Skip to content
American Factbook
Menu

Method

How the county health figures are made

How the pages of the County health hub read what their publishers put out, and what that record can and cannot say. The method every topic shares — where a figure comes from, snapshots, freshness and corrections — is on the site’s methodology page.

The source, and what a row is

These pages read CDC’s PLACES: Local Data for Better Health, County Data, 2025 release, published by the Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Division of Population Health, as the file retrieved on November 4, 2026. The program’s own page is www.cdc.gov/places.

One row of the file is one county, one measure and one value type, keyed by CDC’s location id, the county’s five-digit FIPS code. The file holds 3,144 counties in 51 states (the District of Columbia counted as one), 40 measures and 229,218 county rows, of which 66 carry no value. CDC’s row for the United States is kept apart and never counted as a county.

What a PLACES value is

Every value is a model-based estimate, not a count. The dataset’s description: “This dataset contains model-based county estimates.” CDC’s methodology page: “PLACES uses a multilevel regression and poststratification (MRP) method to generate estimates of each measure”, from its Behavioral Risk Factor Surveillance System survey and the Census Bureau’s population figures.

The estimate is driven by who lives in a county. CDC’s frequently asked questions: “The PLACES approach incorporates age, race or ethnicity, sex, education, and poverty to generate the model-based estimates”. So two counties alike in those come out alike, whatever else differs between them, and an estimate says nothing about why a county is where it is.

Every value is printed with its 95% confidence interval as CDC publishes it, and the pages call it an estimate. No page multiplies an estimate by a population: a percentage of adults times a count of adults would read as a count of people, and CDC publishes none.

Age-adjusted and crude

CDC publishes each measure twice for each county: a crude prevalence, the estimate for the county’s adults as they are, and an age-adjusted prevalence, standardized to one age distribution so that an older county and a younger one can be set side by side. CDC’s frequently asked questions say that “age-adjusted estimates … can be used for county- and place-level comparisons”. Every ranking, median and comparison on these pages uses the age-adjusted value, with the crude value printed beside it.

How counties are ranked and summarized

A ranking orders counties on one measure, highest age-adjusted estimate first. Counties with the same published value share a rank, and the next rank skips the places they take. The FAQ’s caution is on every ranking: “it is not adequate to simply compare point estimates”. Two counties whose intervals overlap may not differ at all, so each row prints its interval.

There is no overall ranking. The FAQ: “the data should not be used for ranking the overall health of any county, place, census tract, or ZCTA”. No page combines measures into a score.

A median is one of the published values, at CDC’s precision: with an even number of counties it is the lower of the two in the middle, so it is always a value CDC published and never an average of two. The national ranking prints the first 500 counties; every county is ranked in its own state’s list.

Which counties have which measures

Not every county has every measure, and nothing missing is filled in from another year or another source. CDC’s release notes: “Estimates for Kentucky and Pennsylvania were not available for measures based on the 2023 BRFSS . However, the five measures based on the 2022 BRFSS were carried over for Kentucky and Pennsylvania.”

In this file, Colorado has 33 of the 40; Florida has 33 of the 40; Kentucky has 5 of the 40; Oregon has 33 of the 40; Pennsylvania has 5 of the 40; South Dakota has 33 of the 40; Tennessee has 33 of the 40; Texas has 33 of the 40; Vermont has 33 of the 40; Washington has 33 of the 40; Wyoming has 33 of the 40. The fewest for any state is 5.

Some states have none of the health-related social needs measures; CDC’s footnote on its United States row for them reads “The estimates are based on 39 states and DC.” A county with too few adults has no estimate at all: “If the count is less than 50, the estimate is not reported” (the FAQ), and CDC’s footnote on each such row is printed as it is. In this file 1 county has no published estimate and 845 lack at least one measure.

Connecticut, Alaska and the counties

The counties are CDC’s. Connecticut is on its nine planning regions, as the Census Bureau now counts it; the release notes of the 2024 release: “First release that uses 2020 Census population data and Connecticut planning regions for county equivalents”. Alaska’s are its current boroughs and census areas, with Chugach and Copper River where older files have Valdez-Cordova. No figure is moved from one geography to another.

The population on a county page is the Census Bureau’s figure CDC printed beside its estimates, and is cited as that. A population is a count of people, so one under 11 is not printed: the page says only that the factbook left it out, never “fewer than 11”, and a county’s adults are left out with its total. CDC publishes these figures, so the cut is the factbook’s, not CDC’s. No page prints a state’s or the nation’s population, so no county’s can be worked out from the others.

What this data cannot show

  • What any county measured. Every value is a model’s estimate from who lives there.
  • Whether a program worked. The dataset’s description: “Because the small area model cannot detect effects due to local interventions, users are cautioned against using these estimates for program or policy evaluations.”
  • Change over time. The FAQ: “the time is not included in the model as a variable, so the data cannot be used to access trends for counties.” Earlier releases are not loaded, and no page draws a trend from PLACES.
  • Overall health. There is no composite score and no healthiest county.
  • The measures CDC did not publish for a state or a county, which are listed as not published and never filled in.
  • Deaths. PLACES has none; death rates by county are a later source of this hub’s.

The hub’s own pages start at county health.

What this data cannot show

  • PLACES values are CDC's model-based estimates, predicted from who lives in a county, not counted there; CDC says they should not be used to rank the overall health of an area or to evaluate a program.
  • CDC says PLACES cannot show trends for counties, so these pages show one release and no change over time.
  • Kentucky and Pennsylvania counties have only the measures CDC carried over from an earlier survey year, some states have no social-needs measures, and a county with too few adults has none; nothing is filled in for any of them.

Sources

Data retrieved