How we decide a Federal IDR determination

Methodology version: nsa_idr_passport_v1 · Effective: 2026-09-06 (v1, pre-certification) · Entity: Medigami LLC, a Delaware limited liability company (sole member: Ripplarity Inc), doing business as Trilatic. It is preparing an application for certification as a Federal Independent Dispute Resolution (IDR) entity under 45 CFR 149.510; on certification it will decide out-of-network payment disputes as described below. This page describes, in plain language, how every determination is made. It is not marketing copy — it is the operational description we hold ourselves to, and the one we expect the Departments, the parties to a dispute, and any member of the public to check us against.

What a determination considers, and in what order

For every dispute, we consider information in the order 45 CFR 149.510(c)(4)(iii) sets out:

  1. The Qualifying Payment Amount (QPA) for the item or service, for the applicable year. This is always considered — it is the anchor of every determination.
  2. Then, only if a party raises it and we determine it credible and relevant to the specific offer at issue: - The provider's or facility's training, experience, and quality/outcomes measurements. - Market share held by the provider/facility, or by the plan/issuer, in the relevant region. - Patient acuity, or the complexity of furnishing the item or service. - The facility's teaching status, case mix, and scope of services, where applicable. - Good-faith network-contracting efforts by either side, and — where the provider/facility was previously in-network — the contracted rates between them for the past four plan years.
  3. Also considered: information we ourselves requested from the parties, and any other additional information a party submits — but only if it is credible, relates directly to the offer amount at issue, and is not already reflected in the QPA or in other information we're already weighing. We do not double-count the same fact under two different headings.

We never select an offer by splitting the difference or by any formula not tied to these factors. The written decision states which factors were relied on and why.

What a determination may not consider

By law (149.510(c)(4)(v)), three categories are excluded, no matter who submits them or how they're framed: usual and customary charges (including rates expressed as a percentage of them); billed charges (what would have been billed absent balance-billing protections); and public payor rates (Medicare, Medicaid, CHIP, TRICARE, VA, or state 1115A rates). Every submission is screened for these before drafting begins; flagged material is removed from the record entirely — not left out of the write-up, but excluded from what the reviewer and the drafting tool ever see.

How we are selected

We accept selection by either party to a dispute — the initiating party naming us directly, or the non-initiating party proposing us after objecting to the initiator's first choice — by joint agreement of both parties, or by random assignment from the Departments. We publish our determination outcomes broken out by which of these routes led to our selection (our quarterly resolver record's selecting-party breakdown), and we maintain no commercial, referral, or fee arrangement with any party, payer, provider group, or filer that could influence which route selects us.

How eligibility is decided

Before we look at any dollar amount, we run a nine-rule eligibility check on every dispute, in this order:

  1. NSA applicability. Is the disputed item or service actually a "qualified IDR item or service" under the statute (an emergency out-of-network service, a non-emergency out-of-network service at an in-network facility, or air ambulance service) for which open negotiation lapsed without agreement?
  2. State law. Is it instead payable by Medicare, Medicaid, CHIP, or TRICARE, or governed by a specified state law or All-Payer Model Agreement — in which case the Federal process doesn't apply?
  3. Open negotiation. Did the 30-business-day open negotiation period run its course?
  4. Initiation window. Was the Notice of IDR Initiation filed within the required window after negotiation ended?
  5. Batching. For batched disputes: the 50-line-item cap, the 30-business-day service-date window, and the three permitted batching bases — see "How we batch", below.
  6. Cooling-off. Is the initiating party barred by the 90-business-day "cooling off" period after a recent, similar determination between the same two parties for the same service code?
  7. Duplicate. Has this same dispute id, or this same claim between these same two parties, already been decided?
  8. Wrong party. Does the named plan or issuer match our plan registry? A mismatch is a cure opportunity, not an outright denial — the initiator may re-file naming the correct party.
  9. Missing information. Is every field the Notice of IDR Initiation requires actually present? Also a cure opportunity, with a 5-business-day cure period.

This eligibility review, together with our conflict-of-interest check (below), is completed within 5 business days of final selection, with a 5-business-day cure period for the parties if something is missing. If we determine the Federal process doesn't apply — to all or part of a batched dispute — we notify both parties and the Departments and stop there.

How we batch

Batching does not change the substantive test — each line item in a batch is still weighed under the same considerations as a single dispute. What batching changes is administrative handling: one fee, one timeline, for the items that qualify. Under the June 2026 IDR Operations final rule (91 FR 33949, June 4, 2026), we apply three checks, in this order, to every batched dispute:

An item that fails all three checks is excluded from the batch; it does not sink the batch, and the remaining items proceed. Parties are notified of any exclusion inside the same 5-business-day eligibility window as every other eligibility determination.

How the case reviewer is assigned and screened for conflicts

Every determination is assigned to a case team drawn from our credentialed roster. Before we ever touch the substance of a case, we run a conflict screen covering the preceding 12 months:

We file an attestation of no conflict — or a decline, if one exists — within 3 business days of selection, as the rule requires. A decline sends the dispute back for reselection; it does not proceed with us. We also keep a standing ledger of every commercial relationship our parent company has had with any payer or provider-side organization ("billing agent/filer" below means any entity that submits or manages a Notice of IDR Initiation on a provider's or facility's behalf, whether or not it is itself a party), reviewed quarterly, so a conflict that only becomes material later doesn't slip through.

If you believe we have a conflict. Either party may object to a named case-team member, with reasons, at any point before a determination issues; contact mailto:compliance@trilatic.com. A late- discovered conflict is handled the same as one caught at selection — a decline and reassignment, never a determination issued around it. The 3-business-day attestation itself (clean or a decline) is available to both parties on request.

What happens if a party doesn't respond (default)

If one party submits no offer or no supporting information by the deadline, we issue a default determination based on the other party's timely, admissible submission and the QPA. A default determination goes through the same human sign-off as every other case — it is shorter because there is less to weigh, not because it gets less scrutiny.

How software is used — and how a human decides and signs

Role of software (summary). Every submission is screened by a prohibited-factor filter before drafting begins, and any flagged material is excluded from the record entirely, not merely omitted from the write-up. The offer selected is always the one nearest the credible value on the record under 149.510(c)(4), never a split-the-difference or formulaic result. Any confidence stated in a draft is a probability, never a certainty, and is disclosed as such in our published calibration record. No determination is ever issued without independent review and signature by personnel holding the required credential — a model never issues a determination. We do not communicate with any party about a pending dispute outside the official portal record. The entity answers any inquiry from the Departments about this software's role in full.

Our case-management system does two things, and two things only, before a human ever makes a decision:

  1. Drafting. It assembles the admissible record — the QPA plus any credible (c)(4)(iii) information — and produces a draft written decision identifying which offer it determines best represents the value of the item or service, with a rationale tied to specific factors.
  2. Prohibited-factor detection. It screens every submission for usual-and-customary charges, billed charges, and public payor rates, and strips flagged material before drafting begins.

No determination is ever issued without a credentialed human — an attorney or senior reviewer holding the required arbitration credentials — independently reviewing and signing it. That reviewer checks the offer selected, confirms the rationale cites only admissible considerations and doesn't double-count against the QPA, and confirms no prohibited factor appears anywhere in the reasoning. The reviewer can accept, edit, or override the draft; any override is logged with a written reason. The software drafts; a licensed human decides.

How every determination is sealed and verified

At issuance, every determination is cryptographically sealed — a hash chain entry that cannot be altered without detection. Anyone — either party, the Departments, or the public — can check that a determination's content matches what we issued, without our needing to be trusted on our word, at https://trilatic.com/verify/{determination-id}. Verification never exposes protected health information; it confirms integrity, not clinical content.

The quarterly resolver record

Every quarter, we publish a public record covering: number of determinations issued; median and p90 turnaround time in business days; the share of disputes resolved by determination versus settlement/withdrawal; the share that were defaults; and, once we have enough volume, our own calibration — how often our drafting tool's initial offer selection matched the final signed decision, and how often a determination was later challenged or reversed on judicial review.

k-floor suppression rule: any cell in the published record representing fewer than 11 determinations is suppressed or aggregated with an adjacent cell, so no combination of filters can be used to identify an individual dispute, provider, or plan. This applies even where the raw number would otherwise be informative — protecting re-identification takes priority over granularity.

Methodology version and change process

This page carries a version string (nsa_idr_passport_v1). Any change to the considerations we weigh, the eligibility checks we run, the conflict-screening process, or the software's role is released as a new version with a dated changelog at https://trilatic.com/methodology, and applies only to determinations selected on or after the new version's effective date. Determinations already issued are never re-evaluated under a later version.

How to report a concern

If you believe a determination we issued reflects a conflict of interest, considered a prohibited factor, or otherwise departed from this methodology, contact mailto:compliance@trilatic.com. Concerns involving a possible conflict of interest are also reportable directly to the Departments at FederalIDRQuestions@cms.hhs.gov, or through the public petition process under 45 CFR 149.510(e)(5). We log every concern received, our investigation, and its outcome, and retain that record for six years alongside the determination itself.


*Plain-language summary of our binding SOPs (application docs 02 and 04); those documents, not this page, govern in the event of any discrepancy. Word count target: ≤1,200 words.*