Methodology

Every figure on this site comes from CDC's PLACES files (currently the 2025 release, data updated 4 Dec 2025), rebuilt daily. This page explains what the site adds and the rules it follows.

What PLACES estimates are

CDC fits a multilevel logistic regression to Behavioral Risk Factor Surveillance System responses (age, sex, race/ethnicity, education, county poverty, county and state random effects), applies it to Census block population counts and aggregates to ZCTAs, places and tracts; for counties it uses annual population estimates. Each estimate is the mean of 1,000 Monte Carlo draws and the 95% confidence interval is the 2.5th to 97.5th percentile of those draws. Nothing is measured locally: a ZIP Code's figure is what the national model expects for a population with that ZIP's demographic make-up, inside that county and state. CDC's methodology page and FAQ are the authority.

The comparison rule

CDC: "it is not adequate to simply compare point estimates. The confidence intervals should also be considered". This site applies one mechanical rule, in the build, to every comparison it makes:

• Two estimates with intervals (a ZIP against its county, an area against the US): higher when the area's whole interval lies above the reference interval, lower when it lies entirely below, otherwise similar.
• An estimate against a state figure, which has no published interval: above or below when the area's interval excludes the state figure, otherwise in line.
• A state figure against the US interval: above, below or within.
• Nothing is said when either side is not published.

The rule is deliberately conservative. Two areas whose intervals overlap may still differ; the model simply cannot show it. Comparing overlapping intervals is not the same as a formal test of difference, but it is the reading CDC asks for and it never claims a difference the intervals do not support.

State figures

PLACES publishes no state estimates. Following CDC's guidance for aggregating areas, each state figure is the sum over its counties of estimate × adult population, divided by the state's adult population, computed separately for crude and age-adjusted values and for each release. CDC's method calls for the population relevant to each measure; this site uses the adult population for every measure, which is an approximation for the measures defined on women aged 50–74, adults 18–64 or adults 65 and over. State figures carry no interval.

National figures

CDC's own United States row from the long-format county file of each release, with its interval, crude and age-adjusted.

Crude and age-adjusted

Crude prevalence is the estimated share of all adults in the area. Age-adjusted prevalence is standardised to the 2000 US standard population. CDC publishes age-adjusted figures for counties and places only; ZCTA pages show crude figures alone. The toggle on each page switches every table, flag and chart between the two.

When a measure is missing from the latest release

CDC does not publish some estimates in some releases: in 2025, Kentucky and Pennsylvania have only the five measures carried over from the 2022 BRFSS, and the seven health-related social needs measures exist only for the states that ran that survey module. Where the latest release has no value, the page shows the most recent release that has one, marks the release year in colour and compares it with the county, state and national figures of that same release. Measures with no value in any release since 2022 are shown as not published.

Estimates by release

The 4 releases since 2022 are listed for each area. CDC states that the modelling "does not support using the estimates to track changes at the local level over time": the 2022 and 2023 releases used 2010 Census geographies, the interval method changed in 2023, survey questions change, and the population base for ZCTAs and places is fixed at a census. The site therefore prints no change words, arrows or percentages between releases.

ZIP Codes, counties and places

"ZIP Code" on this site means a Census ZIP Code Tabulation Area (ZCTA), the unit PLACES estimates; ZCTAs approximate postal ZIP Codes but are not identical. A ZCTA is assigned to the county holding the largest share of its land area (Census 2020 relationship file) and lists the places it overlaps with their land shares. A place is assigned to the county containing the centroid CDC publishes for it, tested against the Census 1:20m county outlines; Connecticut's planning regions are the county equivalents from the 2024 release onward. Places are the Census incorporated places and census-designated places with 50 or more adults.

Thresholds and indexing

Every area CDC publishes gets a page. ZIP Codes and places with fewer than 500 residents are built but marked "noindex" for search engines, because their intervals are very wide and the pages add little beyond the numbers. Census tract estimates (>80,000 areas) are not republished. County lists on state pages and ZIP and place lists on county pages are ordered by point estimate for a chosen measure and always show the interval and the comparison flag; they are not rankings of health, which CDC says the data cannot support.

Why figures can differ from CDC's tools

• Release: CDC's map may show a newer release for some days before this site's daily probe and snapshot catch up.
• Fallback: for an area missing from the latest release this site shows the previous release, labelled; CDC's tools show a blank.
• State figures and comparison words are computed here, not by CDC.
• Rounding: CDC publishes one decimal; nothing is re-rounded.

Measure definitions on measure pages are quoted from CDC's measure-definitions pages, with CDC's burden summaries and references omitted. The site covers 40 measures, 32,520 ZCTAs, 29,923 places and 3,143 counties.