Starfile / Guides / CMS 5-star rating

What the CMS 5-star rating really measures

CMS rolled out the 5-star system in 2008 specifically to help families compare nursing homes. It works, but not the way most families think it does. The overall star is a weighted blend of three different measurements — and the three can tell you very different stories about the same facility.

The three components

1. Health inspection rating.Derived from the facility's last three annual standard surveys plus any complaint investigations within that window. CMS converts citations into a weighted deficiency score (more serious = more weight), then ranks all facilities nationally. The top 10% get 5 stars; the bottom 20% get 1 star; everything between is bucketed.
2. Staffing rating. Based on nurse hours per resident per day (HPRD), case-mix adjusted for resident acuity. Calculated from PBJ (Payroll-Based Journal) data that facilities submit quarterly. There are two staffing sub-ratings — total nurse and registered nurse — and both factor in.
3. Quality measures (QM) rating. Based on ~15 MDS-reported metrics: pressure ulcers, falls, UTIs, catheter use, antipsychotic use, functional improvement, etc. Split into long-stay and short-stay measures. Self-reported by facilities, which is the obvious weakness.

How they combine into the overall star

The algorithm is not a simple average. It starts with the health inspection rating as the base, then applies upward or downward adjustments:

• +1 star if staffing is 4 or 5
• −1 star if staffing is 1
• +1 star if QM is 5
• −1 star if QM is 1
• Overall cannot exceed health inspection + 1, and cannot be 5 unless health inspection is at least 4

So the health inspection rating is effectively the ceiling. A facility with a 2-star health inspection can never earn a 5-star overall, regardless of how well it staffs or how clean its MDS reporting looks.

Why the overall star can mislead

Three systematic issues to know about:

The QM rating is largely self-reported

Facilities submit MDS (Minimum Data Set) assessments on their own residents. CMS has caught systemic under-reporting of falls and pressure ulcers in audits; when the OIG has independently validated facility-reported data against medical records, the rate of under-reporting is material. A facility with a 5-star QM rating and a 2-star health inspection rating is probably a 2-star facility — the QM data is noise, not signal.

Staffing is case-mix adjusted — sort of

The case-mix adjustment uses the facility's own MDS-reported resident acuity to normalize staffing expectations. That means a facility can game both variables simultaneously: over-report acuity to lower the staffing bar, while also reporting favorable quality outcomes. The case-mix-adjusted HPRD is not the number you actually want.

On Starfile's facility pages, the reported (not adjusted) HPRD is the more honest number. Below 3.5 total nurse HPRD is a severe staffing shortage regardless of what the case-mix adjustment produces.

The rating is a rolling three-year blend

A facility that was terrible two years ago but has genuinely improved still carries the old citations in its rating. Conversely, a formerly 5-star facility that was recently acquired by a chain with a bad track record may still show the old rating while already deteriorating. Look at the most recent cycle specifically, not the blended average. Starfile shows all three cycles separately on each facility page.

What the rating is actually good at

Despite the above, the 5-star system reliably distinguishes the worst ~15% of facilities from the top ~15%. Research consistently shows 1-star and 2-star facilities have materially higher resident mortality, more hospitalizations, and more abuse/neglect citations than 4-star and 5-star facilities. The rating is noisy in the middle of the distribution; it is accurate at the tails.

In practice that means the 5-star system is a useful filter — eliminate the 1-star options — but not a useful ranker. Pick your top 3-5 options by geographic convenience and minimum rating threshold, then distinguish between them with citation reading, staffing inspection, and in-person visits.

Gaming the system

The New York Times ran a large investigation in 2021 (Silver-Greenberg & Gebeloff) showing systematic inflation of quality data at many facilities. CMS has since tightened MDS audit requirements, but the incentive structure hasn't changed: a higher rating means more referrals from hospitals (which themselves are penalized for readmissions), which means more revenue.

The metric least vulnerable to gaming is the health inspection rating, because surveyors write the citations from on-site observations, not from facility-submitted data. When the three components disagree, weight the health inspection rating heaviest.

Reading a real rating

An example: a facility at 3-stars overall, 2-stars health inspection, 4-stars staffing, 5-stars QM.

• The facility is closer to a 2-star facility than a 3-star one. Health inspection is the base; everything else is propping it up.
• 4-star staffing combined with 2-star inspection usually means either (a) the staff are present but poorly trained or (b) there's an administrative dysfunction — policies on paper, but not followed.
• 5-star QM with 2-star inspection is a classic signal of MDS over-reporting. A genuinely excellent facility on quality measures also has clean inspections.

What Starfile publishes

For each facility, Starfile shows the overall rating, all three component ratings, the reported and adjusted staffing HPRDs, every F-tag citation in the window, and the inspection cycle history with state and national averages. All fields come directly from the CMS Provider Data Catalog and refresh monthly.

You can use the Special Focus list, abuse watchlist, and your state's worst-rated page as negative filters — facilities to strike off your list before you start the real evaluation.

Algorithm details from CMS's “Design for Nursing Home Compare Five-Star Quality Rating System: Technical Users' Guide.” Critique informed by ongoing peer-reviewed research and CMS / OIG audit findings.