November 13, 2026 update. SMD 26-006 now specifies the audit treatment for determinations that meet the terms of a CMS-approved community-engagement mitigation plan. States can document the applicable approval, covered eligibility action and supporting evidence. Other eligibility requirements continue to apply, and federal funding exposure requires a separate analysis. See the hold-harmless special edition.
A Medicaid eligibility finding should lead to a practical question: who can change the conditions that produced it?
A missing income record, a system defect, an unresolved interface failure, and a disputed federal requirement may all complicate an audit. They require different responses. Training cannot repair every interface. A request for federal clarification cannot excuse a record that a state already had the authority and ability to preserve.
97 or Better™ is our shorthand for the objective of a payment-weighted Medicaid eligibility improper-payment rate below 3 percent. It is an educational aspiration, not a federal certification, a prediction of a state’s results, or a claim that 97 percent of cases are correct. The useful question is how to build and demonstrate improvement before the next measurement.
Start with what the state can change
States can examine whether staff applied the operative rule, retained the supporting evidence, recorded the correct dates, and carried the decision into enrollment and payment systems. They can assign an accountable owner, identify an authorized correction, and test later cases to see whether the defect recurs. Those are management actions, not promises of a particular federal result.
The U.S. Government Accountability Office’s (GAO) 2026 review supplies concrete examples. Delaware audited entries in its document-imaging software after missing-documentation problems. Tennessee changed its application system to capture more complete resource information. Kansas described planned contractor-agreement changes addressing income-verification errors. GAO reported these responses; their existence alone does not establish lasting effectiveness. [1]
The next test should therefore reach beyond completion. Can another authorized reviewer retrieve the evidence and reconstruct the decision without help from the person who made it? Does the revised workflow work during a busy renewal period? If a correction changes eligibility, were required procedures followed and downstream records reconciled?
Workload belongs in this discussion. A process that works only with exceptional attention may fail when queues grow or experienced staff leave. Reviewing capacity, exception queues, handoffs, and uneven performance is part of understanding the cause—not a substitute for fixing it.
A proposed management scorecard for 97 or Better™
Are decisions accurate? — Share passing independent case review.
Can we produce the evidence? — Share of sampled cases reconstructed within a defined response window.
How quickly do we correct defects? — Median days from detection to valid, documented correction.
Do corrections hold? — Recurrence rate in independently retested cases.
Where is performance uneven? — Variation across counties, vendors, or processing teams, with comparable populations and review methods.
These five measures are PERM Insights recommendations and leading indicators. They do not replace the Centers for Medicare & Medicaid Services’ (CMS) payment-weighted Payment Error Rate Measurement (PERM) estimate. Define each sample, denominator, review window, correction criterion, baseline, target, owner, and review date before interpreting a trend. A higher case-review pass rate does not translate directly into a lower PERM payment-weighted rate.
Pair this scorecard with an explicit check for improper denials and timely decisions. Federal corrective-action rules require an eligibility plan to evaluate whether actions to reduce eligibility errors also avoid increased improper denials. Removing an eligible person from coverage is not an acceptable route to apparent improvement. [2]
Some corrections depend on several parties: an eligibility agency, county offices, a system contractor, a data supplier, or CMS. Shared responsibility becomes useful when it identifies the next deliverable and who owes it.
Consider an illustrative data-interface failure. The state can retain request and response timestamps, identify affected decisions, operate a lawful interim procedure, and test reconciliations. A contractor may need to repair the interface. A source agency may need to explain a field or delivery schedule. The state still needs a named owner for coordinating the response.
GAO reported that Pennsylvania officials cited funding and staffing constraints on system improvements, and Delaware described interim guidance while a system fix was underway. These examples show why a corrective-action plan may need both an immediate workaround and a durable repair. They do not establish that every workaround is sufficient or every funding request will be approved. [1]A useful dependency record identifies the unresolved issue, responsible parties, interim control, decision needed, due date, and evidence that will demonstrate resolution.
The CMS mitigation checklist circulated October 21 supplies a concrete example of this shared work. Where a strategy requires approval, the state should connect that approval and its conditions to the affected decisions, implementation dates, reporting milestones, and permanent fix. Track proposed, approved, implemented, and independently tested as different statuses. Our November 6 special briefing examines the evidence needed to support that process. Approval of a mitigation plan does not, by itself, establish its PERM treatment. [4] Preserve what the dependency prevented and what the state did within its own authority. “Waiting on another party” should describe a managed condition, not close the finding.
Separate federal decisions from state corrections
Federal accountability has a concrete place. GAO found gaps in CMS’s review of corrective-action plans and in its analysis of what works across states and reporting years. Its recommendations call for complete plans and systematic analysis of eligibility errors and corrective actions. A state can improve its own evaluation; it cannot create a national learning system on CMS’s behalf. [1]
The federal regulation already requires evaluation of operational improvements, efficiency, error counts, improper payments, and ability to meet CMS-assigned targets. Those five regulatory evaluation elements are distinct from the proposed management scorecard above. An accepted plan and demonstrated correction remain different facts. [2]
Other questions concern interpretation, measurement, or statute. A state can assemble the operative rule, conflicting instructions, affected decision periods, and a precise request for clarification. CMS can address questions within its authority; a statutory change requires federal legislative action. A state’s corrective-action plan cannot amend federal law.
Section 71106 of Public Law 119-21 changes the erroneous-excess-payment framework beginning in federal fiscal year 2030. It does not mean that every reported PERM percentage creates an immediate invoice. Keep a measured estimate, a statutory calculation, a waiver decision, and an actual federal recovery distinct. [3]
Give every finding a next action
For the next management review, select three recurring defects. For each, record the cause, the part the state can change, the external dependency, and any specific federal decision needed. Assign an owner and date. Define the lawful correction and reserve an independent retest.
Retest both the evidence and the outcome. A newly uploaded document may repair retrieval while leaving the original verification failure unresolved. A system release may correct one pathway while creating another. A lower error count may reflect a different sample rather than a more reliable operation.
Keep a second record for questions requiring federal action. Make those requests specific enough to answer, while continuing the improvements that do not depend on the answer.
State and federal accountability belong together. The path toward 97 or Better™ starts with knowing what can change now, who must act next, and what evidence will show that the correction held.
Put this into practice
PERM Corrective Action Effectiveness & Closure Checklist — define the issue, accountable owner, interim control, independent retest, and evidence needed for management closure.
Shadow PERM Implementation Guide — organize a bounded independent review before federal sampling and test whether historical decisions can be reconstructed.
These are PERM Insights educational resources, not CMS forms or certification. Resource access conditions may differ from this article.
97 or Better™ — Before 2029
Education not Advocacy.
Sources
[1] U.S. Government Accountability Office, GAO-26-108017, Medicaid: Improved Oversight Needed of State Eligibility Error Corrective Action Plans, published August 18, 2026; publicly released September 17, 2026. Table 3 and discussion of implementation constraints and CMS oversight.
[2] 42 CFR § 431.992, Corrective action plan, especially paragraphs (a)(2), (b)(3), and (b)(4).
[3] Public Law 119-21, § 71106.
https://www.govinfo.gov/content/pkg/PLAW-119publ21/pdf/PLAW-119publ21.pdf[4] CMS, Proposed State Mitigation and Optional Strategies checklist, p. 1, circulated with the October 21, 2026 state-director email; reviewed November 2, 2026. A public posting of the exact checklist was not located.
Independent educational analysis. Management recommendations and the illustrative interface example are not federal mandates, findings about a named state, or validated performance claims. Sources reviewed November 1, 2026; mitigation context added November 2, 2026.
© 2026 PERM Insights & Perspectives LLC. All rights reserved.