Simple explanations of data-related concepts
These stories will demystify technical concepts and data terms using (almost) real-life situations. We’ll add one every so often, so please check back or sign up for our newsletter so we can let you know we’ve published a new story.
The neighborhood that eliminated asthma?
A community asthma coalition wanted to see what parts of the county had the worst problem with serious asthma among children. They mapped by zip code the total number of child asthma-related hospitalizations in the past year, because hospitalizations are a commonly-used measure of asthma that is uncontrolled. One zip code, actually a small city within the larger city, really stuck out on the map – it had only a handful of child asthma hospitalizations despite having a population size similar to other urban zip codes. The coalition’s planning committee was very curious how this city had managed to bring child asthma under such good control that hospitalizations were rare, and the committee proposed talking to that city’s leadership to learn more and hopefully transfer successful practices to the surrounding area.
When the committee presented the map to the full coalition, though, one member pointed out something the planning committee hadn’t thought of: the city with the very small number of child hospitalizations has a very old population compared to surrounding areas. There are very few child asthma hospitalizations in that zip code because there are actually very few children at all. Once the committee mapped the rate – number of child asthma hospitalizations per 1,000 children – instead of the raw number of hospitalizations, that zip code didn’t look any better than the others.
Key point: different geographic areas – for example, zip codes or counties – have different population sizes and characteristics. So if you want to compare and contrast something across geographic areas, you usually want to compare a rate per population, not a number of events or cases, for a specific period of time. To calculate a rate, divide the number of events or cases, like child asthma hospitalizations, by the number of people who could have had an event or case during the time period. When calculating a rate, we generally multiply by a factor of 100 – 100, 1,000, 10,000, or 100,000 – to get a number that’s not a fraction and is easier to look at. Use the same multiplier for every geography you’re comparing.
Example: Zip code #1 had 22 child asthma hospitalizations in 2014, and a total of 1337 children lived there in 2014; dividing 22 by 1337 gives us 0.01645. Zip code #2 had 14 child asthma hospitalizations among a total of 1051 children in 2014, or 0.01332 hospitalizations per child. Multiply by 1,000 children to get a rate of 16.5 asthma hospitalizations per 1,000 children in zip code #1, and a rate of 13.3 asthma hospitalizations per 1,000 children in zip code #2. These two zip codes can now be compared directly, and we can tell that child asthma hospitalizations are a bigger problem in zip code #1 than in zip code #2.