What PLACES is
PLACES (Local Data for Better Health) is a CDC project, run with the Robert Wood Johnson Foundation and the CDC Foundation, that publishes estimates of 40 health measures for every county, incorporated and census-designated place, census tract and ZIP Code Tabulation Area in the 50 states and the District of Columbia. It began in 2015 as the 500 Cities Project and expanded to the whole country in December 2020; the 2025 release is the current one.
The problem it solves
The national health survey behind it, the Behavioral Risk Factor Surveillance System (BRFSS), interviews around 400,000 adults a year by telephone. That is enough to say something about each state and some large counties, but nothing about a single ZIP Code, which might contain no respondents at all.
Small-area estimation
PLACES gets round this with a method called multilevel regression and post-stratification. For each measure CDC fits a statistical model to the whole national survey that relates a respondent's answer to their age, sex, race/ethnicity and education, the poverty level of their county, and effects for their county and state. That model gives a probability of the condition for every combination of those characteristics in every county. CDC then applies those probabilities to the Census count of people with each combination in each Census block, and adds the blocks up to a ZIP Code, place or tract. For counties it uses the annual population estimates instead.
The result is the prevalence the model expects for that population in that county and state. CDC runs the calculation 1,000 times with random variation to produce the 95% confidence interval reported with each estimate.
What that means in practice
- Two ZIP Codes in the same county with similar age, sex, race/ethnicity and education profiles will get similar estimates whatever is actually happening in them.
- Differences between areas mostly reflect differences in who lives there, plus the county and state context the model learned from the survey.
- A local programme, a new clinic or an outbreak leaves no trace unless it shows up in the state's survey responses. CDC: "users are cautioned against using these estimates for program or policy evaluations."
- The estimates are for adults (18 and over) only, because BRFSS surveys adults.
Validation
CDC reports that internal and external validation studies found moderate to strong correlations between the model-based estimates and direct survey estimates at state, county and place level. Its methodology page lists the published studies.