HCAHPS Scores by Hospital: How to Compare Facilities

By Jason Messerli, Strategic Advisor, Customer Strategy, NRC Health

You receive a forwarded email from your CEO. It is in response to a message they received from their longtime friend who happens to be an executive at your crosstown competitor. They routinely tease one another, but this email struck a chord. It references Care Compare and information that was just shared with them. They recently moved ahead of your hospital on three HCAHPS measures and the question from your CEO is “why”? What you do in the next 30 minutes matters. That conversation either becomes a productive one about peer comparison, or an unproductive one about why your team is “behind.” 

Let’s walk through the workflow I use when that printout shows up. This workflow includes: 

• Two CMS tools you should know how to use.  
• Five steps that turn a side-by-side comparison into an actionable comparison.  
• Three traps that make most facility-to-facility comparisons misleading. 

If your organization is trying to make sense of changes to HCAHPS, redesigned survey requirements, or peer benchmarking data, NRC Health’s CAHPS solutions are built to help patient experience teams connect public reporting back to what’s happening operationally. 

Key Takeaways 

• A lot of bad HCAHPS comparisons start with the wrong peer group. Two hospitals can look close on paper and still be operating in completely different patient environments. 
• Care Compare can provide high level views that compare hospitals to one another. For deeper comparison work you may need to download the Provider Data Catalog files. 
• Top-box gets most of the attention because it’s easy to recognize quickly. The bigger story sometimes shows up in the linear mean before it shows up in the public star. 
• Single-quarter HCAHPS comparisons can create overreactions. Four-quarter trends usually provide a more reliable view of performance against a peer group. 
• Hospitals with similar HCAHPS scores can have very different operational realities underneath them, which is why survey data works best as a starting point for investigation, not a final explanation. 

It means lining up the same measures for two hospitals, in the same reporting window, and looking at how they stack up. Top-box percentages. Linear mean scores. Star ratings. The reporting window has to match. 

CMS uses four quarters of rolling discharge data per release, so a comparison that mixes quarters isn’t really a comparison. 

CMS publishes the data through two free tools. Most people I work with only utilize one of them. Care Compare is the tool your patients see when they Google your hospital. The Provider Data Catalog is the tool your analyst should be downloading from when they want to do the work. 

A singular comparison is never a useful comparison. The hospitals worth comparing yours to are those that look like yours in the dimensions that drive patient experience scores. Case mix. Bed count. Payer mix. Geography. Service line concentration. 

Care Compare 

Care Compare is the consumer tool. It shows star ratings, top-box scores, and national averages for any hospital your patients (or your CFO) might look up. The interface lets you compare up to three facilities at a time, and refreshes quarterly along with the underlying HCAHPS data. 

For most board members, journalists, casual patient experience enthusiasts and patients, Care Compare is the only data source they’ll ever see. 

It’s worth knowing what your hospital’s page looks like to a non-expert. Their is immense merit in opening a fresh browser tab once a quarter and reading the page the way a patient deciding where to deliver their baby would. However, the internal dashboards your team is staring at are rarely the same as shows up there. 

Provider Data Catalog 

The Provider Data Catalog is the analyst tool. It hosts three HCAHPS datasets you can download as CSVs for offline work. 

• Patient survey (HCAHPS) — Hospital: hospital-level top-box, linear mean, and star ratings for every reporting facility. 

• Patient survey (HCAHPS) — State: state-level rollups. 

• Patient survey (HCAHPS) — National: national-level rollups for benchmark work. 

The CSV exports give you what the Care Compare interface won’t. You can filter by bed count, region, or service profile. You can run year-over-year trends across hundreds of hospitals. You can build the peer group your leadership team wants to see.  

The tradeoff is the data dictionary. Each variable name follows a convention you have to learn before the spreadsheet starts making sense. 

If your team has someone who lives in spreadsheets, the Provider Data Catalog is the single highest-leverage tool you have for peer benchmarking. If your team doesn’t, this is the year to make a case for one. 

Step 1: Define Your Peer Group Before You Pull Data 

Pick four to seven characteristics that matter to your hospital. Bed count is usually the place to start. Service mix comes next — and you have to define what that means before going further: OB or no OB, level of trauma designation, cardiac surgery on-site or not. From there, payer mix, region, system affiliation, and whether you’re a rural or urban facility all shape who belongs in the group. 

Step 2: Pull the Most Recent Quarter From the Provider Data Catalog 

Download the hospital-level CSV. Confirm the reporting period in the metadata before you start. 

CMS updates these datasets quarterly. The wrong quarter will produce confident-sounding answers about stale data.  

Step 3: Filter the Dataset to Your Peer Group 

Use the characteristics from Step 1 to filter down. 

A 250-bed Midwestern non-trauma community hospital should be comparing itself to other 200- to 300-bed Midwestern community hospitals without a level-1 trauma designation. Not the academic medical center across town with a 12,000-employee headcount and a quaternary case mix. The peer group is the accurate comparison. 

Step 4: Put Top-Box and Linear Mean Next To Each Other 

Most teams end up staring at top-box first because it’s the number everyone recognizes fastest. 

It’s the percentage of patients who answered “Always.” That number ends up everywhere, dashboards, monthly scorecards, board slides, and nursing reviews. 

The problem is that CMS doesn’t build the public star off top-box alone. 

Linear mean is the weighted score across the full distribution, and it’s what feeds the CMS star calculation. 

A peer hospital can sit a couple of points ahead of you on top-box and a couple of points behind on linear mean. When that happens, the action is moving in the middle of the response distribution, which top-box won’t catch. 

A Across my 2025 partner reviews at NRC Health, roughly 90% of hospitals had at least one care-coordination metric among the top three correlates of overall hospital rating.Care coordination has likely always been important to patients and something they struggle with a great deal. What was useful was confirming that the leverage map has shifted. Communication with nurses and communication with doctors aren’t the only places to look anymore. 

Step 5: Pull More Than One Quarter 

One quarter can send people into a spiral for no reason. 

I’ve seen hospitals celebrate a good reporting period that disappeared the next refresh. I’ve also seen rough quarters look a lot less alarming once the prior three were sitting next to them. 

The four-quarter view is where the pattern starts making sense. 

I recommend rolling 12-month trends rather than month-by-month evaluation. I showed a CMO a year-over-year trend where they improved every single year for five years, to which the CMO replied, ‘I had no idea we were improving like this.’

Short-term fluctuations disguise true performance. They cause leaders to question their approach or pivot strategy prematurely. The same dynamic applies to peer comparison. 

Last spring, I sat through a system-wide review of an integrated network with 11 hospitals. 

The COO asked the team to rank all 11 hospitals against each other on 5 HCAHPS measures and present the rankings as a leaderboard. The slide was clean. The story was clear. Three hospitals were “winning.”  

We pulled the case-mix and bed-count data the next day and rebuilt the slide. 

The hospital ranked sixth on the leaderboard, but turned out to be third in its peer group. The hospital ranked second, below the median for its peer group, just operating in a service mix that flatters HCAHPS scores. Two hospitals at the “bottom” of the original leaderboard moved into the top half once we adjusted who they were being compared to.  

The COO asked which version of the slide she should present to the board. 

The answer was the second one. The first slide would have led to a real intervention at the wrong three hospitals.

1. Comparing Hospitals That Take Care of Very Different Patients

OB-heavy hospitals and high-acuity medical centers do not land in the same place on HCAHPS measures. 

They’re caring for different patient populations under very different conditions. 

A hospital with a large maternity census will almost always look different on communication measures than one managing a heavier ICU or a complex surgical population. That doesn’t automatically mean one team communicates better than the other. 

This is where many hospital rankings start to get shaky. The scores look comparable, but the patient-mix underneath them is completely different. 

Research on nurse staffing and patient experience found that some relationships that appeared strong in broad hospital-to-hospital comparisons weakened when hospitals were tracked over time rather than compared in a single view. 

That’s part of the problem with broad leaderboard-style comparisons. The rankings can feel convincing while quietly comparing organizations operating in completely different environments.

2. Comparing Across Data Versions

The Provider Data Catalog refreshes quarterly. Each refresh includes new measure scores and may introduce methodology changes. 

As I covered in the March 2026 CAHPS Insider, the Care Transitions and Responsiveness of Hospital Staff composites are not currently displayed during the redesigned survey transition. Three new measures are expected to be added to public reporting at the October 2026 Care Compare refresh. 

A peer comparison that mixes pre-transition and post-transition data is comparing different things and calling them the same thing. 

There’s a quieter version of this trap that’s easier to miss. 

I walked through one example during a recent CAHPS Insider Live session. Nationally, about 63% of HCAHPS respondents are age 65 or older. During his 2025 reviews with hospital systems, I  saw organizations where that number had climbed to 75% or even 80%. Some of those same organizations had been close to the national average just a few years earlier. 

Their patient populations didn’t suddenly age that dramatically in a short period. The survey response pool changed, and in many cases, the collection method influenced who actually responded. 

If you’re comparing two hospitals and one of them recently changed survey modes, you may be comparing different respondent populations rather than different patient experiences.

3. Comparing Star Ratings Without Comparing the Measures Underneath

A 4-star hospital and a 3-star hospital can be 1 to 2 percentage points apart on linear mean for several individual measures, with the difference compounding through the clustering algorithm. 

The star is the headline. 

The measure-level scores are the story.  

If you walk into a comparison conversation with stars only, you’ll have a hard time explaining what to do about the gap. 

It can tell you whether your performance is moving relative to a peer group on measures that matter for VBP. It can tell you which underlying questions are dragging or lifting your composite scores. It can tell you whether a competitor’s recent improvement is broad or concentrated in one or two measures. 

It cannot tell you why those gaps exist. 

Two hospitals at 80% and 84% on Communication with Nurses can have very different underlying realities. One may have rounding routines that work well. Another may have a strong charge-nurse model. Another may have low staff turnover.  

The comparison surfaces the question. It doesn’t answer it. 

Answering it requires going underneath the survey data, into open-ended patient comments and real-time feedback. Our research team manager, Sarah Fryda, has been working on the longitudinal relationship between Net Promoter Score (NPS) and HCAHPS for some time now. “We know that NPS and HCAHPS are strongly related,” she said in a recent webcast. “And we can actually predict what future CMS HCAHPS scores will be.” for some time now. “We know that NPS and HCAHPS are strongly related,” she said in a recent webcast. “And we can actually predict what future CMS HCAHPS scores will be.”  

Organizations with strong HCAHPS performance usually aren’t far apart on NPS either. The timelines look different, but the patterns tend to point in the same direction over time. NPS reacts faster to what patients are feeling in the moment. HCAHPS takes longer to reflect operational habits that stay consistent across reporting periods. Sometimes the two measures drift apart for a quarter or two. They usually find each other again. 

That matters for facility comparison because if a peer hospital is outperforming you on a measure, the operational reasons usually show up faster in real-time feedback than in the next HCAHPS refresh. Our work on incorporating HCAHPS changes into the patient experience strategy, and on how NPS and HCAHPS work together to transform patient experience, walks through more of those patterns. 

Where Can I Compare HCAHPS Scores Between Two Hospitals? 

Care Compare lets you compare up to three hospitals side-by-side on HCAHPS top-box scores, star ratings, and other quality measures. For deeper analysis across multiple hospitals, use the Provider Data Catalog and download the hospital-level HCAHPS CSV file. 

Which CMS Dataset Has the Most Current HCAHPS Scores by Hospital? 

The hospital-level dataset at data.cms.gov/provider-data/dataset/dgck-syfz is refreshed quarterly with the most recent four-quarter HCAHPS reporting period. Confirm the reporting period in the dataset metadata before each pull. 

How Do I Build a Peer Group of Hospitals to Compare Against? 

Start with the characteristics that shape the patient experience environment in which your hospital operates. Bed count usually matters. Service mix does too. Geography, payer mix, health system structure, and rural-versus-urban differences can all change what a fair comparison looks like. 

Once those filters are set, narrow the Provider Data Catalog dataset down to hospitals that resemble yours. From there, look at top-box, linear mean, and public star ratings together. 

Why Do Two Hospitals With Similar HCAHPS Scores Have Different Star Ratings? 

Star ratings are assigned by a clustering algorithm that recalculates cut points each quarter based on current national performance. Hospitals with similar raw scores can fall on different sides of a cut-point. That’s one reason measure-level comparisons are more reliable than star-only comparisons. 

Can I Compare HCAHPS Scores Year Over Year? 

Yes. The methodology changed with the redesigned survey, effective January 1, 2025, so pre- and post-transition comparisons require care. Some composites are currently not displayed during the transition, and three new measures are expected at the October 2026 refresh of Care Compare. 

The first leaderboard is usually the least useful one. 

The hospitals that get the most value from HCAHPS comparison work are rarely the ones that react most strongly to a single quarter or a single public star. They’re the ones spending more time asking whether the comparison itself makes sense in the first place. 

Wrong reporting window. Wrong peer group. Different respondent populations. Different service mixes. Most misleading HCAHPS comparisons trace back to one of those problems long before anyone gets to the actual scores. 

That’s the part that people outside of patient experience sometimes miss. HCAHPS comparison work appears quantitative, but a surprising amount of judgment is baked into the setup. 

Which hospitals belong in the peer group? 

Which measure best addresses the question leadership is asking? 

Which trends are worth paying attention to, and which ones are mostly noise? 

The public score matters. The peer comparison matters too. 

Most operational improvement work starts after the comparison, not before it. 

For organizations looking beyond the public star and into the operational patterns shaping patient experience performance, NRC Health’s CAHPS solutions combine benchmarking, real-time feedback, and patient experience insights designed to support improvement work between reporting cycles.