Methodology
Transparency
In scope
57,950
Scored
57,560
99.327% coverage
Unscored
390
each served with a reason
Registry models
76
default models drive the composite
Sources tracked
26
0 stale
Open errors
0
as of 2026-08-05
Coverage ceiling
Coverage is bounded at roughly 98.87 percent by source suppression, not by pipeline completeness. Of the residual unscored facilities, those holding a federal CCN have Care Compare rows whose values CMS withholds below its case-count thresholds, and the remainder have no CCN and sit outside the federal surface entirely. The unscored set is therefore biased toward small providers and is not a random sample. See docs/coverage_ceiling_source_suppression.md.
Grade bands are fixed percentile cuts within peer groups: F ≥ 95, D ≥ 85, C ≥ 60, B ≥ 30, A < 30. Band counts cannot be read as a trend.
Grade distribution
57,560A
17,421
B
17,129
C
14,488
D
5,655
F
2,867
Lowest-coverage peer groups
15| Setting | State | In scope | Scored | Below floor | Coverage |
|---|---|---|---|---|---|
| nf | PA | 52 | 0 | 52 | 0.0% |
| nf | CT | 49 | 0 | 0 | 0.0% |
| nf | FL | 18 | 0 | 12 | 0.0% |
| home health | IN | 379 | 291 | 70 | 76.8% |
| hospice | FL | 79 | 62 | 0 | 78.5% |
| hospice | NY | 42 | 39 | 3 | 92.9% |
| nf | TX | 24 | 23 | 1 | 95.8% |
| home health | NY | 1,350 | 1,295 | 54 | 95.9% |
| snf | SD | 98 | 95 | 3 | 96.9% |
| snf | HI | 43 | 42 | 1 | 97.7% |
| snf | MT | 61 | 60 | 1 | 98.4% |
| snf | CA | 1,180 | 1,163 | 3 | 98.6% |
| home health | CT | 72 | 71 | 1 | 98.6% |
| icf iid | CA | 1,016 | 1,004 | 0 | 98.8% |
| home health | WI | 90 | 89 | 1 | 98.9% |
Groups where scoring is limited by source suppression or missing components. Unscored facilities remain listed in the screener with reasons.
Model registry
76| Score type | Version | Status | Effective | Registry notes |
|---|---|---|---|---|
| billing conduct | v0.1-cacdph | default | 2026-07-25 | California hospice live discharge, over-180-day stay and non-terminal admission mix rates. |
| billing conduct | v0.1-hci | default | 2026-07-26 | Hospice Care Index sub-indicators from CMS Care Compare, claims-derived and independent of the Provider of Services file. Scored as conduct rather than quality because the indicators are live-discharge timing, level-of-care mix, visit gaps and long-stay share. |
| billing conduct | v0.1-pacpuf | default | 2026-07-25 | R21. CMS PAC PUF home health 2023. Outlier share, risk-adjusted episode churn, LUPA rate. |
| billing conduct | v0.1-pacpuf-hos | default | 2026-07-25 | R22. PAC PUF hospice 2023. Long stay vs acuity, visit intensity, charge markup. |
| billing conduct | v0.1-pacpuf-snf | default | 2026-07-25 | R22. PAC PUF SNF 2023. PDPM coding intensity vs acuity, stay length, charge markup. |
| chow risk | v0.1-flahca | default | 2026-07-25 | Recency of the AHCA reported change of controlling owner, owner chain scale, domicile and vehicle naming. |
| chow risk | v0.1-ohodh | default | 2026-07-25 | Ohio DOH facility listing: owner portfolio size and recency of licence opening as an entry and control proxy. |
| chow risk | v0.1-pecoshha | default | 2026-07-25 | Home health change of ownership from PECOS owner association dates, the created-for-acquisition flag, owner cohort churn and investment or holding company layering. |
| chow risk | v0.1-pos | default | 2026-07-25 | Change-of-ownership footprint in the CMS Provider of Services file: control-family, legal-name and address deltas across quarterly snapshots, plus certification reset recency and Medicare tenure. |
| chow risk | v0.1-txhhsc | default | 2026-07-25 | Texas licence reissue timing as an ownership-transfer proxy. |
| chow risk | v0.6 | default | 2026-07-25 | |
| composite | v0.1 | superseded | 2026-07-24 | Pre-registry model iteration, superseded by a later composite rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| composite | v0.2 | superseded | 2026-07-24 | Pre-registry model iteration, superseded by a later composite rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| composite | v0.3 | superseded | 2026-07-25 | Pre-registry model iteration, superseded by a later composite rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| composite | v0.4 | superseded | 2026-07-24 | Pre-registry model iteration, superseded by a later composite rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| composite | v0.5 | superseded | 2026-07-25 | Pre-registry model iteration, superseded by a later composite rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| composite | v0.6 | superseded | 2026-07-25 | Pre-registry model iteration, superseded by a later composite rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| composite | v0.7 | superseded | 2026-07-25 | Pre-registry model iteration, superseded by a later composite rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| composite | v0.8 | superseded | 2026-07-25 | Pre-registry model iteration, superseded by a later composite rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| composite | v0.9 | superseded | 2026-07-25 | Pre-registry model iteration, superseded by a later composite rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| composite | v1.0 | addressable | 2026-07-25 | R20. Higher coverage (72.7% vs 68.1%) but includes ownership_opacity, which runs backwards against measured billing behaviour. |
| composite | v1.1 | addressable | 2026-07-25 | R21. billing_conduct at 0.30; ownership_opacity removed from weighting after it tested inversely related to conduct. |
| composite | v1.2 | default | 2026-07-25 | Rank-based composite. Each component is percentile-ranked within its own model and peer group, the weighted mean of those ranks is itself ranked, and the published score is that position. Fixes the tail compression that v1.1 suffered when a broad low-mean component was added. Grade band sizes are fixed by construction and cannot be read as a trend. v1.1 remains addressable for level questions. |
| financial risk | v0.1 | superseded | 2026-07-24 | Pre-registry model iteration, superseded by a later financial_risk rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| financial risk | v0.1-cacdph | default | 2026-07-25 | California home health operating scale, public payer visit concentration, service intensity and lack-of-funds discharges. |
| financial risk | v0.1-hospiceown | default | 2026-07-26 | Hospice ownership structure from CMS PECOS enrollment and All Owners. Reaches newly certified hospices that no federal quality file covers. Structure only, no quality content; the two score types from this model share one source. |
| financial risk | v0.1-nyacf | default | 2026-07-25 | NY assisted living occupancy, trend and turnover from NYSDOH ACF quarterly report. Disjoint from the cost-report financial_risk models. |
| financial risk | v0.1-txclosures | default | 2026-07-25 | Texas HHSC HCSSA closure footprint: own-licence closure, owner closure count, closure recency and branch churn, as a going-concern proxy. |
| financial risk | v0.1-txhcssa | default | 2026-07-25 | Texas HCSSA current client census and service breadth. |
| financial risk | v0.2 | superseded | 2026-07-24 | Pre-registry model iteration, superseded by a later financial_risk rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| financial risk | v0.3 | superseded | 2026-07-25 | Pre-registry model iteration, superseded by a later financial_risk rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| financial risk | v0.4 | superseded | 2026-07-25 | Pre-registry model iteration, superseded by a later financial_risk rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| financial risk | v0.5 | default | 2026-07-25 | R18. |
| lease risk | v0.4 | superseded | 2026-07-24 | Pre-registry model iteration, superseded by a later lease_risk rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| lease risk | v0.5 | default | 2026-07-25 | |
| ownership opacity | v0.1-cacdph | default | 2026-07-25 | CDPH licensee chain size, parent corporation and domicile. |
| ownership opacity | v0.1-flahca | default | 2026-07-25 | AHCA owner chain concentration, profit status, domicile and public web presence. |
| ownership opacity | v0.1-hospiceown | default | 2026-07-26 | Hospice ownership structure from CMS PECOS enrollment and All Owners. Reaches newly certified hospices that no federal quality file covers. Structure only, no quality content; the two score types from this model share one source. |
| ownership opacity | v0.1-nyacf | default | 2026-07-25 | NY adult care operator concentration and entity form from the NYSDOH ACF directory. |
| ownership opacity | v0.1-nyadhc | default | 2026-07-25 | NYSDOH adult day operator and co-operator structure. |
| ownership opacity | v0.1-nylhcsa | default | 2026-07-25 | NYSDOH operator and co-operator structure. |
| ownership opacity | v0.1-pecos | addressable | 2026-07-25 | R21. Retained as a description of ownership structure. Deliberately NOT in any composite: inversely related to billing conduct. |
| ownership opacity | v0.1-txhcssa | default | 2026-07-25 | Texas HCSSA legal entity chain and domicile. |
| ownership opacity | v0.1-txhhsc | default | 2026-07-25 | Texas HHSC directory ownership structure: chain size, absentee owner, management company, entity form, owner closure history. |
| p adverse action 12m | v0.1 | default | 2026-07-29 | R23/W4-3. Deterministic logistic p(Medicare termination within 12m), SNF only. Label = first POS termination date (any code: voluntary closure, involuntary termination, merger exit); facilities already terminated by the panel date are excluded, not counted as safe. Trained on panels 2024-07/2024-10/2025-01 (n=46,466, 283 positives), evaluated on the held-out later panel 2025-04 BEFORE shipping: AUC 0.971, top-decile lift 9.5x, top-decile capture 95.4% of the 65 terminations that followed. Separability is high because closures announce themselves in the data: collapsing occupancy is the strongest signal, and the learned NEGATIVE weight on 12m deficiency counts is the surveys-stop signature of a facility winding down, not a claim that clean inspections are dangerous. Same 15 as-of-date features and anti-leakage chassis as p_chow_12m; SFF entry rejected as a label (only one month of SFF history exists, so as-of transitions cannot be reconstructed). Score = probability x 100; grade NULL. Not in any composite. |
| p chow 12m | v0.1 | default | 2026-07-29 | R22/W4-2. Deterministic logistic p(CHOW within 12m), SNF only. Trained on panels 2024-07/2024-10/2025-01 (n=50,700, 1,207 positives), evaluated on the held-out later panel 2025-04 BEFORE shipping: AUC 0.874, top-decile lift 5.0x (top 10% of predictions captured 49.6% of the CHOWs that actually closed in the next 12 months). Shipped coefficients are the evaluated ones; no post-holdout refit. Caveat: holdout positives (133) run below the training base rate because recent CHOWs surface in public records with a lag, so measured lift is conservative. Features are as-of-date public records: PBJ staffing level/trend/contract share (quarterly history to 2022Q1), deficiency and CMP counts (12m), ownership tenure and churn from CHOW-evidenced transactions, beds, occupancy, ownership type, chain size. Chain membership and ownership type are current-state (history not published). SFF and star ratings deliberately excluded (no as-of history). Notable learned direction: long ownership tenure RAISES sale odds and a recent prior CHOW lowers them. Score = probability x 100; grade is NULL because a probability is not a quality letter. Not in any composite. |
| quality risk | v0.1 | default | 2026-07-25 | |
| quality risk | v0.1-cacdph | default | 2026-07-25 | California home health discharge to hospital and discharge by death rates, not case mix adjusted. |
| quality risk | v0.1-cahps | default | 2026-07-26 | CAHPS Hospice Survey bottom-box caregiver results from CMS Care Compare. Independent of claims and of state licensure. |
| quality risk | v0.1-nylhcsastat | default | 2026-07-25 | NYSDOH LHCSA Annual Statistical Report: institutional discharge rate and mandatory annual filing continuity, 2021 to 2024. |
| regulatory risk | v0.1 | superseded | 2026-07-24 | Pre-registry model iteration, superseded by a later regulatory_risk rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| regulatory risk | v0.1-caclhf | default | 2026-07-26 | CA CLHF assessed state penalties from the CDPH enforcement file, evidence through 2024-06-30 only. |
| regulatory risk | v0.1-flahca | default | 2026-07-26 | FL AHCA survey deficiencies by severity class, absolute and per bed, plus complaints and emergency action. Browser harvest. |
| regulatory risk | v0.1-nynhp | default | 2026-07-26 | NYSDOH Nursing Home Profile inspection and complaint metrics for non-Medicare nursing homes, resolved on operating certificate. |
| regulatory risk | v0.1-padoh | default | 2026-07-26 | PA nursing care facility sanctions Q2 2026, joined on state facility id. Single-source setting, below composite floor. |
| regulatory risk | v0.1-txhcssa | default | 2026-07-26 | Texas HHSC HCSSA licence status. Flags active state enforcement; records status not severity. |
| regulatory risk | v0.2 | superseded | 2026-07-24 | Pre-registry model iteration, superseded by a later regulatory_risk rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| regulatory risk | v0.3 | default | 2026-07-25 | Deficiency-based, SNF population. Runs alongside v0.5-pos. |
| regulatory risk | v0.4-pos | retracted | 2026-07-25 | R19. Added 20 points for acptbl_poc_sw=No, wrongly penalising 1,391 compliant ICF/IID. Retained for audit, never to be consumed. |
| regulatory risk | v0.5-pos | superseded | 2026-07-25 | R20. Corrects the acptbl_poc_sw error in v0.4-pos; covers provider types 3 and 8. Superseded by v0.6-pos, which extends the population to hospices with no expression change. |
| regulatory risk | v0.6-pos | default | 2026-07-27 | R31. Extends v0.5-pos to hospices (provider type 12): same fields, same weights, no expression change. Hospice regulatory coverage before this was 148 of 6,535 scored hospices, all from the TX HCSSA state model. SNF/NF (type 20) deliberately excluded: the deficiency-based v0.3 already covers 14,315 of 14,720 and is richer evidence than a POS compliance flag. |
| staffing risk | v0.1 | superseded | 2026-07-24 | Pre-registry model iteration, superseded by a later staffing_risk rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| staffing risk | v0.2 | superseded | 2026-07-24 | Pre-registry model iteration, superseded by a later staffing_risk rebuild. Rows retained for audit per project convention (see regulatory_risk v0.4-pos); never consumed because v_cii_api_facility_score and the composite eligibility join both require status = default. Backfilled 2026-07-26 by alert triage on model_registry_integrity. |
| staffing risk | v0.3 | default | 2026-07-25 | Cost-report staffing, SNF population. Runs alongside v0.4-pos, which covers ICF/IID from POS. |
| staffing risk | v0.4-pos | default | 2026-07-25 | R19. Percentile-ranked direct-care staff per bed, ICF/IID. |
| state reg risk | v0.1-cacdph | default | 2026-07-25 | CDPH licence tenure, expiry, branch role and failure to file the mandatory HCAI annual utilization report. |
| state reg risk | v0.1-caclhf | default | 2026-07-26 | CA CLHF licence status and tenure from the CDPH facility roster. |
| state reg risk | v0.1-flahca | default | 2026-07-25 | AHCA licence status, expiry, tenure and out-of-state administration. |
| state reg risk | v0.1-inidoh | default | 2026-07-25 | Indiana DOH QAMIS directory: licence currency and published service area breadth. |
| state reg risk | v0.1-nyacf | default | 2026-07-25 | NY adult care designation score: transitional adult home and Justice Center coverage. |
| state reg risk | v0.1-nyadhc | default | 2026-07-25 | NYSDOH adult day programme age, operator domicile, legal form and operator chain size. |
| state reg risk | v0.1-nylhcsa | default | 2026-07-25 | NYSDOH LHCSA registration continuity 2019-2025. |
| state reg risk | v0.1-nynhp | default | 2026-07-26 | NYSDOH nursing home enforcement stipulations. |
| state reg risk | v0.1-ohodh | default | 2026-07-25 | Ohio DOH facility listing: licence tenure, owner disclosure, owner domicile and owner address identity. |
| state reg risk | v0.1-txhcssa | default | 2026-07-25 | Texas HCSSA licence status, tenure and branch role. |
| state reg risk | v0.5 | default | 2026-07-25 |
Only models with status 'default' contribute to the composite. 'addressable' models are queryable but excluded by design; 'retracted' models are retained for audit and must not be used.
Source freshness
26| Source | Latest edition | Last run | Days since | SLA |
|---|---|---|---|---|
| usda_mfh_exit | 2026-03-31 | 2026-07-27 | 9 | within |
| ca_hcai_ltc_disclosure_2024 | 2025-10-21 | 2026-07-27 | 9 | within |
| census_acs5_rents | 2024-12-31 | 2026-07-27 | 9 | within |
| cms_pos_iqies | 2026-04-01 | 2026-07-27 | 9 | within |
| fl_ahca_healthfinder | 2026-07-27 | 2026-07-27 | 9 | within |
| hud_user_fmr | 2025-10-01 | 2026-07-27 | 9 | within |
| hud_user_safmr | 2025-10-01 | 2026-07-27 | 9 | within |
| il_hfs_ltc_cost_report_2024 | 2025-12-03 | 2026-07-27 | 9 | within |
| usda_mfh_active | 2026-06-17 | 2026-07-27 | 9 | within |
| cms_hospice_pecos | 2026-07-21 | 2026-08-03 | 2 | within |
| cms_hha_pecos | 2026-07-21 | 2026-08-03 | 2 | within |
| cms_hospice_carecompare | 2026-05-01 | 2026-08-03 | 2 | within |
| cms_pac_puf_snf | 2025-08-05 | 2026-08-03 | 2 | within |
| cms_pac_puf_hh | 2025-08-05 | 2026-08-03 | 2 | within |
| cms_pac_puf_hos | 2025-08-05 | 2026-08-03 | 2 | within |
| hud_mf_contract_rents | 2026-06-03 | 2026-08-04 | 1 | within |
| oh_odh_eid | 2026-08-05 | 2026-08-05 | 0 | within |
| nysdoh_lhcsa | 2026-08-03 | 2026-08-05 | 0 | within |
| nysdoh_lhcsa_stat | 2026-07-07 | 2026-08-05 | 0 | within |
| nysdoh_health_data_ny | 2021-12-22 | 2026-08-05 | 0 | within |
| tx_hhsc_closures | 2026-07-20 | 2026-08-05 | 0 | within |
| tx_hhsc_directory | 2026-08-03 | 2026-08-05 | 0 | within |
| tx_hhsc_hcssa | 2026-08-03 | 2026-08-05 | 0 | within |
| ca_cdph_elms | 2026-07-29 | 2026-08-05 | 0 | within |
| in_idoh_qamis | 2026-08-05 | 2026-08-05 | 0 | within |
| nysdoh_adhc | 2026-08-03 | 2026-08-05 | 0 | within |
SLA windows are per source cadence; a stale source is a data quality event, not a cosmetic one.