
|---|---|---|
| Member id | Required | Opaque plan/member id. Not an MRN, government ID, or name. |
| Claim id | Required | Payer claim number or internal claim key. |
| Line number | Optional | Defaults to 1 when omitted. |
| Date of service | Required | Day precision (YYYY-MM-DD, YYYYMMDD, or MM/DD/YYYY). |
| HCPCS / CPT | Required | Procedure code on the line. |
| HCPCS modifiers | Optional | Pipe- or comma-separated. |
| Units | Optional | Defaults to 1. |
| Allowed amount | Optional | Dollars (or cents — say which). Integer cents preferred. |
| Paid amount | Optional | Same unit as allowed. |
| Billing NPI | Optional | 10-digit NPI; leading zeros OK. |
| Rendering NPI | Optional | Performing clinician. |
| Ordering / referring NPI | Required | Item 17b / 837P referring loop. Tier-1 attribution key. |
| Principal diagnosis | Optional | ICD-10-CM; with or without decimal. |
| Other diagnoses | Optional | Pipe-separated ICD-10-CM. |
| Place of service | Optional | CMS POS code (professional). |
| Revenue code | Optional | UB revenue code (institutional). |
| DRG | Optional | MS-DRG when institutional. |
| Claim type | Optional | Professional vs institutional flag if you have one. |
| Provider specialty | Optional | Rendering specialty label or CMS specialty code. |
| Provider state | Optional | Two-letter state of rendering provider. |
This is the actual report you get. It was produced from public data so you can read it before you decide whether to send us anything.
Download sample ledger PDFIf you never buy anything, you keep the report.
What we will not do
- We will not touch your ordering behaviour until your LEAD benchmark is closed.
- ARKA can auto-approve; it can never auto-deny.
- No leaderboard, no rank, no bottom-N.
- ARKA does not store justification text.
From the published catalog — What we will never do.