Reads as: riskier than 17.3% of the 14,684 facilities in the snf peer group. percentile within peer group, 100 = highest risk. Grade bands are fixed percentile cuts (F is the top 5% of peer risk), so a grade is a position, not an absolute level.
weighted_contribution sums to pre_rank_index over rows where counts_toward_index is true. The composite is the facility's percentile position of that index within its peer group, so contributions explain the input, and the rank explains the score.
ML signals — parallel engine
3
Kind
Finding
Signal
Subject
First seen
connection
Owner spans multiple chains
6 chains
SMITH KIMBERLY
9 facilities
connection
Owner spans multiple chains
4 chains
PROACTIVE MEDICAL REVIEW AND CONSULTANTS LLC
12 facilities
connection
Owner spans multiple chains
4 chains
MCINTOSH SARAH
9 facilities
Findings from the always-on scanning layer: anomalies on this facility, connections through its current owners and past buyers. Runs beside the deterministic composite, never inside it. Full board on the ML Lab page.
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.
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.
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.
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.
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.
Rows with counts_toward_composite false come from models that are addressable or retracted rather than default. They exist and may be informative, but they did not enter the published composite and must not be summed as though they had.
p_adverse_action_12m: probability the facility's Medicare participation terminates within 12 months (any POS termination code; already-terminated facilities are excluded, not scored safe). Deterministic logistic over as-of-date public records, SNF only. Holdout (2025-04 panel, evaluated before shipping): AUC 0.971, top-decile lift 9.5x. Closures announce themselves: collapsing occupancy dominates, and fewer recent surveys is a winding-down signature. inputs.features holds raw feature values, inputs.contributions each feature's standardized pull on the log-odds. Predictions never enter the composite. See /v1/models. | p_chow_12m: probability of a change of ownership within 12 months. Deterministic logistic over as-of-date public records, SNF only. Holdout (2025-04 panel, evaluated before shipping): AUC 0.874, top-decile lift 5.0x. inputs.features holds raw feature values, inputs.contributions each feature's standardized pull on the log-odds. Predictions never enter the composite. See /v1/models for full notes.
Probability the facility's Medicare participation ends within 12 months, from public records only. Red bars raise exit odds. A negative pull from deficiencies is the winding-down signature — surveys stop as a facility empties — not evidence that clean inspections are dangerous.
Staffing vs peers
2
2025Q4 HPRD
This facility
KY median
National median
National p25–p75
Total nurse
4.90
3.11
3.28
2.90–3.77
RN
0.46
0.41
0.39
0.25–0.58
Weekend total
4.49
3.01
3.11
—
Weekday total
5.06
3.16
3.35
—
Quarter
Days
Avg census
Total HPRD
RN
LPN
CNA
Weekend
Contract %
2025Q4
92
33.8
4.90
0.46
1.10
3.34
4.49
0.1%
2025Q3
92
9.4
7.93
0.49
2.32
5.12
7.37
0.0%
HPRD is census-weighted: total hours over the quarter divided by total resident-days, not an average of daily ratios. Source is CMS Payroll-Based Journal daily records; only SNFs report PBJ, so other settings legitimately return zero rows.
Penalties
0
No penalties on record for this facility. Absence of a record here means the sources CII ingests contain none, not that none exists anywhere.
Deficiencies
0
No deficiencies on record for this facility. Absence of a record here means the sources CII ingests contain none, not that none exists anywhere.
scope_severity uses the CMS letter grid: A is least serious, L most. J, K and L are immediate jeopardy. Rows are the surveyor's findings of record, not the model's opinion.
Licensure actions
0
No licensure actions on record for this facility. Absence of a record here means the sources CII ingests contain none, not that none exists anywhere.
Ownership
34
Owner
Role
Pct
From
Status
Source
FRANCES, JAY
5% OR GREATER DIRECT OWNERSHIP INTEREST
50%
2025-02-01
current
pecos_ownership
FRANCES, MOLLYANNE
5% OR GREATER DIRECT OWNERSHIP INTEREST
50%
2025-02-01
current
pecos_ownership
ANDERSON, VERONICA
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
BABER, ROY
OPERATIONAL/MANAGERIAL CONTROL
—
2025-02-01
current
pecos_ownership
CALHOUN & COMPANY PLLC
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
CHATHAM, BARRY
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
CHATHAM, BRIAN
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
CHATHAM, STEPHEN
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
FULLER, DEBBIE
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
HARGIS & ASSOCIATES, LLC
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
HARGIS, FORWOOD
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
KREIL, KAREN
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
LEGACY HEALTH SERVICES INC
OPERATIONAL/MANAGERIAL CONTROL
—
2025-02-01
current
pecos_ownership
MARTIN, ARNIE
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
MCINTOSH, SARAH
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
NICHOLAS COUNTY FISCAL COURT
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
PETERSEN, JALME
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
PROACTIVE MEDICAL REVIEW AND CONSULTANTS LLC
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
SMITH, KIMBERLY
INDIVIDUAL IS AN OWNER, PARTNER OR TRUSTEE OF ANY ADP OF THE SNF
—
2025-04-01
current
pecos_ownership
TOADVINE, STEPHEN
ADP OF THE SNF
—
2025-02-01
current
pecos_ownership
JAY FRANCES
5% OR GREATER DIRECT OWNERSHIP INTEREST
50%
2025-02-01
ended 2026-05-18
pecos_ownership
MOLLYANNE FRANCES
5% OR GREATER DIRECT OWNERSHIP INTEREST
50%
2025-02-01
ended 2026-05-18
pecos_ownership
ARNIE MARTIN
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
BARRY CHATHAM
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
BRIAN CHATHAM
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
DEBBIE FULLER
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
FORWOOD HARGIS
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
JALME PETERSEN
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
KAREN KREIL
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
ROY BABER
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
SARAH MCINTOSH
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
STEPHEN CHATHAM
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
STEPHEN TOADVINE
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
VERONICA ANDERSON
ADP OF THE SNF
—
2025-02-01
ended 2026-05-18
pecos_ownership
PECOS ownership and managing-control associations. Current edges have no end date in the source; historical edges are retained.
Related-party costs
0
No related-party cost years on record for this facility. Absence of a record here means the sources CII ingests contain none, not that none exists anywhere.
Cost-report related-party spend by fiscal year. has_propco flags real-estate counterparties.
Transactions
0
No transactions on record for this facility. Absence of a record here means the sources CII ingests contain none, not that none exists anywhere.
Most recent same-state, same-setting trades. Most CHOW-derived trades carry no disclosed consideration, so price, $/bed, and cap rate show only where a sale price was resolved from a deed or filing — blanks are honest, not missing data.
Market rent context
28
Geo
Kind
BR
Amount
As of
Source
county 21181
fmr
0
$619
2025-10-01
HUD_USER_FMR
county 21181
acs median gross
1
$568
2024-12-31
CENSUS_ACS5
county 21181
fmr
1
$696
2025-10-01
HUD_USER_FMR
county 21181
acs median gross
2
$709
2024-12-31
CENSUS_ACS5
county 21181
fmr
2
$866
2025-10-01
HUD_USER_FMR
county 21181
acs median gross
3
$841
2024-12-31
CENSUS_ACS5
county 21181
fmr
3
$1,204
2025-10-01
HUD_USER_FMR
county 21181
fmr
4
$1,311
2025-10-01
HUD_USER_FMR
county 21181
acs median gross
$728
2024-12-31
CENSUS_ACS5
zcta 40311
acs median gross
1
$568
2024-12-31
CENSUS_ACS5
zcta 40311
acs median gross
2
$686
2024-12-31
CENSUS_ACS5
zcta 40311
acs median gross
3
$842
2024-12-31
CENSUS_ACS5
zcta 40311
acs median gross
$693
2024-12-31
CENSUS_ACS5
zip 40311
payment standard 110
0
$792
2025-10-01
HUD_USER_SAFMR
zip 40311
payment standard 90
0
$648
2025-10-01
HUD_USER_SAFMR
zip 40311
safmr
0
$720
2025-10-01
HUD_USER_SAFMR
zip 40311
payment standard 110
1
$979
2025-10-01
HUD_USER_SAFMR
zip 40311
payment standard 90
1
$801
2025-10-01
HUD_USER_SAFMR
zip 40311
safmr
1
$890
2025-10-01
HUD_USER_SAFMR
zip 40311
payment standard 110
2
$1,155
2025-10-01
HUD_USER_SAFMR
zip 40311
payment standard 90
2
$945
2025-10-01
HUD_USER_SAFMR
zip 40311
safmr
2
$1,050
2025-10-01
HUD_USER_SAFMR
zip 40311
payment standard 110
3
$1,573
2025-10-01
HUD_USER_SAFMR
zip 40311
payment standard 90
3
$1,287
2025-10-01
HUD_USER_SAFMR
zip 40311
safmr
3
$1,430
2025-10-01
HUD_USER_SAFMR
zip 40311
payment standard 110
4
$1,771
2025-10-01
HUD_USER_SAFMR
zip 40311
payment standard 90
4
$1,449
2025-10-01
HUD_USER_SAFMR
zip 40311
safmr
4
$1,610
2025-10-01
HUD_USER_SAFMR
ACS median gross rents and HUD FMR/SAFMR for the facility's county and ZIP. Context for underwriting, not facility pricing.