clinicians.build · interactive · September 18, 2026 · built on CMS, Sep 17 2026
The Invisible Half
Washington is sending $50 billion to transform rural healthcare and demanding “accountability for results.” In the states with the most rural hospitals, CMS cannot currently rate most of those hospitals at all.
of critical access hospitals have no CMS star rating. For rural emergency hospitals it is 100% — all 41 of them.
CMS rates hospitals on up to 52 quality measures across mortality, safety, readmissions, patient experience and timely care. A hospital that reports too few gets no star rating — not a bad one, none. That is mostly by design: critical access hospitals are exempt from most mandatory inpatient quality reporting, and the brand-new rural emergency hospital designation has no reporting program spun up yet.
Which is fine, until the moment a program promises to transform rural care and measure whether it worked. Every dot below is one state. Across the horizontal axis: how much of its hospital base is rural-designated. Up the vertical: how much of that base CMS currently cannot rate.
One dot per state
States shown—
Correlation (r)—
Hospitals covered—
Drag that slider from 1 to about 5 and watch the relationship appear. It is not new signal — it is noise leaving. American Samoa, Guam and the Northern Mariana Islands have one hospital each; two of them sit at 100% unrated because a single unrated hospital is 100%. With every state and territory included, r = 0.40. Drop the places with fewer than five hospitals and it is 0.67. Four dots were doing the damage.
the 80/20 lens
The one thing that matters: CMS is about to spend $50 billion on the part of the hospital system it measures least. Nebraska, Kansas, Montana, the Dakotas and Iowa are where rural hospitals actually are — and where 46% to 72% of hospitals carry no star rating. If you are building the evaluation layer for any of this, the baseline you would want to measure improvement against does not exist yet for most of the buildings receiving the money. That is a product gap, not a complaint.
South Carolina, for scale
Yesterday’s $167 million went to a state with 3 critical access hospitals and no rural emergency hospitals — 5.6% of its hospital base, against a national picture where North Dakota is at 84% and Montana 82%. South Carolina’s hospitals report an average of 38.8 of the 52 measures; Nebraska’s report 18.6. By the measure on this chart, South Carolina is one of the most visible states in the country, at 13% unrated.
That is not evidence of misallocation, and this interactive does not claim it is — RHTP money went to all 50 states, roughly 20% of South Carolinians live rurally, and rural South Carolinians are largely served by hospitals that never took the critical access designation. It is evidence of something narrower and more useful: “rural hospital” is not one thing, and the federal designation a hospital holds determines whether the federal government can see its outcomes.
where this chart is thin
Unrated is not unmeasured-anywhere. Critical access hospitals report to MBQIP and to state programs; they are exempt from the Inpatient Quality Reporting program that feeds Care Compare stars. This chart measures federal visibility on one public scoreboard, not care quality and not total data collection.
Unrated is not bad. Nothing here says a single one of these hospitals delivers worse care. A missing star is a missing star.
Rurality is not the only route to invisible. Puerto Rico is the clean counter-example: 54 acute care hospitals, zero critical access hospitals, and still 89% unrated with an average of just 6.9 measures reported. Whatever is happening there is a territory-reporting story, not a rural one — and it sits far off the trend line, which is the honest reason to distrust a single-variable explanation.
One quarterly snapshot. CMS refreshes Care Compare quarterly; this is the May 2026 vintage. The rural emergency hospital count moved 38 → 39 → 41 across the last three vintages, so the designation is growing while its reporting program is not.
Correlation, plotted deliberately. The trend line is a least-squares fit across state aggregates, which is an ecological summary — it describes states, not hospitals, and cannot tell you what any individual hospital reports. The companion explorer does that instead.