Data Sources & Methodology

Data pipeline: We download the raw CMS CSV files monthly, parse and normalize the data, cross-reference each facility with Census demographics and cost data, and generate static JSON files for each facility, city, state, chain, and deficiency code.

Ratings: All CMS ratings are presented exactly as published — we do not modify, weight, or create our own rating system. Rankings (best, worst, most fined, best staffing) are derived directly from CMS data.

Geographic matching: Facilities are matched to cities and counties using the address data provided by CMS. County demographics are matched by county name within each state. Cities with facilities in multiple counties use the county of the first listed facility.

Cost assignment: Cost data is assigned at the state level (not facility level). Individual facility costs are not publicly available in any dataset we're aware of.

State comparisons: A facility's published CMS rating, nursing hours, RN hours, and turnover are compared directly with the corresponding CMS state average. We use small neutral ranges to avoid describing immaterial differences as better or worse. We do not combine these metrics into a proprietary score.

Inspection trends: The current and previous CMS inspection cycles are shown side by side using their recorded deficiency counts. A lower count does not necessarily mean lower severity, so the underlying citations, scope, severity, and correction status remain available.

Nearby facilities: Nearby alternatives are selected from coordinates in the facility data and ordered by approximate straight-line distance. Proximity is not an endorsement, and distance can differ from driving distance.

Corrections: Facility data can change between CMS releases. If a profile appears inconsistent with the source data, please contact us with the facility name and CMS Certification Number so it can be reviewed.