Total Performance Score: How Hospital VBP Turns Four Domains Into a Payment Adjustment

A hospital’s Total Performance Score (TPS) is the single number that decides whether Medicare’s Hospital Value-Based Purchasing (VBP) Program pays it back more, less, or exactly what it withheld. It is not a rating and not a star. It is the output of a defined arithmetic chain: measure points, then domain scores, then weighted domain scores, then a linear exchange function that converts the score into a percentage of every base operating DRG payment for the year.

This page works that arithmetic end to end, with CMS’s own published numbers where they exist. The programme is authorised by section 1886(o) of the Social Security Act and codified at 42 CFR 412.160 through 412.168.

Date this page, because the weights are not permanent

Verified against the eCFR text of 42 CFR part 412 subpart I and against the FY 2027 IPPS/LTCH PPS final rule (91 FR 49570, 4 August 2026) as of 26 August 2026. Everything below describes the FY 2027 programme year unless a different year is named.

The domain weights and the measure set are not in the regulation. 42 CFR 412.165(b)(4) says only that CMS “will weight the domain scores with the finalized domain weights for each fiscal year.” The weights, the measures, the baseline and performance periods and the performance standards are all outputs of the annual IPPS rulemaking cycle, which means they can and do change from one programme year to the next. Do not treat any one year’s weights as a permanent feature of the programme — check the IPPS final rule for the fiscal year you are actually being scored under before you rely on a number here.

First, which programme is this? VBP is not HRRP and not HAC Reduction

Three separate CMS payment programmes sit in the same corner of the IPPS and use overlapping data. Hospital staff conflate them constantly, and the consequences of doing so are real: they have different statutory bases, different arithmetic, different directions and different appeal rights.

	Hospital VBP	HRRP	HAC Reduction
Statute	§1886(o)	§1886(q)	§1886(p)
Codified at	42 CFR 412.160–412.168	42 CFR 412.150–412.154	42 CFR 412.170–412.172
Direction	Up or down — a withhold you can earn back, exceed, or lose	Down only	Down only
Score	Total Performance Score, 0–100	Excess readmission ratio per condition, floored at 1.0	Total HAC Score, worst quartile penalised
Money	2.0% of base operating DRG withheld and redistributed	Up to 3% reduction	1% reduction for the worst-performing 25%

Two consequences worth committing to memory:

• A readmission never touches your TPS. 42 CFR 412.164(a) directs CMS to select VBP measures “other than measures of readmissions.” Readmissions are HRRP’s business, and HRRP’s arithmetic is asymmetric in a way VBP’s is not — see the mechanics and penalty arithmetic of the Hospital Readmissions Reduction Program, where the excess readmission ratio is defined as “not less than 1.0” and outperformance earns nothing.
• The same NHSN infection data feeds two programmes with different maths. CLABSI, CAUTI, colon and abdominal hysterectomy SSI, MRSA bacteraemia and C. difficile infection all sit in the VBP Safety domain and in the HAC Reduction Program. One surveillance decision by an infection preventionist — whether an event meets the NHSN LCBI criteria for a CLABSI or the SUTI criteria for a CAUTI — therefore propagates into two independent payment calculations. The standardized infection ratio is the common currency.

Also distinct: the Overall Hospital Quality Star Rating, codified separately at 42 CFR 412.190. It is a public-reporting construct, not a payment one, and it is not derived from the TPS.

Step 1 — measure points: two ways to score, only one of them counts

Every scored measure is worth 0 to 10 points. CMS calculates the points twice, by two different routes, and keeps whichever is higher.

The three reference points

42 CFR 412.160 defines all three:

• Achievement threshold — the median (50th percentile) of hospital performance on the measure during the baseline period. Everyone is measured against the same national line.
• Benchmark — the arithmetic mean of the top decile of hospital performance during the baseline period. Hitting it is worth the full 10 points; there is nothing above it.
• Improvement threshold — your own performance on that measure during the baseline period. This is the only one of the three that is hospital-specific.

The Efficiency and Cost Reduction exception. For measures in that domain, both the achievement threshold and the benchmark are calculated from the performance period, not the baseline period (42 CFR 412.160). You are being ranked against your peers’ contemporaneous performance rather than against a line fixed in advance, which is why CMS states in the annual rule that it cannot publish numerical performance standards for MSPB Hospital ahead of time.

The two formulas

Both were finalised in the original Hospital Inpatient VBP final rule (76 FR 26490, 6 May 2011, at 26518–26519) and have not changed:

Achievement points = [9 × ((performance period score − achievement threshold) ÷ (benchmark − achievement threshold))] + 0.5, rounded to the nearest whole number.

Improvement points = [10 × ((performance period score − your baseline score) ÷ (benchmark − your baseline score))] − 0.5, rounded to the nearest whole number.

The scoring rules that bound them, per 42 CFR 412.165(a):

• Performance at or above the benchmark: 10 points, full stop. Improvement is irrelevant.
• Performance at or above the achievement threshold but below the benchmark: 1 to 9 achievement points.
• Performance below the achievement threshold: 0 achievement points.
• Performance above your own baseline but below the benchmark: 0 to 9 improvement points.
• Performance at or below your own baseline: 0 improvement points.

Note the asymmetry in the two ranges. Improvement caps at 9 and achievement caps at 10, so the tenth point is only ever available by reaching the top-decile benchmark. Note also the ±0.5 terms: the achievement formula adds half a point and the improvement formula subtracts one, which is what makes the achievement scale effectively run 0.5 to 9.5 and lets a hospital sitting exactly on the achievement threshold round up to 1 rather than sit at 0.

The higher score wins — and that is a codified, appealable fact

You do not choose. CMS takes the larger of the two automatically, and 42 CFR 412.167(a)(3) makes “whether CMS properly used the higher of the achievement/improvement points in calculating the hospital’s measure/dimension score” one of the ten enumerated grounds on which a hospital may appeal. The rule is not merely policy; it is a specific reviewable step.

The practical implication is the one most hospitals miss: a hospital that is far below the national median can still score well. If your baseline was poor, the improvement denominator (benchmark minus your baseline) is wide, and a large absolute gain converts into a large number of improvement points even while your achievement points stay near zero. The reverse is also true and less pleasant — a hospital already performing near the benchmark has almost no improvement room, so its score is entirely hostage to achievement.

CMS’s own worked example

These are not illustrative numbers; they are the figures CMS published at 76 FR 26520 for the hospital it labelled “Hospital I” on a fibrinolytic therapy measure.

Input	Value
Hospital’s baseline period score	0.4297
Hospital’s performance period score	0.8163
Achievement threshold (national median)	0.6548
Benchmark (mean of top decile)	0.9191

Achievement: [9 × ((0.8163 − 0.6548) ÷ (0.9191 − 0.6548))] + 0.5 = 5.5 + 0.5 = 6 points.

Improvement: [10 × ((0.8163 − 0.4297) ÷ (0.9191 − 0.4297))] − 0.5 = 7.9 − 0.5 = 7.4, rounded to 7 points.

Measure score: 7 points, because the higher of the two is used. A hospital that read only its achievement score would have understated this measure by a full point — and, as the domain arithmetic below shows, that is a real amount of money.

Minimum case counts, or the measure is not scored at all

A measure with too few cases is dropped rather than scored badly. Table 1 to 42 CFR 412.165(a)(1) sets the floors:

• HCAHPS: at least 100 completed surveys.
• Mortality and complication measures (MORT-30-AMI, MORT-30-HF, MORT-30-PN, MORT-30-COPD, MORT-30-CABG, COMP-HIP-KNEE): at least 25 cases each.
• The five NHSN healthcare-associated infection measures (CAUTI, CLABSI, colon and abdominal hysterectomy SSI, MRSA bacteraemia, CDI): at least 1.000 predicted infections as calculated by the CDC — a predicted-events floor, not a case count, which is why a small hospital can fall out of an HAI measure while still reporting events.
• SEP-1: at least 25 cases.
• MSPB Hospital: at least 25 cases.

Step 2 — domain scores: sum, normalise, weight

42 CFR 412.165(b) sets out four operations, in order:

1. CMS calculates a domain score only where the hospital reported the minimum number of measures in that domain.
2. Sum all points awarded for each measure in the domain → the unweighted domain score.
3. Normalise that score so it is expressed as a percentage of points earned out of 100.
4. Apply the finalised domain weights for the fiscal year, and sum — the TPS is the sum of the weighted domain scores, up to a maximum score of 100.

The four domains are Clinical Outcomes, Person and Community Engagement, Safety, and Efficiency and Cost Reduction. CMS has weighted them equally — in its own words, “Each domain contributes 25 percent to the total score” — continuously since FY 2020. Again: that equality lives in the annual rule, not the CFR. Confirm it for your programme year.

The normalisation step is where the leverage is, and almost nobody reads it

Normalising to a percentage means a domain’s weight is spread across however many measures happen to be in it. The domains are not the same size:

• Clinical Outcomes carries six measures → 60 possible points → one measure point is worth 25 ÷ 60 = 0.4167 TPS points.
• Safety carries six measures → 60 possible points → also 0.4167 TPS points per measure point.
• Person and Community Engagement is scored out of 100 (an 0–80 HCAHPS base score plus 0–20 consistency points) → 0.25 TPS points per point.
• Efficiency and Cost Reduction carries one measure, MSPB Hospital → 10 possible points → one measure point is worth 25 ÷ 10 = 2.5 TPS points.

A single point on MSPB Hospital is worth six times a single point on a mortality measure, and ten times a single HCAHPS point. That fact falls directly out of the arithmetic in 412.165(b)(2)–(4) and it is the most consequential thing on this page for anyone deciding where to put improvement effort.

HCAHPS scoring, and what changed for FY 2027

The Person and Community Engagement domain is scored differently from the others. Achievement and improvement points are calculated per HCAHPS dimension (the larger of the two kept, as everywhere else) and summed into a base score, to which consistency points are added. Consistency points are driven entirely by a hospital’s single worst dimension:

lowest dimension score = (performance period score − floor) ÷ (achievement threshold − floor), and HCAHPS consistency points = (20 × lowest dimension score) − 0.5, rounded, bounded at 0 and 20 (76 FR 26519–26520). If any dimension is at or below the floor — the worst-performing hospital’s baseline score on that dimension — consistency points are zero. Twenty points, a fifth of the domain, hangs on your weakest dimension alone.

For the FY 2027 through FY 2029 programme years, CMS finalised a modification in the FY 2025 IPPS final rule (89 FR 69507–69508) while HCAHPS survey updates are being publicly reported. Only six unchanged dimensions are scored: Communication with Nurses, Communication with Doctors, Communication about Medicines, Discharge Information, Cleanliness and Quietness, and Overall Rating. These produce a pre-normalised base score of 0–60, which is multiplied by 8/6 (1.3333333) and rounded to restore the familiar 0–80 range. Consistency points are calculated the same way but across those six dimensions. The domain still totals 0–100.

A worked TPS build

Domain	Points earned / possible	Normalised	Weight	Weighted score
Clinical Outcomes	24 / 60	40.00	25%	10.00
Person and Community Engagement	31 / 100	31.00	25%	7.75
Safety	21 / 60	35.00	25%	8.75
Efficiency and Cost Reduction	8 / 10	80.00	25%	20.00
Total Performance Score				46.50

Read the bottom row of that table against the first: 43% of this hospital’s entire TPS came from one measure. Eight points on MSPB Hospital contributed 20.00 TPS points; twenty-four points spread across six clinical outcome measures contributed 10.00. That is not a quirk of the example — it is what equal domain weighting plus unequal domain sizes produces, every year, for every hospital.

Step 3 — TPS to money: the linear exchange function

The programme is funded by taking money away first. 42 CFR 412.162(b)(1): the total amount available for value-based incentive payments in a fiscal year equals the total amount of base operating DRG payment reductions for that year, as estimated by the Secretary. It is a redistribution, not an appropriation. CMS’s own framing: a hospital “may earn back a value-based incentive payment percentage that is less than, equal to, or more than the applicable reduction for that program year.”

The withhold. The “applicable percent” defined in 42 CFR 412.160 rose on a statutory schedule — 1.0% in FY 2013, 1.25% FY 2014, 1.50% FY 2015, 1.75% FY 2016 — and has been 2.0% for FY 2017 and every subsequent fiscal year. The FY 2027 IPPS final rule confirms 2.00 percent for FY 2027 under section 1886(o)(7)(C)(v), with approximately $1.9 billion available for redistribution based on the December 2025 update of the FY 2025 MedPAR file.

What the percentage applies to. Only the base operating DRG payment amount: the wage-adjusted DRG operating payment plus any applicable new technology add-on payments. Per 42 CFR 412.160, that figure is determined without regard to any HRRP adjustment, and it excludes indirect medical education, disproportionate share, outlier and low-volume payments. A hospital estimating its VBP exposure off total Medicare inpatient revenue will overstate it substantially.

The two equations

Per 42 CFR 412.162(b)(3) and (c):

Value-based incentive payment percentage = applicable percent × (TPS ÷ divisor) × linear exchange function slope

Value-based incentive payment adjustment factor = 1 + (value-based incentive payment percentage − applicable percent)

That adjustment factor is then multiplied against the base operating DRG payment amount for each discharge in the fiscal year.

The slope is not a policy choice made in advance. It is solved for: it is whatever value makes the aggregate of all hospitals’ incentive payments equal the aggregate withhold, given the actual distribution of TPSs that year. That is what the definition of “Linear Exchange Function” in 42 CFR 412.160 requires — each hospital’s percentage is based on its TPS, and total payments equal total available funds. Two things follow. First, you cannot know your adjustment until every hospital’s score is known, which is why CMS publishes proxy factors in the proposed rule and actual factors only later. Second, a slope from one year cannot be mixed with a divisor from another — the two are jointly determined.

The divisor: a live discrepancy in the current regulation

This is where the current CFR text does not read cleanly, and a hospital doing its own arithmetic should know why.

42 CFR 412.162(b)(3)(ii) sets the divisor: “For fiscal years before FY 2026, the hospital’s Total Performance Score divided by 100; or… Beginning with FY 2026, the hospital’s Total Performance Score divided by 110.”

The 110 was introduced for the Health Equity Adjustment (HEA), adopted in the FY 2024 IPPS final rule (88 FR 59092–59106), which would have added up to 10 bonus points on top of a 100-point TPS for high-performing hospitals serving high proportions of dually eligible patients — hence a 110-point denominator.

CMS then removed the HEA. The FY 2026 IPPS/LTCH PPS final rule finalised its removal “effective with the FY 2026 program year,” on the stated reasoning that it simplified the scoring methodology and that its impact was small — CMS’s own estimate was an average net FY 2026 payment adjustment of 0.170% with the HEA and 0.168% without it. The codifying amendments at 90 FR 37202 removed the definition of “Health equity adjustment bonus points” from 412.160, removed 412.165(b)(5), and redesignated the following paragraph as the new 412.165(b)(5), which now reads that the TPS “is the sum of the weighted domain scores up to a maximum score of 100.”

Those amendments did not touch 412.162(b)(3)(ii)(B). As the eCFR reads today, the TPS is capped at 100 by one section and divided by 110 by another, and 412.160 still carries the definitions of “Measure performance scaler” and “Underserved multiplier” — the two HEA components — with nothing left that uses them.

We are not asserting what CMS intends here, and this is not a determination a hospital should make on its own. The operative numbers are the ones CMS publishes: the actual value-based incentive payment adjustment factors, exchange function slope and estimated amount available appear in IPPS Table 16B on cms.gov, and your own factor appears in your Percentage Payment Summary Report. If your independently computed factor disagrees with Table 16B, the divisor is the first place to look — and the discrepancy is a review-and-corrections question, not an arithmetic one you can settle yourself.

A worked payment calculation

Using the illustrative TPS of 46.50 from above, the FY 2027 applicable percent of 2.00%, and the FY 2027 proxy slope of 3.4489188481 published with the FY 2027 proposed rule. CMS states plainly that these proxy factors “will not be used to adjust hospital payments” — they were calculated from FY 2026 historical baseline and performance periods using the March 2026 update to the FY 2025 MedPAR file, and Table 16B with the real figures was expected in the autumn of 2026.

	Divisor 100	Divisor 110 (as 412.162(b)(3)(ii)(B) reads)
TPS ÷ divisor	0.4650	0.422727
× slope 3.4489188481	1.603743	1.457948
× applicable percent 2.00%	3.2075%	2.9159%
Adjustment factor = 1 + (3.2075 − 2.00)/100	1.012075	1.009159
Net effect per discharge	+1.2075%	+0.9159%
On a $9,000 base operating DRG amount	$9,108.67	$9,082.43
Implied break-even TPS	28.99	31.89

The last row is the number quality directors actually want and CMS does not publish as such. Break-even TPS = divisor ÷ slope — the score at which your adjustment factor is exactly 1.0000 and you earn back precisely what was withheld. Below it you lose money; above it you gain. Because the slope is re-solved every year against that year’s TPS distribution, the break-even score moves every year too, which is why “we scored 46 last year and 46 this year” does not imply the same payment outcome.

Who is scored, and who is not

The definition of “Hospital” in 42 CFR 412.160 does the exclusion work. A subsection (d) hospital is not in the programme for a fiscal year if any of the following applies:

• It is subject to the Hospital IQR payment reduction under section 1886(b)(3)(B)(viii)(I) for that year.
• It was cited for deficiencies that pose immediate jeopardy during the applicable performance period — defined precisely as immediate jeopardy cited on at least three surveys using Form CMS-2567, with specific rules for which date assigns a citation to a performance period (issue date for EMTALA-only citations, ASPEN survey end date otherwise).
• There are not a minimum number of applicable measures for the performance period.
• There are not a minimum number of cases for the measures that apply.

Maryland hospitals, Puerto Rico hospitals under certain arrangements, critical access hospitals and the specialty PPS-exempt hospitals sit outside the IPPS and therefore outside this programme entirely.

Reviewing, correcting and appealing your TPS

Before the score is final. Under 42 CFR 412.163, CMS gives each hospital confidential hospital-specific reports and discharge-level information used to calculate its measure, condition and domain results and its TPS, then allows 30 days to review and submit corrections. CMS also supplies a Mortality and Complication Hospital-Specific Report and an MSPB Hospital-Specific Report for each performance period, so that performance-period measure results can be challenged before they feed the score. CMS anticipates the Percentage Payment Summary Report for a coming fiscal year being available on the HQR Secure Portal by 1 August, before that fiscal year starts. Results become public in the January Care Compare refresh and on the Provider Data Catalog at data.cms.gov.

Appeal. 42 CFR 412.167(a) enumerates exactly ten appealable issues, and they are all arithmetic and process rather than clinical judgement: denial of a correction request; whether achievement/improvement points were calculated correctly; whether the higher of the two was properly used; whether domain scores including normalisation were correct; whether the proper lowest dimension score was used for HCAHPS consistency points; whether consistency points were calculated correctly; whether the correct domain scores went into the TPS; whether each domain was weighted properly; whether the weighted domain scores were properly summed; and whether the hospital’s open/closed status including mergers and acquisitions is correctly recorded. Appeals must be filed within 30 days of CMS’s denial of a corrections request or of the close of the review-and-corrections period, and must include the CCN, the hospital name, the specific basis, and CEO and QualityNet security official contact details with physical addresses. A dissatisfied hospital may then request an independent CMS review.

What you cannot appeal. 42 CFR 412.167(d) bars administrative and judicial review of the incentive-payment methodology and amount, the funding amount and the payment reduction, the performance standards and performance periods, the measures selected, the scoring methodology and the calculation of scores, and the validation methodology. In practice: you may argue that CMS applied its method to your data incorrectly. You may not argue that the method is wrong.

Extraordinary circumstance exception. Under 42 CFR 412.165(c), a hospital may request an ECE within 60 calendar days of the extraordinary circumstance, via QualityNet. CMS may also grant an ECE to hospitals that did not request one, where a systemic CMS data-collection problem or a region-wide event is involved.

How the measure set changes, and why it keeps shrinking

All VBP measures are drawn from measures already specified under the Hospital IQR Program (42 CFR 412.164(a)), and once selected they stay until CMS removes or replaces them through rulemaking. 42 CFR 412.164(c)(3) lists eight removal factors, of which the first explains most of the historical churn: a measure may be removed where performance is “so high and unvarying that meaningful distinctions and improvements in performance can no longer be made” — topped out, defined operationally as statistically indistinguishable performance at the 75th and 90th percentiles together with a truncated coefficient of variation of 0.10 or less. The other seven cover misalignment with clinical guidelines, availability of a broader or more proximal measure, no link to better outcomes, availability of a more strongly associated measure, negative unintended consequences, infeasibility, and cost exceeding benefit. A separate patient safety exception lets CMS remove a measure immediately, with notice in the Federal Register, where continued data submission itself raises patient safety concerns.

This is why the chart-abstracted process measures that dominated the early programme are largely gone and the surviving set is outcome-, experience- and cost-weighted. It also means a domain’s size changes over time — and, per the normalisation arithmetic above, changing the number of measures in a domain silently changes what one measure point is worth.

Where to verify all of this

• 42 CFR part 412, subpart I — the codified programme (412.160–412.168 for VBP; 412.150–412.154 for HRRP; 412.170–412.172 for HAC Reduction).
• CMS Hospital Value-Based Purchasing Program — reports, Table 16B adjustment factors, and current programme notices.
• QualityNet Hospital VBP — technical measure specification updates, which CMS makes outside rulemaking (412.164(c)(1)).
• The IPPS/LTCH PPS final rule for your programme year — the only authoritative source for that year’s domain weights, measure set, baseline and performance periods and performance standards. FY 2027: 91 FR 49570 (4 August 2026). FY 2026: 90 FR 36536 (4 August 2025).
• The original Hospital Inpatient VBP final rule, 76 FR 26490 (6 May 2011), for the achievement, improvement and consistency-point formulas and CMS’s own worked examples.

For the wider quality-measurement and reporting context this programme sits inside, see the patient safety cluster, and in particular how QAPI plans, QAPI reports and PIP write-ups are documented — the internal improvement machinery that generates the performance a TPS eventually scores — and the NHSN VAE surveillance definition, whose data does not feed VBP but is routinely assumed to.

Frequently asked questions

What is a good Total Performance Score?

There is no fixed answer, and any source that gives you one is guessing. The only threshold that means anything financially is the break-even TPS — divisor divided by that year’s linear exchange function slope — and it moves annually because the slope is re-solved against each year’s distribution of scores. Derive it from the year’s Table 16B, or find the TPS at which the published adjustment factor equals 1.0000.

Which counts, achievement points or improvement points?

Whichever is higher, automatically, per measure or HCAHPS dimension. You do not elect. 42 CFR 412.167(a)(3) makes CMS’s correct application of that rule specifically appealable.

Do readmissions affect my Total Performance Score?

No. 42 CFR 412.164(a) requires CMS to select VBP measures “other than measures of readmissions.” Readmissions are scored separately under the Hospital Readmissions Reduction Program, which is a penalty-only programme with its own arithmetic.

Is the TPS the same thing as the Hospital Compare star rating?

No. The Overall Hospital Quality Star Rating is a separate construct codified at 42 CFR 412.190. It is a public-reporting summary and carries no payment consequence of its own; the TPS is a payment calculation.

Why did our TPS fall even though our quality improved?

Three structural reasons. Achievement thresholds and benchmarks are recalculated from a new baseline period each programme year, so the national bar moves. Your improvement threshold is your own prior performance, so a good year raises the bar you must beat next year. And normalisation means a change in the number of measures in a domain changes what each measure point is worth, independent of your clinical performance.

Where do we find our own TPS and adjustment factor?

The Percentage Payment Summary Report on the HQR Secure Portal, which CMS anticipates publishing by 1 August before the start of the applicable fiscal year, gives your measure scores, domain scores, TPS and adjustment factor. Aggregate factors for all hospitals by CCN are published in IPPS Table 16B on cms.gov, and results are posted publicly in the January Care Compare refresh and on data.cms.gov.

Can a hospital be excluded from the programme for a safety citation?

Yes. Under the definition of “Hospital” in 42 CFR 412.160, a hospital cited for immediate jeopardy on at least three surveys using Form CMS-2567 during the applicable performance period is not a “hospital” for that fiscal year and is neither scored nor paid an incentive.

Is the Health Equity Adjustment still part of the score?

No. It was adopted in the FY 2024 IPPS final rule to begin with the FY 2026 programme year, then removed by the FY 2026 IPPS final rule effective with that same FY 2026 programme year — so in practice it never applied to a scored year. The codifying amendments removed its definition from 42 CFR 412.160 and its calculation from 412.165(b). Note, though, that 412.162(b)(3)(ii)(B)’s 110 divisor and the “measure performance scaler” and “underserved multiplier” definitions remain in the current CFR text; use CMS’s published Table 16B factors rather than computing your own.