Four Medicaid Spending Checks Could Not Run on a Majority of ‘Complete’ Reports

A completed form was not a checkable calculation

CMS labeled 595 state-submitted medical-loss-ratio summaries complete for the 2019 reporting year because they contained the four required fields. Yet four accuracy tests each could not run on more than half. Reconstructing the counts published in the September 21 OIG evaluation exposes the difference between collecting a required answer and collecting the information needed to verify it.

What’s the Scoop With Broach subtracted the checked reports from 595: 320 were unavailable for the numerator check, 337 for the first denominator check, 322 for the second and 331 for the adjusted-ratio check. That is 53.8%, 56.6%, 54.1% and 55.6%, respectively.

These are four separate coverage gaps, not four groups to add. A single report can lack information needed for several checks. The published 57 summaries with at least one potential inaccuracy likewise differ from the 82 flags across all five checks; summing the flags would count some reports repeatedly.

An 85% standard depends on usable financial evidence

MLRs compare patient-care and quality-improvement spending with premium revenue adjusted for specified deductions. CMS’s 2026–27 rate guide requires rates reasonably designed to achieve at least 85%. It also calls for recent plan financial performance and comparison with the assumptions used in setting rates.

The guide therefore connects past evidence to future payments. Its accelerated-review documentation describes up to three years of performance and asks states to address significant prior issues. A complete ratio field helps only if the amounts behind it can be checked.

OIG’s finding was potential inaccuracy, not a finding that the 57 flagged summaries establish illegal profits or that every associated rate was too high. The review did not independently validate each underlying state submission against its original source file, and its central data concerned 2019, not this year’s claims.

Older refunds and newer accuracy are distinct parts of oversight

We also checked the live status page for OIG’s May 2025 remittance review. Two of its five recommendations were listed as implemented; three remained open. The implemented items concerned internal communication and procedures for tracking expected returns. The open items included obtaining outstanding federal shares, systematic tracking and routine confirmation.

The newer evaluation addresses an earlier stage: whether the spending ratios are accurate and whether staff reviewing future payment rates receive the accuracy-check results. It should not be read as proof that no improvements followed the older refund review.

CMS agreed with the new recommendations and emphasized more recent changes, including a managed-care reporting tool that blocks blank required fields, a shared repository and public data release. It planned stronger accuracy checks and encouraged optional fields. The distinction remains consequential: a blank-field block can improve completeness without proving that a filled-in figure is correct.

Our work recalculates the published coverage denominators and compares two oversight records with current rate-development guidance. It does not calculate an unsupported dollar-loss estimate or assume every historical gap survives unchanged in 2026.