Star Ratings Cut Points: Why Your Plan's Score Depends on Everyone Else's

Cut points are a contest, not a test

The quality team at a mid-sized Medicare Advantage plan closes thousands of care gaps each year. Their HEDIS rates climb. Their CAHPS scores edge up. Their Star Rating does not move. Sometimes it drops.

The reason is structural. The Star Ratings are not a test a plan passes or fails against a fixed bar. They are a tournament in which a contract’s score is graded on a curve set by the entire industry. The cut points that determine whether a breast cancer screening rate earns 3 stars or 4 stars derive from the previous year’s distribution of every contract’s performance. If the industry improves faster than the plan does, the rating falls even as absolute performance rises. It is not a bug; it is the design principle CMS chose to drive continuous improvement.

How CMS sets cut points each year

Each Star Ratings measure receives cut points that map a contract’s numeric performance, such as a 78% breast cancer screening rate, to a star rating from 1 to 5. The methodology, detailed in the annual CMS Part C and D Star Ratings Technical Notes, uses a hierarchical clustering algorithm applied to the distribution of all contracts’ scores from the prior measurement period (CMS, 2026 Star Ratings Technical Notes). The process works as follows:

• Data collection. CMS calculates every contract’s performance on every measure using the prior measurement year’s data.
• Clustering. For each measure, CMS applies a clustering algorithm (mean resampling with hierarchical clustering) to the distribution of contract-level scores. This identifies natural breakpoints in the data where contracts’ performance meaningfully differs.
• Tukey outlier deletion. Before clustering, CMS removes Tukey outer fence outliers - contracts whose measure-specific scores fall outside Q1 − 3.0 × IQR or Q3 + 3.0 × IQR (CMS, 2026 Star Ratings Technical Notes). This prevents extreme outliers from distorting cut points for all other contracts. Tukey was finalized in the CY 2021 Final Rule and has been in effect since the 2024 Star Ratings (CMS, 2025 Star Ratings Fact Sheet, Oct 2024; CMS, 2026 Star Ratings Technical Notes).
• Cut point assignment. The clusters are mapped to the 1-to-5 star scale. The thresholds between clusters become the year’s cut points. A contract whose score falls above the 4-star cut point earns 4 stars on that measure; above the 5-star cut point, 5 stars.
• Relative, not absolute. Because cut points derive from the prior year’s distribution, they shift every year based on industry-wide performance. If every contract improves by 2 percentage points on a measure, the cut points rise by roughly the same amount, and a contract that held steady loses stars.

The 2025 Star Ratings Fact Sheet (CMS, Oct 2024) explicitly noted that “changes in measure-level cut points for 2025 Star Ratings were impacted by various factors,” including trend adjustments, the natural movement of the distribution, post-pandemic normalization, and outlier removal effects (CMS, 2025 Star Ratings Fact Sheet, Oct 2024). CMS publishes a Cut Point Trends file (Part C and D Performance Data) showing historical cut point movement across all measures - essential reading for any Star Ratings strategist (CMS, Part C and D Performance Data).

Three imperatives follow from the curve

The cut point methodology creates three strategic imperatives for health plans:

• You must outrun the industry, not just your own baseline. A 2% year-over-year improvement on a measure with a 3% industry-wide gain is a net loss.
• Cut points are predictable but not guaranteed. CMS publishes cut points after the measurement year. Plans forecast them but cannot lock them.
• High-performing measures are harder to move. At the distribution’s upper tail, further improvement requires disproportionately more effort because the plan is competing against the already-excellent tail.

The strategic takeaway: do not benchmark against your own prior year. Benchmark against where the cut point is trending. The Quality Health quality portal tracks your measure performance against estimated cut points in real time, so you know which measures need the most attention before the measurement year closes.

How Tukey deletion reshaped outlier treatment

CMS uses the Tukey outer fence outlier deletion method - finalized in the CY 2021 Final Rule and in effect since the 2024 Star Ratings - to remove extreme outlier contracts before clustering (CMS, 2025 Star Ratings Fact Sheet, Oct 2024; CMS, 2026 Star Ratings Technical Notes). The goal is to prevent a single extreme performer from shifting cut points for everyone else.

A brief codification error occurred in 2022: the May 9, 2022 Final Rule accidentally removed the Tukey language from the regulatory text at 42 CFR 422.166. CMS acknowledged the error and finalized a technical amendment in CMS-4201-F (April 2023) to restore the text, well before any affected Star Ratings were calculated (PQA, Summary of CMS-4201-F, Mar 2023; RISE Health, Aug 2022). The transition is part of the same regulatory thread that has drawn legal scrutiny of cut-point methodology, as the Clover Health lawsuit and earlier SCAN and Elevation challenges illustrate.

Tukey’s outer fence method defines outliers using the interquartile range (IQR) (CMS, 2026 Star Ratings Technical Notes):

• Lower fence: Q1 − 3.0 × IQR
• Upper fence: Q3 + 3.0 × IQR

Contracts whose scores fall outside these fences are classified as outliers and excluded before clustering. Tukey’s outer fence method is more robust to skewed distributions, which describes many Star Ratings measures - particularly those where contracts cluster near the ceiling (such as medication adherence measures with many high performers).

Simulations show material impact on ceiling-effect measures

CMS published Tukey Outlier Deletion Simulations on the Part C and D Performance Data page ahead of the 2024 implementation (CMS, Part C and D Performance Data). These simulations compared cut points under the pre-Tukey method versus Tukey deletion across multiple rating years and measures. Key findings from CMS’s analysis:

• Minimal impact on most measures. For measures with roughly symmetric distributions, Tukey and the prior method produce nearly identical cut points.
• Material impact on skewed measures. For measures where contracts cluster near the top (common for medication adherence and some screening measures), Tukey deletion identifies and removes a different set of outliers, shifting the 4-star and 5-star cut points.
• Directional uncertainty. On some measures, Tukey deletion raises cut points (making it harder to achieve high stars); on others, it lowers them. The direction depends on whether the outliers being removed were pulling the distribution up or down.

What to model under Tukey deletion

• Ceiling-effect measures may see cut point shifts. Medication adherence measures (Part D) and some HEDIS measures with high industry averages are the most affected. Monitor your position relative to the IQR, not just the mean.
• Non-normal distributions are handled differently. Measures with bimodal distributions (some plans do very well, others very poorly with few in between) receive different outlier treatment under Tukey.
• Historical cut point trends from the pre-Tukey era are less predictive. Since Tukey applies to the 2024 Star Ratings and beyond, historical cut points computed under the old method are not directly comparable for skewed measures.

Plans should model their measure distributions under Tukey thresholds. The Quality Health quality portal’s real-time measure tracking lets a plan see its contract’s position in the distribution, so it knows whether it is the outlier being removed or the plan that gains from its removal.

The reward structure shifts: administrative measures removed, HEDIS weights rise

The CY 2027 Final Rule (CMS-4208-F3, finalized April 2, 2026) introduced the most significant Star Ratings structural changes in years (CMS, CY 2027 Final Rule, CMS-4208-F3, Apr 2026). CMS removed 11 measures (mostly administrative process measures) from the Star Ratings program, concentrating the remaining measure set on clinical outcomes and member experience. Key weight changes include:

CAHPS weight advantage: the member experience lever

Individual CAHPS survey measures (Getting Needed Care, Getting Appointments and Care Quickly, Customer Service, Rating of Health Plan, Rating of Drug Plan, Care Coordination) carry a weight of 2 each in the summary score calculation, compared to weight 1 for HEDIS process measures (CMS, 2026 Star Ratings Technical Notes). At the domain level, the six CAHPS measures collectively contribute 21% of the summary score - approximately 4× the contribution of any single HEDIS process measure within the 28% clinical domain. Despite this structural advantage, many plans allocate disproportionate resources to clinical gap closure (closing mammogram and A1C gaps) and underinvest in the member experience improvements that drive CAHPS.

Why CAHPS is hard to move:

• Survey methodology. CAHPS is administered to a random sample of enrollees by an independent vendor. Plans cannot target which members receive the survey.
• Lag time. The survey field period for a given Star Ratings year runs March to May of the measurement year (CMS, 2026 Star Ratings Technical Notes). Interventions need to be in place at least 6 to 12 months before the field period to change member perceptions.
• Multi-component intervention needed. Single interventions (such as a welcome call) rarely move CAHPS scores meaningfully. The evidence supports multi-component strategies: provider-level feedback, member journey mapping, call center quality monitoring, and targeted outreach to members with recent service encounters.
• Competitive dynamic. Because CAHPS cut points are also relative, industry-wide investment in member experience raises the bar for everyone.

Evidence-based CAHPS improvement strategies include:

• Real-time service recovery: Closing the loop on member complaints within 48 hours, with documented follow-up.
• Provider portal with CAHPS-linked metrics: Giving network providers visibility into their attributed members’ experience scores, tied to incentive payments.
• Call center quality assurance with CAHPS-aligned scoring: Monitoring and coaching call center staff against the specific behaviors CAHPS surveys measure.
• Targeted post-encounter surveys: Sending brief (2-to-3-question) surveys after key encounters (prior authorization decisions, specialist referrals, customer service calls) to identify dissatisfaction in real time rather than waiting for the annual CAHPS result.

The improvement measure is a weighted accelerant

The Part C and Part D improvement measures each carry a weight of 5 in the summary score calculation - the highest per-measure weight in the Star Ratings program (CMS, 2026 Star Ratings Technical Notes). The improvement measure is calculated as follows (Attachment I, CMS, 2026 Star Ratings Technical Notes):

1. For each eligible attainment measure, calculate the year-over-year change score.
2. Determine statistical significance using a two-sided t-test at the 0.05 level.
3. Calculate net improvement per measure class (outcome/intermediate outcome, patient experience/access, process).
4. Compute a weighted improvement score: (3 × Net_Imp_Outcome + 2 × Net_Imp_PtExp + 1 × Net_Imp_Process) / (3 × Elig_Outcome + 2 × Elig_PtExp + 1 × Elig_Process).
   Where:
   Net_Imp_Outcome = Net improvement for outcome and intermediate outcome measures;
   Net_Imp_PtExp = Net improvement for patient experience/complaints and access measures;
   Net_Imp_Process = Net improvement for process measures;
   Elig_Outcome = Number of eligible outcome and intermediate outcome measures;
   Elig_PtExp = Number of eligible patient experience/complaints and access measures;
   Elig_Process = Number of eligible process measures.
5. Convert the improvement measure score to a 1–5 star rating using clustering, with zero (no net change) centered at 3 stars.

A contract’s improvement measure star rating is multiplied by weight 5 in the weighted summary average. In a system where half-star thresholds often come down to a few tenths of a point, that weight is material.

The improvement paradox

The improvement measure creates a paradox for high-performing plans:

• Low-performing (2.5 to 3.5 stars): Significant headroom for improvement. Even modest HEDIS gains produce large standardized improvement scores. The improvement measure is a tailwind.
• Mid-performing (3.5 to 4.0 stars): Gains are achievable but require focused effort on specific measures near cut-point thresholds. The improvement measure can be a differentiator from peers at similar star levels.
• High-performing (4.5 to 5.0 stars): Gains are difficult; the plan is already near the ceiling on most measures. The improvement measure can become a headwind if regression to the mean causes even a small drop.

Strategy by star band

For 3.0 to 3.5 star plans: Prioritize the handful of measures closest to the next cut point. A plan moving from 3.0 to 3.5 stars captures not just the improvement measure benefit but potential movement into QBP eligibility. Focus on HEDIS process measures (BCS, EED, Colorectal Cancer Screening) where measurement-year interventions (closing care gaps before December 31) directly improve rates without the CAHPS lag.

For 4.0-star plans: The improvement measure still matters but is harder to move. Focus on Part D medication adherence measures; these are high-weight (3 each (CMS, 2026 Star Ratings Technical Notes), and interventions (MTM, adherence packaging, refill synchronization) produce measurable results within a single measurement year. Protect against regression in CAHPS, which can wipe out clinical gains.

For 4.5-star plans: Defensive strategy. The improvement measure is likely neutral or slightly negative. The priority is holding every measure at or above the 4-star cut point; a single measure dropping to 3 stars can drag the summary score below 4.0. These plans need real-time visibility into measure-level performance during the measurement year.

A half-star shift moves millions in revenue

Quality bonus payments (QBPs) are the financial incentive behind the Star Ratings. Under 42 CFR Part 422 Subpart D and F, MA plans that achieve 4 or more stars receive a bonus added to their benchmark (Congressional Research Service, R48882, Mar 2026; 42 CFR Part 422, Subparts D and F):

A plan with 4.0 stars receiving a 5% QBP on a $1,000 PMPM benchmark earns an additional $50 PMPM. For a 50,000-member contract, that is $30 million per year.

The financial stakes concentrate at the 3.5-to-4.0 and 4.0-to-4.5 star transitions:

• 3.5 to 4.0: The difference is 5% of the benchmark, every year. This is the single largest revenue upgrade in MA, going from zero QBP to full QBP.
• 4.0 to 4.5: Both earn the 5% QBP, but 4.5-star plans receive a 70% rebate share (versus 65%), effectively retaining more of the bid-benchmark difference for supplemental benefits.

The removal of 11 administrative measures creates a new financial dynamic. Contracts that previously relied on high administrative measure scores to buffer weaker clinical performance now face direct exposure on HEDIS and CAHPS measures. The reward factor (which can add up to approximately 0.3 stars for consistent high performance) remains in place unchanged (CMS, CY 2027 Final Rule, CMS-4208-F3, Apr 2026).

The 2027-2028 playbook: six decisions

The assumptions a Star Ratings strategy was built on five years ago no longer hold:

• Old: “If we improve our HEDIS rates, our stars will go up.” New: “We improve our HEDIS rates, but so does everyone else; cut points may rise faster than our gains.”
• Old: “The reward factor is a reliable 0.3-star tailwind.” New: “The reward factor remains in place, but the removal of 11 administrative measures concentrates risk in clinical and experience domains. A contract that previously used administrative measure scores to cushion weaker clinical performance no longer has that buffer.”
• Old: “We can forecast cut points from last year’s values.” New:“ Tukey outlier deletion has been in effect since the 2024 Star Ratings, using the outer fence (3.0 × IQR) rather than prior outlier methods. Historical cut points from the pre-Tukey era are not directly comparable for skewed measures. Plans should model their distributions under Tukey thresholds using CMS’s published simulations.”
• Old: “CAHPS is too hard to move; focus on HEDIS.” New: “CAHPS measures carry twice the per-measure weight of HEDIS process measures (weight 2 vs. weight 1) and collectively represent 21% of the summary score. Underinvesting in member experience is leaving stars on the table.”

1. Track measure performance against estimated cut points in real time. You cannot manage what you cannot see. The Quality Health quality portal tracks every Star measure against historical and forecasted cut points, updated as claims and chart review data flows in throughout the measurement year. By October, you should know, with high confidence, which measures are at risk of dropping below a star threshold.

2. Model your distribution position under Tukey deletion. Download CMS’s Tukey Outlier Deletion Simulations (available on the Part C and D Performance Data page (CMS, Part C and D Performance Data). Tukey has been in effect since the 2024 Star Ratings. Map your contract’s measure scores against the IQR-based outer fence thresholds (Q1 − 3.0 × IQR to Q3 + 3.0 × IQR (CMS, 2026 Star Ratings Technical Notes). Measures where you are near the fence deserve extra attention.

3. Build measure-level performance tracking under the new weight structure. The removal of 11 administrative measures shifts weight to HEDIS and CAHPS. Contracts that previously depended on administrative measure scores to buffer weaker clinical performance need to reassess their entire measure mix. The Quality Health portal tracks every Star measure, so you can see exactly how the weight redistribution affects your summary score projection on today’s data, not last year’s.

4. Invest in CAHPS before the measurement year starts. CAHPS surveys field between March and May (CMS, 2026 Star Ratings Technical Notes). Interventions must be in place by January to influence the survey window. Multi-component programs (call center quality monitoring, provider portal integration, real-time service recovery) outperform single interventions.

5. Prioritize by measure weight and gap-to-cut-point distance. Not all measures are equal. A 2-point gain on a CAHPS measure (weight 2) is worth twice as much as the same gain on a HEDIS process measure (weight 1). Use the measure weights published in the annual Technical Notes to build a weighted priority list: weight multiplied by distance to next cut point equals improvement priority score (CMS, 2026 Star Ratings Technical Notes).

6. Use the improvement measure as a weighted accelerant, not a crutch. If you are a 3.5-star plan, the improvement measure - carrying a weight of 5, the highest in the program - can deliver meaningful contribution to the summary score and help close the gap to 4.0. But it reflects year-over-year delta; once you reach 4.0, sustaining those gains without the improvement tailwind is the real challenge. Build the infrastructure (real-time tracking, CAHPS investment) that sustains performance rather than relying on year-over-year delta.

What the Quality Health portal adds

The Quality Health quality portal was built for this specific problem: Star Ratings management where the measurement year is the only window that matters. By the time CMS publishes Star Ratings in October, the data is 10 to 22 months old. You cannot change it. What you can change is the measurement year that is still open.

The portal provides:

• Real-time measure tracking. See your current performance on every Star measure against historical and forecasted cut points, updated as claims and chart review data flows in throughout the measurement year. You do not wait for the annual report to learn where you stand.
• Cut point projections. See your measure scores against historical cut points and trend-based forecasts. Know which measures are within striking distance of the next star level, and which are at risk of dropping.
• Gap-to-target visualization. For each measure, the portal shows how many additional compliant members you need to cross the next cut point. Stop guessing how many gaps to close; know the exact number.
• Measurement-year closing tracker. HEDIS process measures like BCS and EED close on December 31. The portal tracks gap closure progress throughout Q3 and Q4 so you know whether your outreach campaigns are moving the needle in time.
• Provider-level performance. CAHPS-linked metrics at the provider and practice level, so you can target improvement resources where member experience scores are weakest.

Sources

• CMS, “2025 Medicare Advantage and Part D Star Ratings,” Fact Sheet, October 10, 2024. cms.gov
• CMS, “2026 Part C and D Star Ratings Technical Notes,” last updated September 25, 2025. cms.gov
• PQA, “Summary of CMS-4201-F,” March 31, 2023. pqa.memberclicks.net
• CMS, “Part C and D Performance Data.” cms.gov
• CMS, “Contract Year 2027 Medicare Advantage and Part D Final Rule,” CMS-4208-F3, April 2, 2026. cms.gov
• Congressional Research Service, “Medicare Advantage (MA): Proposed Benchmark Update and Other Adjustments for CY2027,” R48882, March 12, 2026.
• RISE Health, “Star ratings and Tukey’s disappearing act,” August 9, 2022. risehealth.org
• 42 CFR Part 422, Subpart D (Quality Improvement) and Subpart F (Submission of Bids, Premiums, and Related Information). ecfr.gov
• CMS, “Medicare Program; Contract Year 2024 Policy and Technical Changes to the Medicare Advantage Program,” CMS-4201-F, Federal Register, April 12, 2023. federalregister.gov
• CMS, “2026 Medicare Advantage and Part D Advance Notice Fact Sheet,” January 10, 2025. cms.gov