The Seven Machine-Readable Payers: What a Week of Scorecard Reactions Revealed

On Saturday we published the CMS-0057-F compliance scorecard for the March 31, 2026 metrics deadline. We checked 1,308 payers. 9% showed at least partial compliance. 85% showed none.

Four days of responses later, one number keeps coming up in our inbox and in conversations with payer compliance leaders: the seven machine-readable payers. Why that number, why those plans, and what it says about January 1, 2027.

The seven, by name

Only seven payers in our 1,308-payer dataset posted CY2025 PA metrics in a machine-readable format. All seven are state-run programs:

Utah Medicaid FFS
Wisconsin Medicaid FFS
Wyoming Medicaid FFS
Utah CHIP
Wisconsin CHIP
Wyoming CHIP
(one additional small state plan)

Zero commercial payers. Zero of the largest Medicare Advantage organizations. Zero of the major QHP issuers on the federally facilitated exchanges.

The largest payers with no evidence of metrics publication at all include Wellcare (8.9M enrolled), Aetna Medicare S5601 (4.2M enrolled), Ambetter QHP (4.2M), Humana S5884 (3.8M), and BCBS QHP (3.0M). These five plans alone cover roughly 24 million American lives and together account for more member-impact than the seven compliant payers combined by several orders of magnitude.

That is the scorecard in a single comparison. Small state Medicaid programs are leading on transparency. The national commercial carriers are not.

Why state Medicaid programs were ready and commercial payers were not

The obvious hypothesis, that state Medicaid programs have a data advantage, holds up. State Medicaid agencies have been publishing structured program data for years under the Transformed Medicaid Statistical Information System (T-MSIS) and the CMS Medicaid Managed Care reporting regime. Producing a machine-readable PA metrics file is an incremental extension of existing data pipelines they already run.

Commercial plans and Medicare Advantage organizations, by contrast, have spent the last two decades producing PA analytics for internal underwriting, medical policy, and quality committees. That data lives in claims platforms, utilization management systems, and proprietary rules engines. None of it was designed for public reporting, and the organizational question of which system of record owns the “official” external denial rate is a governance problem before it is a technical one.

That governance problem is why the commercial payer count of machine-readable submissions for the first CY is seven, which is seven plans that are not commercial.

What January 2027 actually requires

The January 1, 2027 deadline under CMS-0057-F is not another reporting cycle. It is a set of operational APIs.

Prior Authorization API (PAS) requires real-time submission and status of PA requests between providers and payers using the HL7 Da Vinci PAS implementation guide.
Coverage Requirements Discovery API (CRD) lets provider systems query what a payer requires before submitting.
Documentation Templates and Rules API (DTR) returns structured forms and decision logic for a specific PA scenario.
Provider Access API lets contracted providers pull claims, encounter, and clinical data.
Payer-to-Payer API moves member data between plans when a member switches coverage.

Each of these requires that PA-relevant data, the same data underlying the CY2025 metrics report, be exposed as structured FHIR resources, not as a narrative PDF.

A payer that could not publish a machine-readable annual summary in April 2026 is not organizationally ready to stand up a PAS API for live production use in January 2027. The technical work of implementing the Da Vinci profiles is real and non-trivial, but the harder work, agreeing internally on which system of record authoritatively answers a PA metrics question, is the work most commercial plans have not done yet.

The April 10 signal from CMS

On April 10, 2026, CMS published a proposed update to the CMS-0057-F metrics framework. The timing is telling. The first metrics year closed December 31, 2025. The reporting deadline was March 31, 2026. Eleven days after that deadline, CMS is already proposing changes.

Agencies do not iterate on a reporting framework mid-cycle because everything worked. They iterate because the submissions they received did not produce comparable, usable data. The April 10 proposed update is a tell: CMS saw the first-round reports, and the agency is using the comment window to tighten the requirements before the CY2026 cycle.

Compliance readers should expect the CY2026 reporting requirements to be stricter. Comments on the April 10 proposal are due in roughly 60 days. Any plan that intends to stay in the “published” column next year should be reading the proposal today.

Three actions for this week

If you run PA compliance at a plan in the silent 1,185, publish something this week. Even a PDF is better than silence. You are on a list that becomes a plaintiff’s bar target as soon as someone indexes it.

If you run data engineering at a plan, scope a FHIR MeasureReport pipeline for CY2026. You have roughly 11 months to prototype. If you are still debating who owns the official denial rate, start there. The technology is downstream of that decision.

If you run strategy or government affairs, read the April 10 CMS proposal, file a comment, and model your plan’s position in the next scorecard. The first public scorecard existed. The second one will exist. The question is whether your plan shows up on the left side or the right side of the table.

The 7 machine-readable payers led. They are not going to lead for long. The question is who joins them.

This analysis is part of Artificer Health’s ongoing monitoring of CMS-0057-F compliance. The raw dataset is available at artificerhealth.com/payer_publication_status.json and the full scorecard is at artificerhealth.com/compliance.

Sources