Where the numbers come from
Quality measures are calculated from two sources: the standardised clinical assessments that nursing homes complete for residents on a regular schedule, and Medicare claims data. Neither is an inspection. A surveyor does not observe or verify a quality measure the way they observe care during a health inspection.
That is worth holding onto, because the three components of the star rating have quite different provenance. Health inspections are produced by outside state surveyors. Staffing is drawn from payroll-based data rather than facility self-report. Quality measures rest substantially on assessments the facility itself completes and submits.
This is not an accusation about any facility — it is a structural feature of the data, and CMS audits and validates submissions. But it does mean the three components are not equally independent of the facility being measured, and quality measures are the component where that matters most.
Two populations, measured separately
CMS reports quality measures separately for short-stay and long-stay residents because a nursing home commonly serves both at once — people in rehabilitation after a hospital stay who expect to go home, and residents who live there.
Short-stay measures speak to recovery and rehabilitation: how residents' function changed, whether they returned home, whether they ended up back in hospital. Long-stay measures speak to ongoing residential care: mobility, pressure ulcers, falls with injury, antipsychotic medication use, and similar.
Reading the wrong set is the most common error here. If the expected stay is three weeks of rehabilitation, the long-stay measures describe a different group of people in the same building, and averaging your impression across both tells you less than reading the relevant half carefully.
Risk adjustment, and what it does and does not fix
CMS risk-adjusts several measures so that facilities serving sicker or more complex residents are not penalised simply for that. Without adjustment, a facility taking on harder cases would look worse than one that avoids them.
Adjustment helps, but it does not make every facility's numbers interchangeable. It corrects for measured characteristics, and no adjustment captures everything about who a facility serves. Treat measures as a signal to investigate rather than a settled verdict — particularly where a facility's population is unusual.
Check the denominator before you weigh the number
A percentage is only as stable as the number of residents behind it. In a small facility, one or two residents can move a measure by a large margin, and a figure that looks dramatic can rest on a handful of cases.
CMS suppresses measures where there is not enough data to report reliably, so a blank is a gap in the record rather than a zero or a finding. Where a measure is present, look at how many residents it covers before drawing a conclusion — a difference of a few percentage points between two facilities may not be a real difference at all.
How much they move the star rating
CMS builds the overall rating by starting from the health-inspection rating and then adjusting it using staffing and quality measures. Quality measures are an adjustment, not the anchor — which is why a facility can carry strong quality measures and still hold a middling overall star if its inspection record is weak.
When components disagree, that disagreement is usually the most informative thing on the page. A facility with a weak inspection history and strong quality measures is telling you something specific, and the inspection record is where the substance of it lives.