PDGM Functional Points for 2027: The 8 OASIS Items That Set Your Payment

Buried in the CY 2027 proposed rule is a line that generated a round of alarming headlines: CMS is changing the points assigned to the OASIS items that drive PDGM payment.

That is true. It is also routine, and the alarm is misplaced.

What CMS proposed is the annual recalibration it performs every year — re-deriving point values from newer claims data using the same methodology finalized back in CY 2019. It is not a redesign, the items are not changing, and nothing about how you document is being rewritten.

What is worth your attention is how few clinicians can name the eight items that decide their agency's case-mix, or explain how the scoring actually works. That gap costs real money every year, recalibration or not.

Source: CMS-1844-P (CY 2027 HH PPS proposed rule), 91 FR 41216, July 6, 2026. Current point values are from Table 8 of the CY 2026 HH PPS Final Rule (CMS-1828-F). OASIS item labels verified against the OASIS-E2 Guidance Manual.

The Eight Items

Your functional impairment level comes from exactly eight OASIS items — seven ADL items plus one risk item:

That is the entire list. Every other OASIS item you complete serves quality reporting, clinical care, or eligibility — but these eight, and only these eight, set the functional component of your payment.

How the Scoring Works

Each response gets a point value. The points are summed. The sum maps to a low, medium, or high functional level — and the thresholds differ by clinical group.

Here are the point values in effect for CY 2026:

Two things jump out of that table.

Bathing and ambulation dominate. M1830 at its top band is worth 17 points and M1860 at its top band is worth 20 — together nearly half the range. M1800 Grooming, by contrast, tops out at 3. If you have limited time to train clinicians on accuracy, bathing and ambulation are where the money is.

M1033 is all-or-nothing. It counts how many hospitalization risk factors are marked, excluding the "None of the above" and "Unknown" options. Three marked risks score zero. Four marked risks score 12 — the same as a fully dependent lower-body dressing response. Under-marking a legitimate fourth risk factor is one of the most expensive small omissions in OASIS.

The Quirk Nobody Talks About

Look at M1860 again:

• Response 2 → 5 points
• Response 3 → 1 point

A more impaired ambulation response scores fewer points than a less impaired one. That is not a typo on our end. CMS published it that way, and acknowledged it in the CY 2026 Final Rule, noting that the point value for response 2 is worth more than the point value for response 3.

It is a genuine artifact of the regression — the point values are derived statistically from what actually predicted resource use in the claims data, not from a clinical judgment about which response ought to score higher. When the data says patients at response 3 consumed fewer resources than patients at response 2, the table reflects that.

This is exactly what annual recalibration does, and it's the best argument for reading Table 19 of the proposed rule rather than assuming the values you memorized still hold.

Thresholds Vary By Clinical Group

The same point total produces different functional levels depending on the clinical group. CMS designed the bands so roughly one-third of periods in each clinical group land in each level.

A sample of the CY 2026 thresholds:

The spread matters. A 44-point patient is high functional impairment in Endocrine and only medium in Neuro Rehab. Same documentation, same points, different payment — because the comparison group is different.

The Rule Most Agencies Get Wrong

Your functional impairment level is locked for both the first and second 30-day periods of care.

The only thing that changes it mid-certification is a significant change in condition warranting an "other follow-up" assessment before the second 30-day period. Medicare's claims system looks for occurrence code 50 on the claim, matched to the M0090 date of the applicable assessment.

Two consequences worth internalizing:

1. The SOC assessment carries 60 days of payment weight, not 30. An hour of extra care on that assessment is the highest-return hour in your operation.
2. A genuine decline mid-certification is worth capturing properly. If a patient truly deteriorates, the other follow-up assessment exists for exactly that reason. Agencies routinely leave it unused and eat the difference.

What Changes for CY 2027

CMS proposed to update functional points and functional impairment levels using CY 2025 claims data, applying the same methodology previously finalized. The proposed point table appears as Table 19 in the rule, and the thresholds by clinical group as Table 20.

Expect the numbers to move once more before they're final. CMS says explicitly of the LUPA thresholds that they "will be updated based on more complete CY 2025 claims data in the final rule," and the functional points are derived from the same claims data on the same annual cycle. Treat anything in Table 19 as provisional until November.

So: the items stay the same. The methodology stays the same. The numbers move. That is the whole story, and it is meaningfully less dramatic than "CMS is changing how PDGM pays."

What To Do Monday

1. Print the eight items and put them somewhere clinicians see them. Most cannot list them from memory. That is a five-minute fix with an annual payoff.
2. Audit M1830 and M1860 specifically. They carry the largest point swings. Pull ten recent SOC assessments and check whether the documented response matches the narrative in the visit note.
3. Check your M1033 marking discipline. Look for assessments with exactly three risk factors marked. Some are correct. Some are a fourth legitimate risk nobody recorded, and each one is 12 points.
4. Find out whether your team knows the level locks for two periods. If they think it recalculates at day 31, they are under-documenting the SOC.
5. Stop treating GG items as payment items. They are quality items. Both matter — for different reasons.

Comment Before August 31

CMS explicitly solicited public comment on the proposed updates to functional points and functional impairment levels by clinical group.

If you have looked at Table 19 and something in it does not match what your clinicians see in the field — particularly another non-monotonic band — that is worth putting on the record. Comments close at 5 p.m. EDT on August 31, 2026, referencing file code CMS-1844-P, submitted through Regulations.gov.

The people who write these comments are overwhelmingly large chains and trade associations. A working clinician describing what a scoring band gets wrong at the bedside is rarer, and it reads differently.