Nobody has measured CMS HWR against HEDIS PCR on one population, but MedPAC and Yale CORE have each measured what one definition change does, and the numbers are large
The question
"Has anyone published a measured gap between the CMS Hospital-Wide Readmission rate and NCQA HEDIS PCR for the same provider population, rather than just a specification difference?"
Context: the 2026-10-02 brief established ten specification differences between CBE/NQF #1789 and HEDIS PCR. This brief looks for the next step up: a published number showing how far apart the two answers land on the same patients.
What we already know (from the vault)
- [[2026-10-02-snowflake-semantic-views-cortex-gateway-agent-identity-audit]] mapped the spec differences (facility vs plan attribution, 65+ Medicare FFS vs 18+ with commercial stopping at 64, inpatient-only vs inpatient plus observation, a risk-standardized rate vs an observed-to-expected ratio). Its open follow-up was this exact question: "A measured gap would be stronger evidence than a specification gap alone, and none was located."
- That brief also flagged that the Hybrid HWR added Medicare Advantage (MA) patients for 2026, so the CMS definition drifts under an unchanged name. One of the measured results below quantifies exactly that drift.
- The essay draft [[2026-10-02-agents-need-your-definitions-v1]] currently illustrates the point with invented numbers (14.8% vs 12.1%, labeled "Hypothetical example, invented numbers"). This brief supplies real, citable replacements.
- [[operational-definitions]] and [[2026-04-15-commoncog-whats-operational-definition]] are the concept anchors: a metric is not defined until its counting procedure is.
What the web says
- Null result on the exact pair. Three focused searches across PubMed-indexed literature, MedPAC, Milliman and state all-payer reports turned up no study that computes CMS HWR and HEDIS PCR on the same patients or the same hospitals. The closest practitioner treatment, Milliman's 2017 brief, describes the two measures side by side but computes nothing; it says only that the measures "are not easily comparable because they employ different methodologies and focus on different populations" (Alston & Berrios, Milliman, Oct 2017). A structural reason the gap is absent: HWR is attributed to a hospital and PCR to a health plan, so a head-to-head requires re-attributing one measure to the other's unit. That is a research design choice nobody appears to have published.
- Closest measured analogue, HEDIS side (definition alone, same data). MedPAC applied the 2021 HEDIS PCR specification to MA encounter data for HMO and PPO enrollees. Result: "When applied to encounter data, the HEDIS specifications excluded 45 percent of index hospitalizations and 71 percent of readmissions." Same patients, same data, and the definition alone removes nearly three-quarters of the return admissions before any rate is computed. The exclusions driving this include hospice use at any point in the year, four or more index stays ("outliers"), and the continuous-enrollment rule (MedPAC, June 2024 Report to the Congress, ch. 3, p. 112).
- Same MedPAC analysis, reporting gap rather than definition gap. Among 2.3 million beneficiaries present in both sources, "encounter data included 11 percent more hospitalizations and 19 percent more readmissions than HEDIS data did"; only 73% of encounter-data hospitalizations (86% inpatient, 40% observation stays) appeared in plan-submitted HEDIS data (Table 3-4). MedPAC adds that this "does not suggest that readmission rates would have been higher" under encounter data, because extra index stays can push beneficiaries over the outlier exclusion. Label this as a data-completeness gap under one definition, not a definitional gap.
- Closest measured analogue, CMS side (one definitional change inside HWR). Yale CORE and CMS authors ran the HWR measure with and without MA admissions on 11.0 million July 2018 - June 2019 admissions. Mean hospital risk-standardized readmission rates barely moved, "15.5% vs 15.3%," yet "1489 hospitals (33.1%) had their performance quintile ranking changed," rising to 45.3% among hospitals with the most MA patients (Kyanko et al., JAMA Netw Open 2024;7(6):e2414431). This is the change the Hybrid HWR adopted for 2026. Abstract read via PubMed; the PMC full text returned a bot-check page and was not retried.
- Two-steward linkage on overlapping Medicare patients (the nearest thing to a same-population CMS-vs-HEDIS comparison). Linking MedPAR (the CMS claims file) to HEDIS for patients 65+ hospitalized for AMI, heart failure or pneumonia, 2011-2014: 29.2%, 38.0% and 37.2% of MedPAR admissions had no HEDIS record, and those missing stays were readmitted more often (18.9% vs 14.4% for AMI; 23.7% vs 18.4% for CHF; 18.3% vs 13.9% for pneumonia). The direction of the MA-vs-traditional comparison also flipped with method: MA looked better unadjusted (16.6% vs 17.1% for AMI) and worse after standardization (17.2% vs 16.9%) (Panagiotou et al., Ann Intern Med 2019;171(2):99-106). Condition-specific cohorts, not HWR or PCR. A follow-on research letter compared contract-reported HEDIS rates against encounter-data rates (Kim et al., JAMA Netw Open 2020;3(4):e203555); only its one-line abstract was retrieved, so no numbers from it are cited here.
- Two readmission definitions on one hospital population (VA data). CMS all-cause vs 3M Potentially Preventable Readmissions on VA FY08-10 data: "moderate correlations"; "After controlling for all methodological differences but preventability, correlations increased to >90%"; the preventability rule alone changed public-reporting outlier status for 7% of hospitals and would have affected HRRP payment for 30% (Mull et al., Med Care 2013;51(7):589-96). HWR vs condition-specific on VA FY07-10 data: "among those hospitals classified as poor performers by the heart failure readmission measure, only 28.6% were similarly classified by the HWR measure," and "The HWR measure penalized only 60% of those hospitals" flagged by the condition-specific measures (Rosen et al., Med Care 2016;54(2):155-61). Abstracts only; full texts paywalled and not pursued.
Convergences and contradictions
- Convergence: the vault's spec-level claim survives contact with data. Every measured comparison found shows that definitional choices the 2026-10-02 brief listed (payer scope, observation stays, exclusions, preventability) move results materially. The most extreme is MedPAC's 71% of readmissions removed by PCR exclusions alone.
- Contradiction worth respecting: averages can agree while rankings disagree. Kyanko et al. found a 0.2-point move in the mean rate alongside a one-in-three quintile reshuffle. An essay that says "the two numbers come out far apart" would overstate the aggregate case. The defensible line is that the ranking and who gets flagged move, sometimes a lot, even when the headline number barely does.
- Gap the vault inherited, now confirmed as a gap in the literature. The 2026-10-02 brief "located none." This search also located none, and the attribution mismatch (hospital vs plan) explains why: the comparison is not a lookup, it is a study someone would have to design.
Synthesis for RDCO
The honest answer for the essay is a null with strong neighbors. Nobody has published HWR-versus-PCR on one population, and the post should not imply otherwise. But that null is itself on-thesis: the two stewards' measures cannot be lined up without first deciding which unit (hospital or plan) owns a readmission, which is a grain decision, the exact thing a semantic layer exists to make explicit. "No one has measured the gap, because you have to pick a definition before you can even ask" is a sharper line than an invented number, and it is true.
For a concrete number, replace the draft's invented 14.8% vs 12.1% with two measured results. First, MedPAC: applying the HEDIS PCR specification to the same 2021 Medicare Advantage encounter data excluded 45% of index hospitalizations and 71% of readmissions. That is one population, one dataset, and a definition doing the work. Second, Yale CORE and CMS's own authors: adding Medicare Advantage patients to the CMS hospital-wide measure moved the national average by 0.2 points (15.3% to 15.5%) and changed the performance quintile of one hospital in three. That second result is the governance story in miniature: the measure kept its name, the average looked stable, and a third of hospitals got a different grade. Both sources are primary (a MedPAC Report to the Congress and a CMS-coauthored JAMA Network Open paper), which matters for a healthcare reader who will test the claims.
Wording guardrails. Say "the HEDIS readmission definition, applied to the same claims, drops 71% of readmissions" rather than "HEDIS says readmissions are 71% lower," because MedPAC computed exclusions, not a comparable rate. Say "one in three hospitals changed quintile" rather than "one in three hospitals' rates changed significantly." Keep the VA results (Mull 2013, Rosen 2016) as backup for a second paragraph or a footnote; they are older and VA-specific, but "only 28.6% of hospitals flagged as poor performers by one CMS measure were flagged by the other" is a clean, quotable illustration of definition-driven disagreement.
For phData credibility work, this also gives a cheap original artifact if the founder ever wants one: the essay's semantic-view demo could implement a stripped-down PCR exclusion set and an HWR-style cohort on a public synthetic dataset (for example CMS's synthetic Medicare claims) and publish both answers. That would be, as far as this search can tell, the first public side-by-side, but it is a build decision for the founder, not a research finding.
Why this is in the vault
It replaces the invented numbers in the "Agents don't need your data warehouse. They need your definitions." essay draft ([[2026-10-02-agents-need-your-definitions-v1]]) with two primary-source measured figures, and closes the open follow-up left in [[2026-10-02-snowflake-semantic-views-cortex-gateway-agent-identity-audit]] with a documented null.
Open follow-ups
- Does the Kim et al. 2020 research letter (JAMA Netw Open e203555) report contract-level readmission rates under HEDIS-reported vs encounter-recomputed data, and how large is the per-contract spread? Only its stub abstract was retrieved.
- Does any state all-payer claims database report (for example Massachusetts CHIA's adapted Yale/CMS all-payer hospital-wide measure) publish rates for the same hospitals under both a Medicare-only and an all-payer cohort, giving a payer-scope gap on one hospital population?
- How much does including observation stays (a PCR feature that HWR lacks) change measured readmission rates on one population? Published work on observation-stay exclusion in HRRP would quantify one of the ten spec differences directly.
- What did the Hybrid HWR's 2026 public-reporting rollout show for hospitals' rate and rank changes versus claims-only HWR, if CMS or Yale CORE published a dry-run comparison?
Related
- [[2026-10-02-snowflake-semantic-views-cortex-gateway-agent-identity-audit]] - parent brief with the ten-element spec comparison this brief upgrades
- [[2026-10-02-agents-need-your-definitions-v1]] - essay draft whose invented readmission numbers these measured figures replace
- [[2026-09-29-agents-need-your-definitions]] - earlier version of the same essay
- [[operational-definitions]] - concept: a metric is defined by its counting procedure
- [[2026-04-15-commoncog-whats-operational-definition]] - Cedric Chin source behind the concept
- [[semantic-layer]] - concept the essay argues for
- [[citation-version-discipline]] - why the HWR measure's MA change must be cited with its year
Sources
Vault:
- 06-reference/research/2026-10-02-snowflake-semantic-views-cortex-gateway-agent-identity-audit.md
- 01-projects/personal-brand/drafts/2026-10-02-agents-need-your-definitions-v1.md
- 01-projects/personal-brand/drafts/2026-09-29-agents-need-your-definitions.md
- 06-reference/concepts/operational-definitions.md
- 06-reference/2026-04-15-commoncog-whats-operational-definition.md
- 06-reference/concepts/semantic-layer.md
- 06-reference/concepts/citation-version-discipline.md
Web:
- https://www.medpac.gov/wp-content/uploads/2024/06/Jun24_Ch3_MedPAC_Report_To_Congress_SEC.pdf - MedPAC June 2024, ch. 3, "Comparing Medicare Advantage encounter data with other data sources" (pp. 101-113): HEDIS PCR spec applied to 2021 MA encounter data excludes 45% of index stays and 71% of readmissions; encounter data shows 11% more hospitalizations and 19% more readmissions than HEDIS among 2.3M shared beneficiaries; Tables 3-3 and 3-4. Primary, full text read.
- https://pubmed.ncbi.nlm.nih.gov/38829614 - Kyanko K, et al. Incorporating Medicare Advantage Admissions Into the CMS Hospital-Wide Readmission Measure. JAMA Netw Open 2024;7(6):e2414431. Abstract via PubMed E-utilities; PMC full text (https://pmc.ncbi.nlm.nih.gov/articles/PMC11148674/) blocked by a bot check, not retried.
- https://pubmed.ncbi.nlm.nih.gov/31234205 - Panagiotou OA, et al. Hospital Readmission Rates in Medicare Advantage and Traditional Medicare. Ann Intern Med 2019;171(2):99-106. Abstract only.
- https://pubmed.ncbi.nlm.nih.gov/32343350 - Kim D, et al. Assessment of Completeness of Hospital Readmission Rates Reported in MA Contracts' HEDIS. JAMA Netw Open 2020;3(4):e203555. Stub abstract only; no figures cited.
- https://pubmed.ncbi.nlm.nih.gov/23604016 - Mull HJ, et al. Comparing 2 methods of assessing 30-day readmissions (CMS vs 3M PPR), VA. Med Care 2013;51(7):589-96. Abstract only; full text paywalled.
- https://pubmed.ncbi.nlm.nih.gov/26595224 - Rosen AK, et al. Does Use of a Hospital-wide Readmission Measure Versus Condition-specific Readmission Measures Make a Difference? Med Care 2016;54(2):155-61. Abstract only; full text paywalled.
- https://media.milliman.com/v1/media/edge/images/millimaninc5660-milliman6442-prod27d5-0001/media/Milliman/importedfiles/uploadedFiles/insight/2017/differences-in-details.pdf - Alston M, Berrios M. The differences are in the details. Milliman, Oct 2017. Qualitative comparison only; no computed gap.
- https://www.chiamass.gov/assets/docs/r/pubs/18/Readmissions-Tech-Appendix-2018-05.pdf - Massachusetts CHIA all-payer hospital-wide readmissions technical appendix (seen in search results only, not read; basis for a follow-up).
Method note: 3 WebSearch queries, 3 WebFetch attempts (Milliman and MedPAC PDFs read via local text extraction; PMC blocked), plus PubMed E-utilities abstract pulls for five papers.