Are Nursing Homes Safe? What CMS Data Reveals

There are 14,703 Medicare-certified nursing homes in the United States. CMS publishes quality data on every one of them: star ratings, health inspection results, staffing levels, fines, and deficiency citations. We analyzed all of it.

The picture is not reassuring.

The Numbers

Our nursing home analysis covers the full CMS dataset:

• $471 million in fines issued across all facilities
• 418,000+ health citations on record
• 70% of facilities are understaffed relative to CMS benchmarks
• 1-star homes bill Medicare 7.2% more than 5-star homes

That last number is the most troubling. The lowest-quality facilities, those with the worst inspection records and lowest staffing levels, bill Medicare at higher rates than the best facilities. Lower quality, higher cost.

What Star Ratings Actually Mean

CMS assigns each nursing home a rating from 1 to 5 stars based on health inspections, staffing, and quality measures. The system is designed to help families compare facilities.

But the ratings have known limitations. Health inspections happen infrequently and are sometimes announced in advance. Staffing data is self-reported by facilities. And quality measures rely on clinical data that facilities themselves submit.

Our analysis cross-references star ratings with other data points: fine amounts, citation severity, complaint volumes, and Medicare spending per resident. The correlation between star ratings and these independent measures is positive but imperfect. Some 4-star facilities have worse fine histories than some 2-star facilities.

The Private Equity Question

A growing share of nursing homes are owned by private equity-backed chains. Our PE nursing home analysis compared PE-structured facilities against non-PE facilities using the same CMS data.

The differences are measurable:

• PE-structured chains average 2.26 stars vs 3.03 for non-PE facilities
• 1.9x higher fines per facility for PE-owned homes
• Lower staffing levels across nursing, aide, and physical therapy categories

These findings align with published academic research from NBER, GAO, and PERI. The pattern is consistent: when ownership prioritizes financial returns, quality metrics decline.

This does not mean every PE-owned facility is unsafe. Some perform well. But at the portfolio level, the data shows a clear gap.

Geographic Hotspots

Nursing home quality varies significantly by state. Some states have stronger enforcement, higher staffing requirements, and more frequent inspections. Others have minimal oversight.

Our analysis maps citation density, fine amounts, and staffing ratios by state. The variation is large enough that choosing a facility in a high-enforcement state is, statistically, a better starting point than choosing one in a low-enforcement state.

What Families Should Look For

CMS star ratings are a starting point, not the final answer. Here is what the data suggests matters most:

Staffing levels. Facilities with higher registered nurse hours per resident day have better outcomes across virtually every quality measure. This is the single most predictive indicator in the data.

Fine history. A facility with repeated fines for the same type of violation (infection control, fall prevention, medication errors) is showing a pattern, not an isolated incident.

Complaint-driven inspections. CMS conducts inspections in response to complaints. Facilities with frequent complaint-driven inspections may have issues that standard inspections miss.

Ownership changes. Rapid ownership changes, especially involving complex corporate structures, correlate with declining quality in the CMS data. When a facility changes hands multiple times in a few years, staffing and care often suffer during transitions.

The Data Is Public

Every data point in our analysis comes from CMS Nursing Home Compare, CMS Provider of Service files, and state inspection records. CMS publishes this data specifically so that families and researchers can evaluate facility quality.

The problem is not data availability. It is data accessibility. Raw CMS files are not easy to navigate. Our analysis compiles, cleans, and visualizes the data so that the patterns are visible without a statistics background.

View the full nursing home analysis

Compare PE vs non-PE ownership data

Explore all 50+ fraud investigations