A “top hospital” is really two different scores, and they barely agree

By Paul Gruensfelder · Healthcare Growth Strategies

A hospital gets evaluated through several overlapping lenses: formal quality ratings, patient-experience scores, and public reputation. Those measures are related, but they aren’t interchangeable, and they often produce very different pictures of the same institution. That’s the practical problem I want to work through here, because “top hospital” gets used in marketing as if it were a single settled fact, and it isn’t.

A quick caution before the framework, because it’s easy to overstate. These lenses aren’t cleanly separated. CMS’s Overall Star Rating already gives patient experience about a quarter of the total score, and Leapfrog’s safety methodology includes HCAHPS communication and responsiveness measures alongside its clinical ones. So the “quality” ratings already contain some of what patients feel, and the “reputation” signals partly reflect real care. The lenses overlap. They just don’t line up.

I’m not building an HGS ranking here. I read the published systems and the peer-reviewed work on how they relate, and pulled out the question I think an operator should actually sit with: if the badge you win and the reputation you earn measure different things, which one are you competing on?

The formal quality ratings often disagree with each other

Start with the ratings, because even here “top” fractures the moment you look closely. Three prominent national systems illustrate the problem, and each answers a different question. (They aren’t the only ones; Healthgrades and others rate hospitals too.)

The CMS Overall Star Rating is the broad utility. It combines 52 measures into five groups, weighting mortality, safety, readmission, and patient experience at about 22% each and timely care at 12%¹. The April 2026 CMS release awarded five stars to 385 of 3,203 rated hospitals¹. Beginning with that release, hospitals initially assigned five stars were capped at four if they fell in the lowest Safety of Care quartile and reported at least three safety measures². The Leapfrog Hospital Safety Grade asks a narrower question, whether the hospital hurts people by accident, using up to 22 safety measures across process/structural and outcome domains, with each domain contributing half of the final grade³ (12 process/structural and 10 outcome measures by count, not an even split). U.S. News Best Hospitals ranks specialty excellence. In its 2026–27 methodology, risk-adjusted outcomes received greater weight across adult specialties; in cardiology, 41 outcome measures now determine 80% of the score, while expert opinion and patient volume no longer contribute⁴. That edition was released August 4, 2026⁵.

Now the part that should give any “top hospital” campaign pause. When researchers compared these systems across 2,384 acute-care hospitals using 2023 data, they disagreed constantly. CMS star ratings and Leapfrog grades differed by at least one grade about 70% of the time, and by two or more grades about 25% of the time⁶. Only 77 hospitals, about 3.2%, simultaneously held any U.S. News regional or national ranking, an A from Leapfrog, and five stars from CMS⁶. One caveat the authors are clear about: to compare letter grades against stars, they mapped both onto a common five-point scale, so the 70% figure reflects their analytical mapping rather than a direct equivalence the rating bodies established. An earlier five-system analysis reached a similar conclusion, with the closest agreement, still modest, between Leapfrog and CMS⁷.

This isn’t sloppiness. The systems use different time windows, hospital populations, risk adjustment, and definitions of the same event, so disagreement is what you’d expect. But the consequence is concrete: the badge you lead with is only as strong as the single system behind it, and a competitor can usually produce an equally valid badge from a different one. There’s no consensus “best.” There are several defensible bests, and they mostly aren’t the same hospitals.

Patient experience and public reputation are related, but distinct

Patients cannot directly observe risk-adjusted mortality. They can see how they were treated and what other people say, and by the mid-2020s those signals had moved to the front of how people research care.

In a 2024 Press Ganey survey of 1,000 healthcare consumers, ratings and reviews ranked first among the factors people weighed during the provider-research phase, ahead of facility reviews and physician referrals⁸. Respondents read about 4.7 reviews on average, and 83.5% said they at least occasionally used reviews to validate a referral they already had⁸ (the public write-up gives limited detail on sampling and fieldwork). Worth being precise about what that does and doesn’t establish: it shows reviews are central to how people research and validate providers. It does not establish that reviews outweigh network status, geography, urgency, access, or cost in the final decision, and the research on hospital choice specifically notes those other factors carry real weight.

Brand sits a layer above reviews. Monigle’s Humanizing Brand Experience, now in its ninth annual healthcare edition (May 2026, covering more than 200 brands)⁹, is a large annual brand-perception study. Monigle reported that rushed and impersonal care was a leading reason consumers lost trust, and that about a quarter changed providers after trust broke¹⁰. Press Ganey’s consumer research points the same way: compassion, communication, and availability were reported as the strongest drivers of recommendation and return¹⁷, while a separate analysis argues that trust begins forming through search, listings, reviews, and other touchpoints before clinical care occurs¹².

Notice what these consumer signals mostly capture. Communication. Responsiveness. Feeling known. Convenience. They’re weighted toward the experience of care, not its clinical result.

Reviews track experience more than they track safety

Here’s where the lenses meet, or rather where they don’t fully line up. When researchers compared Google star ratings against CMS ratings across thousands of hospitals, they found a weak but statistically significant overall relationship, with correlations around 0.23, and no significant association with the safety or readmission components after CMS’s 2021 methodology revision¹¹. Google ratings did carry some signal for patient-experience, mortality, and timeliness measures, so it’s not a clean split. But the pattern is clear enough: at the hospital level, crowd ratings align better with the parts of care patients can perceive than with risk-adjusted safety. That’s an association across hospitals, not evidence that any individual review accurately measures experience.

That pattern is roughly what you’d expect. Patients can directly evaluate communication, responsiveness, and aspects of the care environment, but they generally can’t assess risk-adjusted safety or readmission performance from a single encounter. So a hospital can be strong on safety and underrated online, or well-reviewed and unremarkable on outcomes, and neither case is a contradiction.

For an operator, that reframes the exercise. A CMS star rating and a Google profile operate as different communications assets: one is a formal quality summary, while the other is embedded in common consumer-search behavior. And a wall of positive reviews isn’t clinical validation, because reviews don’t reliably track safety or readmission measures. Treating either as a stand-in for the other is the most common and most expensive mistake I see in healthcare positioning.

Workforce conditions may be one lever that touches both

If quality ratings and consumer reputation mostly don’t line up, the practical question is whether anything moves both. Workforce conditions may be one of the clearer shared levers, though the evidence is associational and shouldn’t be oversold.

A systematic review of 55 studies found consistent positive associations between patient experience, patient safety, and clinical effectiveness across a range of conditions and settings¹³. At the hospital level, one study associated stronger patient-experience scores with lower readmission for heart attack, heart failure, and pneumonia, and with lower mortality for heart attack and pneumonia¹⁴, though these are observational findings and the direction isn’t clean, since a bad outcome also tends to produce a worse experience. On staffing specifically, a cross-sectional study of 146 South Korean hospitals found that general hospitals with the strongest nurse-staffing grade had the highest patient-experience scores across six measured domains, though the adjusted associations varied by hospital type¹⁵. And employee engagement has been associated with better patient safety performance, in a review that included only 15 studies, found substantial heterogeneity, and described the evidence base as early-stage, with several studies measuring safety through employee perception rather than objective adverse events¹⁶.

Read together, these studies support a plausible operational pathway, but not a proven causal mechanism. Patient experience is associated with some clinical outcomes; nurse staffing is associated with patient experience in certain settings; and employee engagement has been associated with aspects of safety. The evidence does not establish that changing any one workforce factor will reliably improve both sets of measures.

Even so, I’d treat workforce conditions as a lever worth disproportionate attention. That’s an HGS interpretation, not a finding from the studies. The reasoning is that workforce factors plausibly support both the experience patients report and the safety that shows up more slowly in the clinical data, and they’re among the most operationally actionable inputs a leader controls. Workforce health also has fewer standardized public scorecards than clinical quality does, which may be why it gets underweighted in strategy conversations that start from published rankings. Other shared levers plausibly matter too, including care coordination, access and scheduling, clinical process design, safety culture, discharge communication, and service recovery.

Why reputation weighs more the more “shoppable” the care is

The hospital data is the richest, but the logic doesn’t stay in the hospital, and this is where it should matter to anyone running a dental group, a specialty platform, or a med-spa rollup.

The pattern is directional: the more elective and consumer-paid the service, the more public reputation tends to carry the consideration decision, and the less the formal quality ratings are even visible to the buyer. A patient choosing a tertiary cancer center leans partly on referral and institutional reputation. A consumer choosing a dentist, an orthodontist, or an aesthetic provider is weighing reputation and reviews much more heavily. I’d treat this as a plausible operating hypothesis rather than a settled research finding, because the consumer surveys measure research behavior, not the full purchase decision across these very different contexts. (I’ve dropped a specific dental review-share figure that circulates in vendor content; I couldn’t trace it to an original survey with a named sponsor, sample, and question wording, and it doesn’t belong in a fact-driven piece without that.)

Two shifts are worth watching, framed as emerging rather than settled. Consumers often consult only a small number of review surfaces, so a thin or negative profile may remove a practice from consideration early. And AI-generated search summaries increasingly synthesize provider information, ratings, and reviews into a compact answer. Vendor consumer research suggests this behavior is emerging; Press Ganey reported that nearly one in five consumers had used AI to search for care¹⁷. But evidence on how often these tools surface a single recommendation, and how much that changes patient choice, is still developing, not something the current research establishes.

What I’d take from this as an operator

The operating implications may extend beyond hospitals to dental and specialty care, although the evidence base reviewed here is strongest for hospitals and the mix likely shifts with how elective and consumer-paid the care is.

Know which lens you’re competing on. Referral-driven, clinically complex care still trades partly on credentialed badges and specialty rankings. Consumer-paid, elective, and local care leans more on reputation and experience. Spending clinical-badge money to win a consumer-consideration game, or the reverse, is how budgets get wasted.

Don’t let one lens vouch for another. A CMS star rating by itself is unlikely to fill an elective service line, and a strong review profile does not protect an organization from a safety failure or a reputation event. Each has to be earned on its own terms, and claiming quality you can’t substantiate is both a credibility and a brand risk.

Put real weight on the plausible shared levers. The evidence reviewed here associates staffing with patient experience and engagement with aspects of safety; communication and access are especially relevant to experience. Before buying more demand, it’s worth checking whether the real constraint is capacity, access, or experience, because additional demand may worsen waits and experience when capacity is already strained.

Treat reputation as an ongoing operating function, not a campaign. Reviews are an important part of the digital front door now, refreshed continuously and increasingly synthesized by AI. That’s closer to quality management than to advertising.

None of this tells you which hospital is “best,” because that was never one measurement. Formal quality ratings, patient experience, and public reputation each capture something real, and none can stand in for the others. The strategic task isn’t to pick one definition of best. It’s to know which measure is being used, what it can legitimately tell you, and what it leaves out.