Source data
The register is built directly from the annual Star Ratings data tables published by the Centers for Medicare & Medicaid Services (CMS) on the Part C & D Performance Data page. CMS publishes the data tables as ZIP archives of CSV files, one set per plan year. Branmoor ingests:
- Summary Ratings — one row per contract; columns include Part C summary, Part D summary, and Overall star ratings for the plan year
- Domain Stars — one row per contract; columns are the five performance domains (HD1–HD5) that roll up into the summary ratings
- High Performing Contracts — the list of contracts that earned the 5-star high performer designation
- Low Performing Contracts — the list of contracts that received a Low Performer flag from CMS, including the multi-year history that triggered the flag
What's published here
The MA Star Forensics register includes every Medicare contract that appears in the 2025, 2026 CMS Summary Ratings tables. For each contract, Branmoor presents:
- Plan-year ratings — overall, Part C summary, Part D summary stars for each available year, with the CMS non-numeric labels preserved (e.g., “Not Applicable”, “Not enough data available”, “Plan too new to be measured”)
- Domain stars — the five domain-level scores published in the 2026 Domain Stars table
- Trajectory — a Branmoor-computed signal comparing the contract's current-year overall star against the prior year: up, down, stable, new (no prior rating), or gone (prior rating but no current-year rating). Numeric comparison only; non-numeric labels do not produce a trajectory
- CMS designations — the High Performer and Low Performer flags published by CMS for the current rating year
What's not here (yet)
The free public register is intentionally a primary-source navigation surface. It does not yet include the elements that comprise the paid Forensics product:
- Measure-level cut-point forensics — how each measure's cut-points moved between plan years, and which measures are getting harder to clear
- Next-year projection model — a calibrated forecast of the upcoming plan year's overall rating from this year's underlying measure data, with confidence bands
- Dollar-at-stake analysis — the bonus payment and rebate dollars riding on each individual measure improvement, by contract and parent
- Peer cohort comparisons — performance against named peer sets, controlling for SNP status, geography, and parent organization
These are available under named institutional license. Contact [email protected].
Update cadence
CMS publishes the annual Star Ratings data tables in October of each year for the upcoming plan year. Branmoor re-ingests the source files within hours of publication. Mid-year corrections issued by CMS (a small number of contracts have their ratings updated post-publication) are picked up on the next ingest run.
No part of this register depends on any non-public source. Every value can be reconstructed by an external party who downloads the same CMS files.
Limits
Star Ratings methodology is set by CMS and is updated periodically. CMS has, for example, introduced and retired measures, modified the Tukey outlier removal procedure for cut-points, rebalanced measure weights, and adjusted for natural-disaster impact in declared counties. The Branmoor register reflects what CMS published; it does not retrospectively re-baseline prior years against the current methodology. Where a contract shows a non-numeric label (“Not Applicable”, “Not enough data available”, etc.), the label is preserved verbatim from the CMS file.