Humana will exit Medicare Advantage plans covering about 600,000 members in 2027 — “mostly plans rated 3.5 stars or lower.” CMS publishes the rating and the enrollment for every contract in the country. Here they all are.
Across all 881 Medicare Advantage contracts CMS reported enrollment for in June 2026, 11,165,479 people sit in a contract rated 3.5 stars or lower under the 2026 Star Ratings — the ratings that set 2027 bonus payments and, per Becker's, the ones the exits track. 4,233,892 of them are Humana's, spread across 25 contracts. That is 59.6% of Humana's entire 7.10M Medicare Advantage book.
Against that, 600,000 is 14.2% of the low-star book and 8.4% of the whole book. Humana's single largest contract — H5216, 2,373,762 members across 2,501 counties, rated 3.5 — is by itself four times the size of the entire announced exit. And the company expects to recapture just over 40% of affected members into its own other plans, so the number of people actually landing somewhere else is closer to 360,000.
Every member a payer sheds is revenue somebody's finance office already planned around. But the exposure isn't the 600,000 — it's the 4.23 million still sitting in contracts that are, by the company's own framing, the ones on the margin-recovery list. This year's announcement is a trim, not a retreat.
610 contracts with 1,000+ members (99.7% of all MA enrollment). Each dot is one contract; area is enrollment. The shaded band is at or below 3.5 stars. Filter by parent organization, drag the size floor, and switch the horizontal axis to stress-test whether the pattern survives.
Leave the axis on star rating, org filter on All, and drag the size floor. At 1,000 members the correlation between a contract's star rating and the log of its enrollment is r = +0.35 across 486 rated contracts — “bigger plans rate better.” Push the floor to 50,000 and it falls to +0.22 on 133 contracts. Most of that headline relationship is carried by contracts too small to matter to anybody's revenue model. Then switch the axis to counties served and watch what happens to a variable that looks like it should predict quality and mostly predicts geography.
Of the 881 MA contracts with June 2026 enrollment, 378 carry no overall star rating — CMS records them as “plan too new to be measured” or “not enough data available.” They are easy to ignore because they hold only 1,190,059 members, 3.3% of the total. But they are 42.9% of the contracts, and every ranking, average, and “X% of plans are 4 stars or better” claim you read silently picks a side on what to do with them. Toggle Include unrated above and watch the readouts move.
Becker's reports that “twenty percent of the insurer's members are in plans rated 4 stars or higher for 2026.” Weighting the published CMS files by June 2026 enrollment at the contract level, we get 38.6% (2,744,370 of 7,103,935) under the 2026 Star Ratings, and 43.0% under the 2025 ratings. Neither is 20%.
That gap is not evidence anybody is wrong. It is evidence that “percent of members in 4-star plans” is at least four different metrics depending on whether you count contracts or individual plan benefit packages, which rating year you mean, whether group and Puerto Rico business is in, and whether the membership base is today's or next year's projection. If you cannot rebuild a company's headline metric from the public file, you do not yet know what the metric is measuring — and neither does the board deck that quotes it.
Pull the CPSC enrollment file for your service area, join it to the 2026 star ratings, and compute one number: the share of your Medicare Advantage volume sitting in contracts rated 3.5 or lower. That is your exposure to the next round of margin-recovery exits, and it is a fifteen-minute join. Then do it by county, because the answer moves a long way — in Montana, 57% of every MA member in the state is in a Humana contract rated 3.5 or below.
Contract counts and members, June 2026 enrollment against 2026 star ratings. Kaiser has no members in a contract rated below 4 stars. Humana has 4.23 million.
| Parent org | ≤3.5 members | ≥4.0 members | Unrated | ≤3.5 share |
|---|---|---|---|---|
| Humana | 4,233,892 | 2,744,370 | 125,673 | 59.6% |
| All other orgs | 2,935,449 | 6,967,445 | 609,216 | 27.9% |
| UnitedHealth | 1,842,323 | 7,193,746 | 187,526 | 20.0% |
| Elevance | 761,395 | 991,295 | 120,037 | 40.6% |
| CVS / Aetna | 728,306 | 3,310,147 | 47,264 | 17.8% |
| Centene | 664,114 | 153,212 | 97,623 | 72.6% |
| Kaiser | 0 | 2,050,697 | 2,720 | 0.0% |
Centene's book is 72.6% low-star, but it is 815,000 members total — a fifth of Humana's low-star exposure. The question a health system should be asking isn't who rates worst. It's who has enough low-star volume that trimming it is a strategy.