Understanding the data

Not all quality numbers mean the same thing.

The Index draws on several federal quality programs. They sound similar — "colorectal screening rate," "blood-pressure control" — but each is a different measure, on a different population, built to a different specification. Here's how to read them, and what they don't cover.

The programs in this Index

Each profile labels which program its numbers come from. They are not interchangeable.

Health centers

HRSA UDS

Who reports
Every federally funded health center (FQHC), annually.
Population
All patients the center serves, every payer — the safety-net population.
Specification
HRSA Uniform Data System clinical measures (Tables 6B/7), self-reported.
Updated
Annual (prior calendar year, ~8-month lag).
ACOs

Medicare Shared Savings Program

Who reports
Medicare ACOs, annually.
Population
Assigned Medicare fee-for-service beneficiaries — older, Medicare-only.
Specification
MIPS CQM / eCQM specifications via the APP reporting set.
Updated
Annual (per performance year).
Hospitals

CMS Care Compare

Who reports
Medicare-certified hospitals, quarterly.
Population
Largely Medicare claims plus chart-abstracted and survey samples.
Specification
Hospital Compare measures — many risk-adjusted; HCAHPS patient survey.
Updated
Quarterly refresh.

Why you can't compare across programs

A health center's colorectal-screening rate and a Medicare ACO's colorectal-screening rate are not the same measure. They use different age ranges, different denominators and exclusions, and — most importantly — different populations. A health center reports on its whole patient panel, including young and uninsured patients; an ACO reports on attributed Medicare beneficiaries, who are older and more consistently engaged in care.

So a health-center median near 40% and a Medicare-ACO median near 80% on a similarly named measure reflect who is counted and how the measure is defined — not that one type of organization is "twice as good." We never rank one program against another, and neither should anyone reading this data.

The rule we hold ourselves to: compare a hospital to other hospitals, a health center to other health centers, an ACO to other ACOs. Comparisons within a program and population are meaningful; comparisons across them are not.

What this Index does not cover

These are federal quality programs, and they overwhelmingly reflect Medicare and Medicaid populations (HRSA UDS is the exception — it covers all of a health center's patients). That is the most complete quality picture available to the public.

Quality results for commercial / private-insurance members are largely not public. Commercial health plans report quality through proprietary measure sets maintained by private organizations, and those underlying results are licensed, not openly released. Where state All-Payer Claims Databases exist, access is restricted and inconsistent. So when a figure here says "Medicare," it means Medicare — we label the scope rather than imply a number covers everyone.

We only republish open, US-government public-domain data (HRSA, CMS, IRS). We never ingest or republish licensed or proprietary ratings.

Data vintage & refresh

Every figure carries the reporting period it was measured over, and every dataset is stamped with the date we retrieved it and a hash of the exact file. We refresh on each source's own schedule: Care Compare quarterly, UDS and ACO results annually. Older snapshots are archived so trends stay reproducible.

See sources, retrieval dates & file hashes →