The 2027 Medicare Star Ratings changes quality leaders should plan for now.

On April 2, 2026, CMS issued the Contract Year 2027 Medicare Advantage and Part D final rule. It is part of a multi-year reshaping of the Star Ratings program that began with the 2020 and 2024 rulemakings. The headline framing is "simpler." For a quality leader, the practical reality is "less room for error." Below are the verified, dated changes that matter, and what to do about each.

Tukey outlier deletion is in effect and it lifted the bar

Tukey outlier deletion removes statistical outliers, frequently a handful of very low performers, before CMS sets cut points for non-CAHPS measures. It was finalized in 2020 rulemaking and has applied since the 2024 Star Ratings, per the CMS Star Ratings technical notes. The effect is mechanical: when low-end outliers no longer drag the distribution down, cut points rise, and the score needed to reach 4 stars on a given measure climbs.

This is not a future risk. It already happened. If your internal targets were set against pre-Tukey cut points, they are stale.

Guardrails are still on, and six 2027 measures do not have one

The 5 percentage point guardrail is still in the regulation and CMS is still applying it. Under 42 CFR 422.166(a)(2)(i), a measure-threshold-specific cut point for a non-CAHPS measure cannot move more than 5 percentage points in either direction from one year to the next. For measures not scored on a 0 to 100 scale the cap is 5 percent of the prior year's restricted range, calculated after outer-fence outliers are dropped. The 2026 Star Ratings Technical Notes document the applied order of operations: Tukey outlier deletion, then mean resampling with hierarchical clustering, then the guardrail on the resulting cut points.

CMS did not propose or finalize any change to the guardrail in the Contract Year 2027 rule. The word appears once in the rule, in a comment summary where commenters asked CMS to preserve stability tools in the methodology. CMS said it would consider the comments and that any further methodology change would have to go through rulemaking.

The part that matters for a forecast is where the guardrail does not reach. It is not applied to the Part C and Part D improvement measures, which carry a weight of 5. It is also not applied to any measure that has been in the program 3 years or less, and a measure returning after a substantive specification change counts as new. For the 2027 Star Ratings that leaves four measures with no cap on cut-point movement: the respecified Colorectal Cancer Screening measure, Care for Older Adults Functional Status Assessment, Concurrent Use of Opioids and Benzodiazepines, and Polypharmacy: Use of Multiple Anticholinergic Medications in Older Adults.

So the planning rule splits in two. On a guardrailed measure, the worst case next year is last year's cut point plus 5 points, and a team can build to that ceiling. On the four uncapped measures and the two improvement measures there is no ceiling, and the only safe posture is to perform well clear of where you think the boundary sits.

Hitting last year's cut point is no longer a plan. On most measures the target can move 5 points. On the new ones it moves as far as the field does.

The health equity reward was reversed, not added

This is the change most likely to be misremembered, so be precise. The 2024 final rule finalized a Health Equity Index reward, later renamed the Excellent Health Outcomes for All reward, to begin with the 2027 Star Ratings. It was designed to reward plans for outcomes among enrollees with social risk factors and to replace the existing reward factor.

In the Contract Year 2027 final rule, CMS finalized its decision not to implement that reward and to continue the historical reward factor instead, the one the equity reward was set to replace. CMS put it plainly: for the 2027 Star Ratings it is not implementing the Excellent Health Outcomes for All reward and will continue the historical reward factor that encourages consistently high performance for all enrollees across all quality measures. The rule was published at 91 FR 17384 on April 6, 2026 and took effect June 1, 2026, and the reward-factor decision applies beginning with the 2027 Star Ratings.

The current regulation confirms the reversal landed. 42 CFR 422.166 now carries the reward factor at (f)(1) and the Categorical Adjustment Index at (f)(2), with no (f)(3). Reward factor values are unchanged at 0.4, 0.3, 0.2, 0.1 and 0.0. If your 2027 forecast assumed an equity-based reward was coming, remove it.

The 2027 measure set: four in, three out

CMS counts four new or updated measures added beginning with the 2027 Star Ratings: a respecified Colorectal Cancer Screening, Care for Older Adults Functional Status Assessment returning after a substantive specification change, Concurrent Use of Opioids and Benzodiazepines, and Polypharmacy: Use of Multiple Anticholinergic Medications in Older Adults. In CMS's words, the last three are process measures developed from evidence-based clinical treatment guidelines, each with a weight of one.

Three measures come out for 2027: Care for Older Adults Pain Assessment, Medication Reconciliation Post-Discharge, and MTM Program Completion Rate for CMR. MTM CMR sits on the display page for measurement years 2025 and 2026 and returns to the Star Ratings as a new measure beginning with the 2029 Star Ratings.

Fewer measures sounds like relief. It is the opposite. Removing measures concentrates the program's total weight onto the measures that remain, so a surviving measure matters more than it did the year before.

Two outcome measures triple in weight for 2027

Improving or Maintaining Physical Health and Improving or Maintaining Mental Health came back into the 2026 Star Ratings at weight 1 after a substantive respecification. For the 2027 Star Ratings both move to weight 3. Two HOS-based outcome measures tripling in weight in a single year is a real forecasting event, and it is easy to miss because nothing about the measures themselves changed between the two years.

The rest of the 2027 weight structure, from Table IV-1 of the 2027 Advance Notice: Health Plan Quality Improvement and Drug Plan Quality Improvement at weight 5; Plan All-Cause Readmissions, Diabetes Care Blood Sugar Controlled, Controlling Blood Pressure and the three Part D adherence measures at weight 3; patient experience, complaints and access measures at weight 2, where they landed after the cut from 4 in the 2026 ratings; process measures at weight 1.

The CY2027 final rule cut 11 measures, and almost none of it hits the 2027 ratings

This is the biggest confirmed story of the year and the one most summaries date wrong. CMS proposed 12 measures for removal and finalized 11. It declined to remove Diabetes Care Eye Exam, which stays in the Star Ratings.

The removals apply to the 2027 measurement period and the 2029 Star Ratings, except for the two Call Center Foreign Language Interpreter and TTY Availability measures and Statin Therapy for Patients with Cardiovascular Disease, which apply beginning with the 2028 Star Ratings. Nothing on this list changes the ratings you are being scored on for 2027.

Removed measures move to the display page and stay publicly reported on cms.gov, so the data collection does not stop even though the measure stops counting toward a star.

The rule adds one measure. Depression Screening and Follow-Up (Part C) starts with the 2027 measurement year and the 2029 Star Ratings. After all of this, CMS says MA-PD contracts are rated on up to 43 measures, MA-only contracts on up to 33, and Part D plans on up to 12.

CMS also put a number on the money. In the rule's regulatory impact analysis, the Star Ratings updates carry a net impact of $5.02 billion in 2028 falling to $1.89 billion in 2036, $18.56 billion over ten years, or 0.21 percent of Medicare payments to private health plans for 2027 through 2036. CMS adds that it is considering removing additional measures in the future, and that any further removal has to go through rulemaking.

You can check CMS's cut point math before the release

Section V.E of the CY2027 final rule amended 42 CFR 422.166(h)(2) and 423.186(h)(2) to codify something CMS had been doing as a courtesy. During the second plan preview, CMS gives sponsors de-identified contract-level sample data for one measure of each type: one CAHPS measure, one Part C clustering measure, one Part D clustering measure, and any measure that needs a different calculation. That is enough to replicate the measure-level cut point calculation on your own numbers. It applies beginning with the 2027 Star Ratings. CMS declined commenters' request to release sample data for every measure.

Put it on the September calendar. The second preview window is the last point where a plan can question a cut point calculation before the ratings publish.

Part D adherence: case-mix adjustment starts with the 2028 Star Ratings

The three Part D medication-adherence measures, covering diabetes medications, hypertension RAS antagonists, and cholesterol statins, are triple-weighted for the 2027 Star Ratings. Table IV-1 of the 2027 Advance Notice lists all three at weight 3 for the January 1 to December 31, 2025 measurement period. Anyone who told you 2027 is the single-weight year has the year wrong.

The change comes after that. In the Contract Year 2027 final rule CMS stated that the Part D medication adherence measures will be adjusted for case mix beginning with the 2028 Star Ratings. The risk-adjusted versions (ADH-Statins RA, ADH-RAS RA, ADH-Diabetes RA) are display measures today. Under the technical notes, a measure returning after a substantive specification change is treated as new, and new measures enter the ratings at weight 1 for their first year. That mechanism is where the widely repeated single-weight year comes from. CMS has not published a 2028 weight table, so treat the weight-1 year as a consequence of the new-measure rule rather than a published figure, and verify it against the technical notes for your measurement year before a forecast depends on it.

Either way, do not read a lower weight as a reason to ease off adherence. The clinical work, and the member outreach behind it, takes months to move. Building adherence capacity in the low-weight year is exactly how you are ready when the weight comes back. We break down the underlying mechanics in our PDC math guide, and the operational playbook in how to improve Part D adherence.

Confirmed as of August 4, 2026, and what is still open

Confirmed: the Excellent Health Outcomes for All reward is not being implemented and the historical reward factor continues for the 2027 Star Ratings. The 5 percentage point guardrail and Tukey outlier deletion both remain in the regulation and in the applied methodology. Eleven measures are finalized for removal on the 2028 and 2029 timelines. Depression Screening and Follow-Up arrives for measurement year 2027 and the 2029 Star Ratings. The plan preview sample-data release is codified beginning with the 2027 Star Ratings.

Still open: CMS says it is considering further measure removals, and any methodology change needs rulemaking. The 2028 weight for the case-mix-adjusted adherence measures has not been published in a CMS table. And CMS has not announced a release date for the 2027 Star Ratings. Expect them on Medicare Plan Finder in October 2026, ahead of Open Enrollment, the way the 2026 ratings published on October 9, 2025. Those ratings will drive 2028 quality bonus payments.

What the changes add up to

Taken together: harder cut points, more volatility, fewer measures carrying more weight each, and a reward methodology that snapped back to where it was. For a plan sitting near a star boundary, a small miss on a high-weight measure now has outsized financial consequences.

The financial mechanism is what makes this a CFO problem as much as a quality-department one. In Medicare Advantage, reaching the 4-star threshold earns a quality bonus payment that flows into plan revenue and rebate dollars, and the difference between 3.5 and 4 stars at the contract level is measured in many millions of dollars. When cut points rise by the full 5 points a guardrailed measure allows, or move without a cap on a measure that has no guardrail, the same clinical performance that cleared the bar last year can fall short this year. The exposure did not change because your members got sicker. It changed because the scoring did.

Why volatility is the hard part

A fourth-quarter sprint only works when the target sits still. Even with the guardrail on, a cut point can move a full 5 points in either direction, and on the four unguardrailed 2027 measures it can move further than that. So the bar has to be cleared with room, all year, on every measure that carries weight. That is a different operating model from sprinting in December to clear a threshold you already know. The plans that struggle most under the new methodology are the ones still managing to last year's numbers.

What quality leaders should do now

1. Re-forecast with the guardrail where it applies, and without a ceiling where it does not. Rebuild your projections post-Tukey: last year's cut point plus 5 points on guardrailed measures, and no ceiling on the four new or respecified 2027 measures and the two improvement measures. Identify every measure where your contract sits within a few points of a star boundary.
2. Triage by weight and proximity. Concentrated weight means a weight-3 outcome measure one point under the cut is worth more attention than three comfortably-passing weight-1 measures. Sequence the work accordingly.
3. Model the two HOS measures at their new weight. Improving or Maintaining Physical Health and Improving or Maintaining Mental Health move from weight 1 to weight 3 for 2027. If your model still carries them at 1, your summary rating projection is wrong.
4. Shift to year-round execution. Volatile cut points punish the year-end scramble. The measures move with continuous member contact: gap closure, adherence, follow-up after acute events. Build the operating cadence that runs all year.
5. Use the second plan preview to check CMS's math. CMS now has to release de-identified sample data for one measure of each type during the second preview, beginning with the 2027 ratings. Replicate the cut point calculation before the October release, not after.
6. Verify every weight against current technical notes. Weights and the measure set are moving year to year. Confirm against the CMS technical notes for your specific measurement year before locking a plan.

This is where execution beats analytics. A dashboard can show you the boundary measures; it cannot make the calls, book the visits, and document the closures that move them. Across Pelica deployments, year-round execution closes roughly 90% of BCS and KED gaps in-year and 70% or more of transitions-of-care gaps within 30 days, with no added headcount.

Sources and further reading

• CMS: Contract Year 2027 Medicare Advantage and Part D Final Rule fact sheet, April 2, 2026 (equity reward not implemented; historical reward factor retained; 11 measure removals; measure counts)
• Federal Register: 91 FR 17384, published April 6, 2026, effective June 1, 2026 (removal timing, plan preview sample data, regulatory impact analysis)
• CMS: 2027 Advance Notice, Table IV-1 (2027 measure list, weights, measurement years)
• 42 CFR 422.166: Calculation of Star Ratings (guardrail text, Tukey, reward factor at (f)(1))
• CMS: 2026 Part C & D Star Ratings Technical Notes (Tukey outlier deletion, guardrail exceptions, new-measure weighting)
• CMS: 2026 Star Ratings fact sheet, November 18, 2025 (2026 ratings published on Medicare Plan Finder October 9, 2025)
• CMS: 2024 Medicare Advantage and Part D Final Rule (CMS-4201-F) fact sheet (original health equity reward finalization)