Calling All Plans: Advance Notice Released For 2026 MA and Part D Rates And Other Policies

Some relief with positive hike, but rate increase still inadequate

On Friday, the outgoing Biden administration released its Advance Notice for Medicare Advantage (MA) and Part D rates and other policy changes for 2026. I have reviewed the 180-page Advance Notice, CMS Fact Sheet, and CMS Press Release. Below are the key highlights. The advance notice will be finalized by early April.

I will publish a comprehensive Stars roadmap blog on Thursday in conjunction with Lilac Software. This will detail all the proposed Star measure changes, updates, and information discussed in the advance notice.

MA rate proposal

It is likely that MA plans and investors will see the proposed rates for 2026 with some relief. After two years, it appears MA plans could see a real revenue increase rather than the decreases they saw in 2024 and 2025. But as plans and financial analysts take a deeper look, the proposal will likely be seen as a very skimpy increase given all the MA world is facing.

In my mind, the net revenue increase is so low that we will be in store for a third year of MA benefit cuts in 2026. MA rates will have gone up just 0.95% over 3 years when the industry is battling a return of utilization, inflation, and increased costs due to new laws and regulations. Benefit reductions and possible contraction of geographic footprints in some form will be needed by the big national plans in order to right their financial ships and meet long-term margin returns demanded by investors.

Let’s break the proposed MA rate hike down.

As you can see in the table below, the proposed rate hike for MA in 2026 is a 4.33% average increase if you use the Centers for Medicare and Medicaid Services (CMS) view or a 2.23% average increase if you use the plans’ view of things. What do I mean by this?

Each year, CMS computes the expected risk score trend into what it views as the overall rate hike. But plans do not see it that way and the difference is really tied up in the whole overpayment and risk adjustment debate. Plans view the risk score trend from year-to-year as representing increased risk of the population. They believe these increased risk adjustment revenues will go out in claims costs. CMS and some detractors of MA see the annual risk score trend as more evidence of overpayments and the risk score trends represent a real revenue increase to plans.

I stand somewhere in the middle. Perhaps some of the trends in risk scores could be part of overpayments that need reform. On the other hand, the margins that MA plans have seen of late would show that risk and costs are increasing robustly. I would also note that the projected risk score trend increase is slated to be up just over 2% for 2026, which is well below the over 4% trend on average for 2024 and 2025.  This would seem to bolster the plans’ position (at least for 2026).

Let’s break down the component parts of rate-setting that derive the 4.3% (CMS) or 2.23% (Plan) views of the 2026 MA increase:

• The effective growth rate (EGR) is derived from the growth in the per capita costs in the traditional fee-for-service (FFS) program and MA for the non-end-state-renal disease (ESRD) populations. It is a national estimate.  Trends can be different in each county and region. Remember that the eventual county benchmarks are adjusted from there based on several parameters, including whether a county is at 95% up to 115% of FFS costs. For 2026, we saw a huge increase in the EGR of 5.93% compared with just over 2% in 2024 and 2025. Utilization and other factors lag a bit in rate-setting. The 2024 and 2025 EGRs had some of the depressed COVID utilization years in them.  Now we are seeing more robust utilization and trends in the EGR calculation.  This is likely good news for the future as well. Note, though, that a medical education formula change in rate-setting actually reduced the growth of the EGR (by about 1.4% for the FFS component and 0.8% for the MA component.) Otherwise, the EGR would have grown much more.
• Rebasing relates to changes in the average geographic adjustment index (GAI), which adjusts for spending differences across regions/counties. This usually is a small impact. The rebasing could be positive or negative in a given year and actual impacts are really seen at the local county or regional level. The rebasing number simply reflects the nationwide impacts of changes in the GAI. The GAI will be set in the final announcement in April.
• Star revenue changes are a global estimate as well. Each plan will have a different result based on its actual Star performance by contract and eligibility for bonus revenue. The reduction due to Star performance is -0.69% for 2026 – the third year in a row of a decline.  This is based on the 2025 Star ratings as there is a one-year lag between the rating year and payment year. Those losing Star ratings may see a much bigger year-over-year reduction in revenue. Those maintaining Star could see additional revenue due to enrollment growth. Those gaining bonus revenue for the first time or increasing their ratings would see a great jump. The negative number for Star revenue is a sign that program-wide plans lost Star power and 4 Star or above Star ratings of late.
• The coding pattern adjustment is a statutory/regulatory reduction taken to MA rates for the perceived differences in risk adjustment scores between MA and FFS. This gets to the controversy about whether MA plans over-score members. You know my thoughts on this – it is a bit of a mixed bag in my opinion. The coding intensity adjustment has been set at negative 5.91% for years now and CMS will not change it for 2026. This is despite pressures from critics who say over-reimbursement has risen well beyond that negative factor. Further, CMS likely did not change it again due to the risk model changes being implemented (see below).
• There is the normalization factor adjustment captured within the Risk Model Revision and Normalization category. The normalization factor ensures the average risk score stays at 1.0 and adjusts from the data year in the model to the payment year. This is different than the coding pattern adjustment.
• The risk model change is responsible for the biggest impact in the Risk Model Revision and Normalization category. In 2024, CMS announced it was updating the risk model and made many major changes, including:
  • Condition categories are restructured from ICD-9 to ICD-10.
  • The underlying FFS data years are revised.
  • There are some revisions focused on reducing the sensitivity of the model to conditions with more coding variation. This has generated some controversy.
  • The new model should reflect more current costs associated with various diseases, conditions, and demographic characteristics.

In a concession to plans in 2024, CMS decided to phase in the new risk model change by using one-third in 2024, two-thirds in 2025, and 100% in 2026. CMS is not backing off the final phase-in for 2026. 

The risk model change certainly hurts. The three-year impact from the model change combined with normalization (they are intertwined) will be a negative 7.62% when fully phased in as of 2026. This is the biggest issue plans have – the risk model change, along with poor Star performance, has basically zeroed out the effective growth rate increase from 2024 to 2026. On the positive side, the amount reported as overpayments should decline as well.

MA plans had hoped that CMS might suspend the last year of phase-in of the risk model change. But CMS said suspending the final year of the risk model phase-in as well as one related to medical education formula changes would have added $10.4 billion to Medicare costs. The increase as outlined by CMS alone will cost about $21 billion in 2026. Further, CMS estimates the federal government will spend $9.2 trillion over the next decade on MA payments to plans, including $1.3 trillion alone on MA supplemental benefits and premium buy-down.

• The anticipated risk score trend is a national number calculated by CMS. Individual plans will see various increases or decreases. I noted above that plans and CMS view this trend very differently and why plans do not compute it into the rate hike year-to-year. Note that the projected risk score trend for 2026 is about half of what it was in 2024 and 2025 annually. This is a sign that the new model could be changing risk scoring patterns.

In the next table, I go further on the troubles MA plans have seen and will see from 2024 through 2026. The industry came off of high annual rate hikes early in the 2020s. In 2024 and 2025, these hikes were basically zeroed out due to the risk model change and low effective growth rates. Compounding the problem was a major drop in Stars starting with 2023 ratings and running so far through 2025 ratings. On the cost side, utilization and other trends (such as drug costs) leapt coming out of COVID.  As well, CMS has placed new costly burdens on plans, including requiring them to follow FFS prior authorization processes (I say taking the managed care out of managed care) and new unpaid costs from the Inflation Reduction Act’s (IRA) Part D cost-sharing changes. All this created a perfect financial storm for MA plans.

Future caution

Another potential caution to MA plans: CMS says it has been working to calibrate the risk adjustment model using MA encounter data (diagnosis, cost, and data submitted to CMS by MA plans). CMS says it could start phasing in an encounter data-based risk adjustment model as early as 2027. The issue for plans is it would break with setting rates in part on what have been traditionally very high FFS program cost trends.

Other financial factors

Other financial changes include:

• Benchmark rate percentages against FFS remain the same at 95% to 115% of FFS in the county.
• The Star bonus remains at the same 5% level and applies only to those with 4 Star ratings or greater.
• The rate rebate percentage — the difference between the plan base bid and the county benchmark which goes back to members via added benefits — remains the same based on Star performance:
  • Less than 3.5 Star = 50%
  • 3.5 and 4 Star = 65%
  • 4.5 and 5 Star = 70%
• No changes to Star calculations for low-enrollment or new plans.
• Part D risk corridors remain the same.
• Given the low rates in Puerto Rico, poverty, and penetration in MA, CMS would continue basing the MA county rates in Puerto Rico on the relatively higher costs of individuals in FFS who have both Medicare Parts A and B and applying an adjustment regarding the propensity of individuals with zero claims. 
• CMS is proposing to begin a transition in calculation of risk scores for PACE organizations. It will begin in 2026 with a blend of 10% of the risk score calculated using the 2024 CMS-HCC model and 90% of the risk score calculated using the 2017 CMS-HCC model. It outlines the full transition timeframe in the advanced notice.
• The advance notice notes the changes to Part D cost-sharing in the IRA and their continuing impact on rate-setting. Of note is that the first ten drugs in Part D will have negotiated prices going into effect in 2026.
• CMS says it will continue the special premium stabilization program it set up for free-standing Part D (PDP) plans. I think this is an extra-legal program by CMS and was done in the heat of the election season because the agency and Democrats in Congress did not anticipate the huge unfunded spending in the IRA levied on MA and PDP plans. Without the special program, premiums would have soared.
• CMS has posted Part D plan-level risk scores on HPMS. The Part D risk scores provided were calculated using the risk adjustment model currently in place for payment year 2025 (i.e., the 2025 RxHCC model) as well as the proposed and alternate RxHCC models described in the advance notice. Risk scores calculated with these models use diagnoses from encounter data and FFS claims filtered with the HCPCS-based filtering logic. Risk scores and accompanying technical notes can be found in the risk adjustment module of HPMS under the heading “Proposed Risk Scores.” Select “Proposed PY 2026 Part D Model Risk Scores” from the drop-down menu and enter a Contract ID to access information related to a specific contract number. The risk scores are payment year 2023 risk scores, calculated with diagnoses submitted for 2022 dates of service.

Could Trump come to the rescue on rates

As I said, some speculated that the risk model might be paused for some time due to the financial issues the MA plans are facing and the erosion of benefits and plan footprints.  The outgoing administration cites various aggregate statistics it says shows the MA program is stable and why the suspension is not needed. But CMS ignores underlying changes, including the huge displacement suffered by enrollees during the 2025 enrollment season. As many as 2 million people had to find new plans due to plan changes or terminations caused by the low rates and model change. This compares with 100,000 people on average each year.

Some hope Donald Trump will come in and make some positive rate changes before the advance notice is finalized in April. While many cite the need to pause the risk model phase-in, the medical education change phase-in actually has a bigger impact. Pausing the risk adjustment model phase-in would result in $3.4 billion in additional payments to MA plans in 2026. Pausing the adjustment to growth rates related to medical education costs would result in additional payments of $7.0 billion to MA plans in 2026.

Could Trump concede one or both of these in the final announcement of rates? Perhaps. But the incoming administration is busy searching for major spending reductions for extending the 2017 tax cuts. 

NOTE: in Thursday’s blog, I will go through all of the Star changes in the advance notice.

#medicareadvantage #partd #pdp #healthplans #rates

— Marc S. Ryan