Start with the exact certified provider
Search Medicare Care Compare by name, address, ZIP code, and distance. Confirm the legal name, address, phone, certification number, ownership, bed count, Medicare and Medicaid participation, and any hospital or chain relationship before interpreting data.
A campus may contain independent living, assisted living, memory care, and a certified nursing home under different legal entities. Care Compare information for the nursing home does not rate the assisted-living building next door. A recent sale or name change can also make older records look disconnected when they are not.
Save the date of your research. CMS data update on different schedules, and inspections or ownership changes may appear after the first search. Record the reporting period for every measure rather than labeling it simply “current.”
Treat the overall stars as a doorway
CMS created the Five-Star Quality Rating System to help people compare nursing homes and identify questions. It provides an overall rating plus separate health-inspection, staffing, and quality-measure ratings. CMS also cautions that no rating system covers every consideration and recommends using the site with visits and other sources (CMS).
Do not read four stars as an 80 percent grade or a promise of care. The overall rating is constructed from component rules rather than a simple average. Open every domain. Note when a rating is unavailable, suppressed, downgraded, or based on limited data.
The health-inspection rating starts the overall calculation. Staffing and quality-measure ratings can adjust it under CMS rules. A facility can therefore have the same overall stars as another while its component pattern is very different. One may have stronger inspections and weaker measures; another may have the reverse.
Read health inspections as narratives
Open the full standard, complaint, infection-control, and follow-up reports. For each deficiency, record:
- inspection type and date
- regulatory requirement
- facts observed, interviewed, or documented
- residents affected
- severity and scope
- immediate-jeopardy or substandard-quality designation, if any
- correction date and follow-up evidence
- recurrence in later surveys
- whether the citation is disputed or final
- enforcement connected to it
Count alone hides seriousness. Several low-level documentation findings are not equivalent to one finding involving actual harm or immediate jeopardy. Conversely, a corrected severe finding should not be erased from history; examine what changed and whether the pattern returned.
Read what the facility failed to do, not just the category heading. Connect it to the prospective resident’s needs. A transfer, wound, medication, nutrition, communication, abuse-prevention, or discharge finding may be especially relevant to one person and less directly relevant to another.
Distinguish surveys, complaints, and allegations
A standard survey samples facility operations at a point in time. Complaint and facility-reported-incident investigations can occur separately. Public data may be added later as investigations and revisits conclude (CMS Provider Data).
A complaint count does not establish that every allegation was substantiated. A deficiency is an official finding, but it may be under dispute. A plan of correction is the facility’s proposed or required response, not by itself proof that the correction endured.
Ask the state survey agency whether more current records exist and how to request them. Ask the Long-Term Care Ombudsman about publicly shareable information and the complaint process. Protect resident confidentiality; do not seek private case details that cannot lawfully be disclosed.
Understand the health-inspection star context
CMS’s health-inspection star cut points are based on relative performance within a state and are recalibrated. The current technical material explains that facilities are distributed into rating groups based on inspection deficiency scores within each state (CMS).
This means a health-inspection rating is not a national absolute safety certificate. State survey practices and relative cut points matter. Use the star to locate the underlying history, then read severity, scope, date, recurrence, and correction.
Do not compare an old screenshot with today’s rating without noting methodology and data refresh changes. CMS periodically revises measures and technical rules.
Examine staffing by role and period
Care Compare reports staffing hours per resident day, weekend staffing, nurse turnover, RN turnover, and administrator departures. Medicare explains that the staffing measures are averages and do not show who was present at a particular time or how much care one resident received (Medicare).
Record separately:
- registered-nurse hours
- licensed-practical or vocational-nurse hours
- certified-nursing-assistant and aide hours
- total nurse staffing
- weekend levels
- resident census used in the denominator
- turnover and administrator changes
- reporting quarter or year
- missing, unverifiable, or penalized data
An average can hide day-to-night and weekday-to-weekend variation. Administrative nursing time may also be included. Ask the facility how many direct-care workers in each role are physically present on the relevant unit during every shift, what vacancies exist, how agency workers are used, and how call-outs are covered.
Connect staffing to the person’s tasks
Higher hours do not automatically mean the right skill is available at the right time. Map each essential task to role, frequency, timing, number of workers, equipment, expected response, and backup.
For example, determine whether two trained workers can perform a nighttime transfer, whether an RN is available for a complex clinical change, who supports meals and hydration, and how therapy schedules align with the person’s tolerance. Ask how workers learn communication preferences, pain signals, routines, and accepted risks.
Turnover can affect continuity, but a percentage does not reveal why people left or current team stability. Ask how long the administrator, director of nursing, unit managers, medical director, infection preventionist, social-services leader, and direct-care workers have served.
Read quality measures as defined populations
Care Compare’s quality measures address selected outcomes and processes. Medicare separates short-stay and long-stay measures and describes them as averages for defined resident groups (Medicare).
For each measure, ask:
- Is it short-stay, long-stay, claims-based, or assessment-based?
- What population and exclusions apply?
- Which dates feed the result?
- Is higher or lower better?
- Is the sample sufficient?
- Does it measure a process, condition, utilization event, or outcome?
- How directly does it relate to this person’s needs?
A measure can identify a useful question without proving cause. Residents differ in health and preferences; measures use adjustment and specifications, but no model captures every relevant difference. Facilities also report assessment data used in some measures, while other measures use claims.
Avoid selecting isolated favorable numbers and ignoring the domain. Look for consistent patterns across periods, measures, inspections, staffing, and visits.
Check enforcement and special designations
Review fines, payment denials, termination actions, directed plans, temporary management, or other remedies shown by CMS or the state. Identify the triggering finding, date, status, amount or restriction, correction, and recurrence.
CMS maintains a Special Focus Facility program for nursing homes with a history of poor survey performance requiring additional attention (CMS). Check the current list and candidate information rather than relying on an old news story.
An enforcement action is serious evidence, but interpret its scope and current status. A fine does not automatically prove that every present service is unsafe; absence of a recent fine does not prove excellent care.
Research ownership and operational control
Record direct owner, chain, management company, real-estate relationship, related organizations, and recent changes shown in official data. CMS publishes chain-level performance information for groups sharing ownership or operational control (CMS Data).
Chain data can reveal broader patterns, but it should not replace facility-specific evidence. A local home’s leadership, workforce, building, resident population, and inspection history may differ from chain averages.
When ownership recently changed, ask what changed in administrator, nursing leadership, staffing vendor, therapy, pharmacy, policies, debt, renovations, or certification. Determine which historical findings attach to the location and which entity was responsible at the time.
Compare hospital use and other available indicators carefully
Care Compare and CMS datasets may include hospitalization, emergency-department, successful community-discharge, infection, vaccination, antipsychotic, pressure-injury, function, or other measures depending on current specifications.
Do not assume lower use is always better. A necessary hospital transfer can protect a resident; avoidable transfers can signal weak prevention or coordination. Ask how the facility recognizes acute change, contacts clinicians, communicates with hospitals, reconciles medicines, and accepts residents back.
Similarly, a medication measure does not establish why one person receives a drug or whether it is appropriate. Individual treatment questions belong with the prescriber and care team.
Visit more than once
Public data cannot show today’s call response, meal assistance, staff communication, odor source, noise, privacy, respect, or whether the actual unit fits the person. Visit at different times, including a meal, evening, or weekend when possible and permitted.
Observe:
- whether workers respond and explain before care
- whether residents have water, mobility aids, hearing aids, glasses, and call access
- assistance with meals, toileting, positioning, and movement
- privacy, visitors, phones, outdoor access, and meaningful activity
- staffing visible on the actual unit
- cleanliness, maintenance, temperature, lighting, and emergency routes
- respectful response to distress or requests
One observation is not proof of a pattern. Record date, time, place, and what occurred. Ask for an explanation and compare it with later visits and records.
Include the resident’s account
Speak privately with the person who may live there and, where appropriate, current residents who choose to talk. Ask what matters, what feels respectful, how quickly help arrives, whether food and routines work, and whether complaints can be raised safely.
Do not treat cognitive impairment as a reason to dismiss an account. Communication may require time, accessible questions, hearing or language support, and corroboration. Also do not turn a statement into a public accusation without verification.
The Ombudsman can help residents understand rights and work through complaints. Its role is resident-directed advocacy, distinct from state survey enforcement.
Create a dated decision record
For each candidate, preserve:
- provider identity and certification
- component and overall ratings with refresh dates
- full inspections, complaint findings, corrections, and enforcement
- staffing roles, periods, averages, turnover, and facility explanation
- quality-measure definitions, periods, and relevance
- ownership and recent changes
- repeated observations and resident preferences
- task-level fit, contract, complete price, payer acceptance, and discharge rules
- unresolved contradictions and who must answer them
Classify essential needs as verified fit, conditional fit, unresolved, or not fit. A high rating cannot cure an unresolved need for specialized equipment, dialysis, ventilator support, dementia practice, language access, two-person transfers, or a particular payer.
Recheck after admission
Research does not end at selection. Compare the care plan with actual delivery, bills, staffing, health changes, and the resident’s experience. Check refreshed inspections and enforcement periodically and after a serious incident or leadership change.
Care Compare is most useful when it sharpens questions and reveals official evidence. It is least useful when a single star becomes a substitute for reading, visiting, listening, and verifying whether this nursing home can support this person now.
Sources
- Centers for Medicare & Medicaid Services: Five-Star Quality Rating System
- Medicare: Nursing-Home Staffing
- Medicare: Nursing-Home Quality Measures
- CMS Provider Data Catalog: Nursing-Home Data Sources
- Centers for Medicare & Medicaid Services: Nursing Homes
- CMS Data: Nursing Home Chain Performance Measures