Welcome to an HTN guest post! This is a new offering we’re tinkering with where we tap into the brain of a thoughtful community member to answer a top-of-mind question for us. You’ll see that question just below. For this post, we invited Peter Yates to share his perspective on the upheaval in the Medicare Star Ratings world.

Lenny Rachitsky’s guest posts from product leaders provided inspiration for this series — you’ll see us borrow some cues from his posts in the structure below. Our goal in these posts is to provide community members with tactical insights as they navigate the changing healthcare landscape.

If you want to join 5,500+ fellow nerds thinking deeply about questions like this, join the HTN community!

Hi, Kevin here with a brief intro. I feel like I’ve been reading and writing constantly about 2025 Stars Ratings for the past month since CMS released its 2025 Medicare Advantage and Part D Star Ratings. It’ll come as no surprise for regular newsletter readers that many health plans are upset with the higher bar CMS has set.

The reaction from plans has been swift — it feels like a new lawsuit has come out every week challenging the rating changes. Beyond the lawsuits, I’ve also seen earnings calls and investor updates with companies describing how they’re attempting to mitigate the damage (or, for some, take advantage of the opportunity). For instance, check out this excerpt from Humana’s Q3 earnings call on October 30th, describing how it is currently on a sprint to impact 2027 financial results:

To summarize: the 2025 Stars score released on October 10th, 2024 was impactful enough for Humana that it caused the organization to increase the number of primary care visits it schedules by 5,000 a week less than a month later. It is scheduling those visits in a scramble that will ideally impact its financial situation in 2027. Humana’s response to the Stars changes seems indicative of just how big of an impact Stars has on plans.

Given all that, we thought it’d be worth digging into this topic more fully. To do so, I teamed up with Peter Yates, who has graciously agreed to share his analysis of the changing Medicare Advantage Star Rating landscape and the practical implications it will have. Peter is the VP of Care Coordination and Partnerships at Signify Health, where he has worked for the past three years. As he notes below, Peter is writing this post in a personal capacity and the views he expresses below are not necessarily reflective of his employer. As the 2025 Star Ratings and subsequent responses from payors have come out, I’ve appreciated Peter's analysis on HTN and LinkedIn. His insights have helped give me context as to why various plans are reacting to the Stars Ratings the way they have.

For folks reading this, I find it to be a helpful 201-level discussion of why the Star Rating system exists, what’s driving the changes in scores today, and the potential future implications of the program. As Peter notes below, it’s intended to bridge the gap between the 101-level overviews and the PhD level discussions of the topic.

Now, onto Peter’s take…

For more from Peter, make sure to follow him on LinkedIn, where he breaks down the latest in the Medicare Advantage market.

In this article, I aim to explain how aspects of the inner workings of the Medicare Advantage Star Ratings program have led to recent big news stories (and major stock price swings). Much of it boils down to the methodology CMS uses to get from ~40 quality measure scores to a single 1 to 5 Star Rating for each plan. Then we will look ahead to what impact this will have on MA plans and possible reactions. I find most Stars content is written at either the 101 level or the PhD actuary level – my goal with this piece is to help bridge that gap.

Disclaimer: The views and opinions expressed in this article are solely my own and do not necessarily reflect the official policy or position of Signify Health, CVS Health, or any of its affiliates. My employment with Signify Health is acknowledged, but the content provided here is written in my personal capacity and is independent of my professional role.

Quick Star Ratings Primer

Brief History of the Stars Program

Back in 2006, the Centers for Medicare & Medicaid Services (CMS) created the MA Star Program. The idea was simple: reward plans that provide the best combination of care and service, and make it easier for consumers to evaluate the quality of their MA plan options. Over time, what started as a transparency initiative evolved into a major driving force for quality improvement and plan competitiveness. Since 2012, the Star Rating Program has been the basis of the MA Quality Bonus Program (QBP), tying major financial incentives to the results.¹ 

In CMS’s own words, the intent of the Stars Ratings System is:

How It Works 

We are all familiar with receiving letter grades in school. Somehow all the scores from quizzes, problem sets, and exams coalesced into a single letter grade at the end of the term. This analogy is helpful to understanding the Stars program, both mathematically and in terms of anxiety of waiting for the report card to show up in the mail. In fact, the letter grade analogy was employed by SCAN Health Plan’s attorneys in their successful lawsuit against CMS – more on this later.  

Before you dive into MA regulations, it is critical to understand the four-year cycles of the Stars Program: 

1. Measure Year (MY): The year during which the Measure data is collected. Think of it as the school year when performance on quizzes and exams is measured.³ This is sometimes referred to as the “Performance Year”, but I will use Measure Year (MY) through the rest of the piece.  

2. Reporting Year (RY): The year when CMS is processing all the data and calculating Star ratings based on the MY. CMS then shares the final Ratings in October of the Reporting Year. This is when you get an early preview of your report card and have a chance to challenge any of the scores or the scoring methodology before it is mailed home in October.  

3. Star Year (SY): The year the ratings based on the data are assigned. This is the year when you receive any special privileges for having Straight A’s based on your MY performance. 

4. Payment Year (PY): This is the year during which the financial implications of the Star Ratings are applied to the MA plan’s payments from CMS. If your parents increased your allowance for getting straight A’s, this is when you collect on that hard work. 

Unlike the academic year, the Stars program stretches this cycle over four years. As an example, the 2025 Star Ratings (SY2025) that were just announced are mostly based on data captured during 2023 (MY2023). The data was released in October of 2024 (RY2024). The financial impact of the SY2025 occurs in calendar year 2026, the Payment Year. (PY2026). This is depicted in the orange row of the diagram below. So there is a two-year lag between MY and SY, and a three-year lag between MY and PY. This lag is important to understanding what levers are available to plans during the Measure Year (e.g. performance improvement programs to increase scores) compared to when the ratings are released the October before the Star Year (e.g. appeal and sue based on the data and/or methodology).  

Next you need to understand the hierarchy from individual measures to that single Overall Star Rating. Imagine a pyramid:

1. At the foundation are the Measures that each capture performance on a specific outcome or process. Each measure has an assigned weight. Cut Points are used to map scores to whole Stars from one to five.

2. Measures are grouped into Domains like “Staying Healthy” or “Managing Chronic Conditions”.

3. The Summary level is the weighted average of the Measure Stars grouped into Part C or Part D.

4. The top of the pyramid is the single Overall Rating that drives the marketing and bonus payouts.

Star Ratings have significant financial, marketing, and enrollment ramifications

Stars are a big deal to Medicare Advantage plans across several key factors:

• Financial Impact: Here’s why every MA plan in America cares deeply about Stars: Money. According to MedPAC's 2024 report, the annual Quality Bonus Program (QBP) payments exceed $15 billion a year. Plans that score 4 Stars or higher get a 5% bonus from CMS, and their rebate percentage increases, ranging from 50% (for 3-Star plans) to 70% for those high-flying 4.5 Star and above plans. This means better benefits, lower premiums, and generally happier enrollees—but only if you’re scoring well.

• Marketing Advantage: Star Ratings are featured on nearly every marketing piece for MA plans, at least for those plans scoring well. Why would a member join a 3-Star plan when there is a 4-Star plan for roughly the same price? It has a direct impact on membership retention and new enrollment. See the screenshot of the Medicare.gov Plan Finder below. The Star rating is the first attribute displayed!

• Enrollment Boost: Speaking of new members, there's also the elusive 5-Star Special Enrollment Period. If a plan hits 5 Stars, it is allowed to enroll members year-round, not just during the usual windows. Q1Medicare.com estimates that 6MM MA beneficiaries live in counties covered by one of the 11 5-Star plans. Any of those members who enroll in another plan have the option during the year to switch to the 5-Star plan.

2025 Star Ratings demonstrate the increasing difficulty of achieving a 4-Star or higher rating

There is quantitative data that it is more difficult each year to achieve a 5-Star rating. As part of this year’s Star Ratings data release, CMS posted a table with trend by Star Rating since 2022. I created the chart below to show the decline in 5-Star contracts. Despite the number of contracts increasing slightly since 2022, the number of 5-Star contracts has decreased by 10x, from 74 to just 7. During the same period, the count of ≥ 4-Star contracts has decreased by 1.5x, from 322 to 209.

The lag between the Measure Year and the release of the cut points leads to big surprises

We’ve all been there. It’s the end of the academic term and the professor has announced several extra credit projects. You did well on the term paper but forgot to submit a couple homework assignments. The final exam was hard, but everyone said it was. Do you do the extra credit project or take it easy? You have a pretty good idea of your performance, but you have no idea where you are on the curve relative to your classmates.

Medicare Advantage plans know this scenario well. They generally have a good idea of how well they are scoring (e.g. 80%) on measures during the Measure Year, but how those scores convert to Star Ratings remains a mystery until September of the following year. At this point, the measurement period is long gone - there’s nothing you can do to change your performance. Plans have a couple opportunities during this period to appeal their data and scores, but in early October the Star Ratings are posted publicly to inform potential members during the Annual Enrollment Period (AEP).

The lag between the Measure Year and the release of the cut points creates an opportunity for big surprises. This year’s big news was four of Humana’s largest contracts dropping by a full Star. Those contracts, with ~4MM lives, represent the majority of Humana’s MA book.

The Morgan Stanley Research team estimates the overall decrease in Stars performance will decrease Humana 2026E EPS by ~$23 (during the Payme

nt Year!).⁴ At 120MM shares outstanding, that represents a maximum unmitigated financial impact of ~$2.8B. The Morgan Stanley report is worth a read, as they discuss their assumptions and model. This is a great example of the financial implications of Star Ratings on MA plans and the cost of low performance.

The impact was large enough that Humana filed an 8K on October 2, 2024 (emphasis mine) and has since sued CMS⁵ after their appeal was rejected:

Elevance provides us another financial impact data point in their lawsuit against CMS. They claim damages of “at least $375 million” from CMS’s methodology rounding contract H3655 down to a 3.5 Star instead of up to a 4 Star rating:

That particular contract has an enrollment of ~272k, so they are estimating a negative impact of ~$1,400 per member for a half-star decrease.

(PS - If you are an MA plan, you always do the extra credit if you still have budget!)

 The Root Cause: Discretization and Cut Points

The Star Ratings program grades on the curve, therefore CMS must define the curve based on the performance of all the Medicare Advantage plans. The methodology CMS uses for discretizing from numerical score data to Five Star values is a large contributor to the chaos and controversy surrounding the program. I’ll briefly describe how CMS determines the oh so important cut points that divide the winners from the losers. Then we’ll break down a few recent lawsuits that MA plans have filed against CMS specifically targeting the discretization methodology.

Discretization is how CMS converts the various measure scores into Star Ratings

Discretization is the process of converting a continuous variable or function into a discrete form. In statistics, it is widely used when dealing with continuous data (e.g. measure scores ranging from 0% to 100%) that needs to be represented in a finite, discrete set of values (e.g. One to Five Stars). We are all familiar with this from school, where exam scores (89%) were mapped to letter grades (B+).

There is no one right discretization method. CMS has stated that their preference is to maximize the separation of the clusters for each Star score.⁶ See the excerpt from CFR 42 that describes the methodology for all measures except those based on the CAHPS survey (emphasis mine):

Those goals and methods sound complicated but innocuous. Let’s see what it looks like in practice.

A deeper look at the Part D Measure cut points

The chart below shows the cut points for all Part D measures based on percentages where higher is better. For each row, you see how a percentage maps to a 1 to 5 Star score. Notice on D01 (bottom row), that a 5-Star score requires 100%! You can also see Measures like D09 (Medication Adherence for Hypertension) with an extremely narrow 4 Star cut point range. An 89% gets you a 3 Star, but a 92% earns 5 Stars. That is a triple-weighted measure, so those 3% have huge financial implications! Many plans would argue that is not a meaningful difference in performance, at least as a member would experience it.

The histogram below shows the distribution of the 2025 scores for the D01 contact center measure color-coded by the assigned Star score (i.e., within the cut points). You can see the clustering based on the color. But you can also see tall bars right next to the border between colors. These are contracts that scored just a hair below the cutoff, which dropped them by a whole Star. Unlike a high school exam, just barely missing the cutoff means a swing of tens or hundreds of millions of dollars for that MA contract.

The D01 histogram is also helpful in understanding how the Tukey outlier deletion methodology results in higher cut points. I will not get into the statistical techniques here, but you can clearly see the D01 scores skew high (i.e. most of the data is 80% or more). This means there are not any outliers on the upper end. In the 2025 Stars technical notes, CMS indicates the Lower Cutoff for D01 outliers is 74%. Any scores below that are “outliers” and eliminated before calculating the cut points. When the lower values of a distribution are removed, it naturally shifts the cut points upward.

Vince Lombardi said that football is a game of inches. One could say Stars are a game of minutes. A great example of the wild scenarios caused by discretization can be found in UnitedHealth’s recent lawsuit against CMS. UnitedHealth alleges that a single 10-minute contact center call caused several of their contracts to fall below the cut point on the 4x weighted D01 measure, resulting in a decrease in their overall Star Rating. They estimate the “collective harm to Plaintiffs from lost potential and actual customers” from those ten minutes “will be millions of dollars”!⁷ The treemap view below shows the D01 measure Star scores across all MAPD plans for SY2025. You can see that some of those contracts on the cusp between four and five stars in the histogram above have huge enrollments.

What can we expect going forward?

With Star Ratings determining the allocation of $15B+ each year, expect continued focus on this program across the industry. Incentives matter, and the industry is constantly adapting to changes. It is unfortunate that so much of the focus is on the technicalities of the Star Rating calculations, but that is the only major lever plans can pull once the scores are released. That said, I have personally worked with many health plan employees working every day to ensure members get the care and resources they need, which is the basis of the Star Measure numerical scores.

Here are some trends I expect to play out in the next few years:

MA regulations will continue to change

CMS has signaled that Stars will increase in difficulty and payouts will decrease over the next decade based on guidance and proposed rulemaking. Last October, Milliman put out a great whitepaper summarizing the Star Ratings system changes on the horizon and their estimated financial impact.⁸ Many of the changes are well-intentioned but will unfortunately result in additional layers of calculation complexity.

Here are some relevant, upcoming changes (though please check out the Milliman paper – I am omitting several big changes like the Health Equity Index (HEI) for brevity):

Measure Weight Adjustments:

• We discussed in the first section that the aggregate of single measure stars to Summary and Overall ratings is the weighted average of the measures. Recall, the weight as a multiplier for how much a measure impacts the overall average. The treemap below shows Humana’s big H5216 contract measure-level Stars sized by measure weight.

  
  

  

  2025 Star Rating breakdown for Humana’s H5216 contract

  
  You can see the Patient Experience, Complaints, and Access measures are all big boxes, as they have been quadruple-weighted measures. These measures will decrease from four to two in SY2026, a significant reduction that shifts the focus away from these experience measures. Instead, more emphasis will be placed on outcome measures and clinical process measures such as controlling blood pressure and diabetes care.

  
  What does this mean? The measure weight adjustments will determine the relative priority and investment plans place on improving performance at the measure level. You will probably also see many fewer lawsuits targeting measures like D01, where a single phone call for a quadruple-weighted measure can drag down the entire Star Rating for a plan by a half Star.

Cut Point Methodology changes:

• CMS will continue to tinker with the cut point methodology. In the 2024 proposed rulemaking, CMS considered removing the guardrails for non-CAHPS measures. These guardrails prevent drastic shifts in cut points year over year by capping the change.⁹ Removing the guardrails would introduce more volatility into the Star Ratings, as plans will no longer have the predictability and protection from large shifts in cut points driven by outlier performance during unusual years like the COVID-19 pandemic. As of this writing, there is no final rule or timeline to implement this change. Any additional tweaks to the discretization and cut point methodology increase complexity, Star rating volatility, and potential health plan litigation.

MA Plans will delegate more Stars performance via Value-based Contracts

• Given the weight decrease of the Member Experience and Complaints measures in MY2025, clinical process and outcomes measures will determine more of the overall Star rating going forward. Examples of this category of measures that will be triple-weighted in SY2026 include:

  • Diabetes Care - Blood Sugar Controlled

  • Controlling Blood Pressure

  • Plan All-Cause Readmissions

  • Medication Adherence

  
  Providers will have more direct ability to impact these measures than a health plan will. I would expect health plans to increase the incentives and penalties tied to Stars performance within their value-based contracts (VBC). This shift will likely be most extreme in VBC’s with providers like Oak Street Health, Centerwell, and Chenmed that are already taking capitated risk.

Increasing legal challenges to the CMS Star Rating methodology

• With billions of dollars and millions of potential enrollees on the line plans are resorting to suing CMS as their last option. Since the Star Ratings were released on October 10, 2024, multiple plans have sued CMS over various aspects of the Star Ratings program and how it is being administered. We mentioned the UnitedHealthcare lawsuit over a single ten-minute call that caused them to drop from 5 to 4 Stars on the D01 measure. Multiple other plans have sued for substantially similar scenarios related to D01, likely because Scan and Elevance’s lawsuits about call center measures were successful earlier in the year.
  

  Elevance’s new lawsuit (filed 2024-10-31) attacks CMS’s rules that round star scores up or down to the nearest half-point at the millionth decimal¹⁰: Specifically, CMS has established a complex system through sub-regulatory guidance that purports to calculate an MAO’s overall contract raw score to, and round at, the millionth (i.e., sixth) decimal to assign an overall Star Rating.

  
  Most of these lawsuits take issue with aspects of the discretization methodology! Somehow, CMS must convert all those measures and numbers into a single Star rating, and plans on the lower side of that cut point will not be happy.
  
  While lawsuits about Star ratings are not uncommon, this year may be different. The Elevance lawsuit quoted above may hint at another angle of attack given the recent Supreme Court Loper Bright Enterprises v. Raimondo decision that overturns what is known as the “Chevron Doctrine”. I am not an attorney, so this is not legal advice. Though CMS has significant statutory authority to manage the Medicare Advantage program, CMS likely has exposure given it publishes hundreds of pages of rules and regulations annually. I believe this is the “sub-regulatory guidance” to which the Elevance attorneys refer.
  
  CMS's statutory authority to regulate Medicare Advantage comes from a combination of federal statutes, dating back to the Social Security Act. The Balanced Budget Act of 1997 created the original Medicare+Choice program, which became Medicare Advantage. Other major laws shaping Medicare Advantage include the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 (MMA), the Affordable Care Act (ACA 2010), and the Medicare Advantage regulations in 42 CFR Part 422.
  
  The case Loper Bright Enterprises v. Raimondo directly challenges the Chevron deference doctrine, which has guided administrative law for over four decades. The primary legal question in Loper Bright was whether courts should defer to federal agencies' interpretations of ambiguous laws. Historically, under Chevron, courts deferred to agencies like CMS when statutes were unclear, provided the agency’s interpretation was reasonable. The Supreme Court's decision now removes this deference, meaning courts must independently interpret statutes without automatically siding with agencies. Courts may no longer defer to CMS’s expertise, particularly if the statutory language lacks explicit clarity on these issues. This could lead to more legal challenges from Medicare Advantage organizations disputing CMS’s regulatory decisions.

Citations

[1] “March 2024 Report to the Congress: Medicare Payment Policy – MedPAC,” accessed October 15, 2024, https://www.medpac.gov/document/march-2024-report-to-the-congress-medicare-payment-policy/. 

[2] “2025 Medicare Advantage and Part D Star Ratings,” CMS Newsroom, October 10, 2024, https://www.cms.gov/newsroom/fact-sheets/2025-medicare-advantage-and-part-d-star-ratings. 

[3] Technically, the CAHPS survey is actually sent at the start of the Reporting Year, but it is effectively surveying performance during the Measure Year.

[4] “Expanded Analysis of MA Star Ratings (Updated for Latest Enrollment Data)- HUM Stars Shortfall Largest of the Group; CVS Up y/y; UNH & ELV Stars Ebb,” Morgan Stanley Research, October 11, 2024.

[5] “Form 8-K. U.S. Securities and Exchange Commission,” Humana Inc. Investor Relations, October 2, 2024, https://humana.gcs-web.com/node/39041/html.

[6] 42 CFR § 423.186(a)(2) https://www.ecfr.gov/current/title-42/chapter-IV/subchapter-B/part-423/subpart-D/section-423.186

[7] UNITEDHEALTHCARE BENEFITS OF TEXAS, INC., UHC OF CALIFORNIA, CARE IMPROVEMENT PLUS SOUTH CENTRAL INSURANCE COMPANY, PREFERRED CARE PARTNERS, INC., UNITEDHEALTHCARE COMMUNITY PLAN, INC., PEOPLES HEALTH, INC., and SIERRA HEALTH AND LIFE INSURANCE COMPANY, INC. v. CENTERS FOR MEDICARE & MEDICAID SERVICES and U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES, No. 6:24-cv-00357-JDK (n.d.).

[8] Hayley Rogers, Matthew H Smith, and Mike Yurkovic, “Future of Medicare Star Ratings: The Reimagined CMS Bonus System,” Milliman, 2023.

[9] CAHPS is the Consumer Assessment of Healthcare Providers and Systems, a survey sent to a sample of plan members. It is the data source for measures like “Getting Needed Care”, “Care Coordination”, and “Annual Flu Vaccine”.