The Score Is the Revenue: Why HEDIS Gaps Are Now a Bonus Problem

A Bonus That Turns on One Threshold

I want to start with the money, because in this case the money is unusually clear.

Under Medicare's quality bonus program, a Medicare Advantage plan rated 4 stars or higher receives a 5-percentage-point quality bonus applied to its benchmark. An established plan below 4 stars receives no quality bonus (new and low-enrollment contracts receive a 3.5-percentage-point bonus). In qualifying low-spending counties, that increase doubles to 10%. There is no partial credit at the threshold: an established plan at 3.5 stars gets nothing; the same plan at 4.0 gets the full 5%.

That is why the ratings CMS released on October 8, 2026 matter so much — and why the timing matters too. Those are the 2027 Star Ratings, and they set bonus payments for 2028, not for 2027. A rating released this fall is next year's bonus, decided by this year's record.

Here is what those ratings showed: 37% of Medicare Advantage prescription-drug contracts earned 4 or more stars, down from about 44–45% for 2026. Fifteen contracts earned 5 stars, down from 22. The enrollment-weighted average rating was 3.99 stars.

Read those two percentages together, and read them carefully, because they measure different things. Thirty-seven percent is contracts. About 71% of enrollees are in 4+-star contracts — contracts and enrollees are not the same thing, and a small number of large contracts can hold a large share of the people. Both numbers are true. They just answer different questions.

The scale of what turns on that threshold is large, and it is still growing:

KFF, "Medicare Will Spend More Than $13 Billion on the Medicare Advantage Quality Bonus Program in 2026," July 1, 2026, updated August 12, 2026.

And the dollars per enrollee are not uniform. In 2026, the average increase in payment per enrollee from the bonus program was $466 for employer/union-sponsored plans, $381 for individual plans, and $318 for special needs plans. Same program, same threshold — very different dollars depending on the plan type.

At the same time, the pool is getting harder to stay in. In 2026, 68% of Medicare Advantage enrollees were in bonus-eligible plans — the lowest share since 2018 — down from 75% in 2025, because CMS raised the performance thresholds. The bar went up, and fewer plans cleared it.

So this is not a story about a quality score sitting in a report somewhere. It is a story about a score that decides bonus dollars.

Don't Confuse the Two Scoreboards

Before going further, one distinction, because confusing these two systems is the easiest way to misread the whole story.

The Healthcare Effectiveness Data and Information Set — HEDIS, maintained by NCQA — covers more than 90 measures across 6 domains of care: Effectiveness of Care; Access/Availability of Care; Experience of Care; Utilization and Risk Adjusted Utilization; Health Plan Descriptive Information; and Measures Reported Using Electronic Clinical Data Systems (ECDS). As of fall 2024, approximately 236 million people were enrolled in health plans that reported HEDIS.

NCQA's own Health Plan Ratings run 0–5 stars, to the nearest half-star, as a weighted average of a plan's HEDIS, CAHPS, and HOS measure ratings, plus bonus points if the plan is NCQA-accredited. In NCQA's 2026 ratings, based primarily on 2025 performance, 18 plans earned 5 stars — a 64% increase from 11 plans in 2025 — and 53 plans earned 4.5 stars.

Those NCQA ratings are not the CMS Star Ratings. Different scales, different schedules. The money story in this article is the CMS system. NCQA is the measure-level evidence underneath it — the record of what care was delivered, documented, and counted.

Two Different Problems, One Score

Here is the part I want health system leaders to hold onto, because most conversations about gaps in care blur it: HEDIS punishes two different failures, and they need two different fixes.

1. Care that didn't happen

The first problem is straightforward. HEDIS counts members who received the care. Each open gap lowers the measure rate, because a missed screening, a missed follow-up, or an unclosed referral loop sits in the denominator without reaching the numerator.

The clinical stakes here are not abstract. A Harris Poll survey of 250 oncologists conducted for Quest Diagnostics found 79% had patients whose cancer recurrence was missed while still in the early disease stage, and 68% of those oncologists attributed it to patients missing, skipping, or delaying follow-up care and monitoring appointments. A missed follow-up is a quality gap and a human one at the same time.

And the system is starting to score follow-up it previously could not see at all. NCQA added Follow-Up After Positive Colorectal Cancer Non-Invasive Screening Test (COF-E) to HEDIS for measurement year 2027. Follow-up colonoscopy rates after a positive stool-based test range from 24% to 75%, per NCQA. That is a wide range, and it is now a scored measure — the follow-up after the screening counts, not just the screening itself.

The most recent NCQA ratings point the same way. Follow-up and continuity measures led the gains: Follow-Up After Hospitalization for Mental Illness rose 8.9 percentage points in commercial plans, 9.0 in Medicaid, and 11.9 in Medicare Advantage, while Kidney Health Evaluation for Patients With Diabetes rose 7.3 points in Medicare Advantage. Twenty-two of 28 chronic-disease indicators improved year over year, and average scores rose 1.6–1.8 points. Closing follow-up gaps is where scores are moving right now.

2. Care that happened but can't be proven from the record

The second problem is quieter, and in my experience it is the one most leaders underestimate. HEDIS only counts what the data shows. Care that was delivered but never documented — or documented in a way claims and EHR data cannot capture — does not count.

"If the clinical evidence isn't in your data, then as far as your score is concerned, the care didn't happen."

That is not my line. It is how 1upHealth put it, and it is the cleanest statement of the record-completeness problem I have seen. HEDIS rewards documented, structured data, not just care delivered: a missing lab code, a blood-pressure reading in the wrong EHR field, a result buried in an unstructured document — any of those can leave a patient who received care listed as an open gap.

This problem is about to get worse, because the safety net is being retired. NCQA is phasing out hybrid HEDIS reporting — the year-end chart review that used to rescue scores after the fact — by calendar year 2030. The size of that safety net is measurable: hybrid colorectal screening rates ran about 10% higher than purely electronic clinical data system rates for Medicare Advantage plans. In other words, chart review used to find care the claims and EHR data missed. Once it is gone, a gap in claims or EHR data is a gap in the score, with no year-end paper chase to fix it.

Keep those two limbs separate in your head, because they fail differently. One is care that didn't happen. The other is care that happened and left no usable trace. A plan, a practice, or a health system can be losing points on both at once — and fixing only one leaves the other untouched.

Why the Weight Is Only Growing

If this were a shrinking part of the rating, it would still matter. It is a growing one.

Going forward, HEDIS measures will account for approximately 30% of Star Ratings weighting as CMS removes 11 measures, including the administrative domain. Under CMS's 2029 restructuring, HEDIS and pharmacy measures will represent about 48% of a contract's rating, with CAHPS and HOS surveys at about 36%, improvement measures at about 16%, and operational measures at 0%, down from 21%. Clinical outcomes, including HEDIS and medication measures, will account for approximately half of rating points by 2029.

Weight is also concentrated inside HEDIS itself. CMS weights outcome measures at 3x: Controlling High Blood Pressure, Glycemic Status Assessment, and Plan All-Cause Readmissions. Patient-experience measures are weighted 2x in 2026, down from 4x; process measures are weighted 1x. A small set of outcome measures therefore moves a Star Rating disproportionately — and those measures are exactly the ones that depend on complete blood-pressure, diabetes, and readmission records: care delivered, documented, and traceable.

The measures that move when records are complete are not obscure. They are the core of Effectiveness of Care and follow-up: blood-pressure control, diabetes care including HbA1c control, eye exams, kidney evaluation and blood-pressure control, breast and colorectal cancer screenings and the new colorectal follow-up companion, follow-up after hospitalization for mental illness, and medication adherence for diabetes, hypertension, and cholesterol.

What the Market Just Told Us

Humana's 2027 ratings are the cleanest published example of what crossing the threshold is worth. Ninety-five percent of Humana's Medicare Advantage members are in plans rated 4 or more stars for 2027, up from 20% for 2026. Eighteen contracts sit at 4 or more stars — 12 at four stars, 6 at 4.5 stars — and its member-weighted average rating is 4.17 for bonus year 2028, versus 3.84 for bonus year 2027.

Humana's own October 9, 2026 SEC filing estimates its ratings are worth about $160 per member per month in Stars revenue for 2028, versus a peer median of about $105 and its own top-quartile goal of about $115 — under standardized assumptions, with Humana itself noting actual 2028 figures will differ. Outside estimates put the scale in context: Evercore ISI estimated about $4.8 billion in 2028 bonus payments, and S&P Global Ratings estimated about a $3.6 billion year-over-year payment increase. Those last two are outside estimates, not Humana figures, and I am labeling them as such.

The lesson is not that every organization can replicate a turnaround like that in one cycle. The lesson is that the distance between 3.5 stars and 4 stars is not a rounding error. It is the difference between no bonus and the full bonus, between a rating that funds improvement and one that does not.

Methods to Determine Gaps in Care

The way to attack both problems is to get better at determining where the gaps actually are — and that starts with more accurate, more complete electronic records.

One method: bring a patient's clinical history together from the different systems and settings where it lives into one more complete record, and have it reviewed before it is relied on. A record assembled that way does two things the fragmented version can't. First, it surfaces gaps so they can be closed: the follow-up that never got scheduled, the screening that never got its follow-up, the referral loop that never closed. Second, it documents care so it counts: care already delivered, captured where it can be measured, so it reaches the numerator instead of sitting invisible in a chart somewhere.

Methods like this don't replace a quality team. They hand that team a more complete record to work from, instead of a chart chase.

I am not telling you where your scores stand — I cannot assess that, and no article should pretend otherwise. What I am saying is narrower, and I think it is stronger for being narrower: if gaps in your records are holding your scores back, the record is the place to start, because the score is now the revenue.