Source data
The register is built directly from the CMS Skilled Nursing Facility Quality Reporting Program (SNF QRP) provider dataset, refreshed quarterly by CMS. It covers every Medicare-certified skilled nursing facility in the United States (14699 facilities at last refresh).
What v1 reports
- Potentially Preventable Readmissions (S_004_01_PPR) — risk-standardized readmission rate, with the CMS comparison against the national rate
- Discharge to Community (S_005_02_DTC) — the rate at which residents are successfully discharged back into community living, an indicator of rehabilitation effectiveness
- Medicare Spending Per Beneficiary (S_006_01_MSPB) — an efficiency index where 1.0 is the national average; lower means the facility delivers care below the national average per-beneficiary cost
- Hospitalization Rate (S_039_01_HAI) — rate of resident hospitalizations during the SNF stay, with comparison against the national rate
Each measure carries the CMS-published classification (Better / No Different / Worse than the National Rate). Branmoor aggregates the classifications into a single per-facility signal flag.
What v1 is not
v1 is the published quality-measure surface. It is not the mortality-cluster-detection product the full Nursing Home Death Cluster Detection scope (per the original spec) delivers. v1 does not include:
- Payroll-Based Journal (PBJ) staffing-adjusted expected mortality — the day-level direct-care staffing data CMS publishes for every facility, regressed against case mix to compute a facility-specific expected mortality rate. The cluster-detection algorithm runs on observed-vs-expected. Multi-GB ingest; planned
- Ownership-network overlay — CMS publishes ownership and related-party records. The PE-roll-up signature (where the same operator's facilities all show the same pattern) is detectable from this overlay. Planned
- Medicare claims-derived mortality — full mortality data requires a CMS Research Data Use Agreement (8–12 weeks). Planned for the paid layer once the DUA is in place
- State inspection findings — supplements CMS deficiencies with state-specific citations. Per-state ingest planned
Branmoor's signal aggregation
Across the four published measure families, Branmoor aggregates the CMS classifications into a single per-facility signal:
- Elevated outcome signal — Worse than the National Rate on 2 or more measures
- Mixed signal — Worse than the National Rate on exactly 1 measure
- Better than national — Better on 2 or more measures and Worse on none
- Outcomes at the national rate — otherwise
The aggregation is a navigation aid. The CMS-published measure-level classifications are the authoritative read.
Update cadence
CMS refreshes the SNF QRP dataset quarterly. Branmoor re-ingests on each site deploy.
Limits
The CMS measures are risk-standardized using documented methodology, but no risk adjustment is perfect. The data lag is typically 6–12 months from the measurement period. The v1 register does not yet include staffing data, ownership chains, or mortality cluster detection — those are the differentiating elements of the paid product.
Facilities below CMS's minimum-volume threshold for a measure are flagged as Not Available rather than scored.