IPW Home

IPW Search

What's New

Hosted Sites





Injury Data - USA
National, Regional, State


Notes about Rates and Comparisons

Age-adjusted rates are useful when making a comparison of the rate of suicide in one defined geographic area (such as a county) with the rate in another. Age-adjusted rates are used when we compare a single summary rate in one county with a single summary rate in another. We don't want to use the crude death rate as the basis for this comparison.  Why not?

Most health conditions (suicide included) do not occur at the same rate in every age group. Certain age groups are at higher or lower risk. If we were to compare the crude death rate from suicide in one county with the crude death rate in another, any difference may be simply due to differences in the population age distribution. If a higher proportion of of the residents of one county is from an age group that is at greater (or lesser) risk of suicide, any difference in the crude suicide death rate may be due to a difference in age distribution not some other factor. If there is some other factor that increases or decreases the risk of suicide in a particular area, we want to know about that factor. We don't want the effect of age distribution to be the most influential thing that leads to our conclusion. As stated above, age adjustment is useful when making comparisons using a single summary rate in each county. Comparisons using a single rate, even if we use age-adjusted rates, are not as informative as some other methods of comparison. Comparisons of a single rate from one area with a single rate from another are merely convenient. A better way of making comparisons would be to compare the age-specific rate in one county with the age-specific rate in another (for example, if we were to compare the suicide rate among adolescents 15 - 19 years old in one county with the suicide rate of 15 - 19 year olds in another).

The crude death rate may be calculated more than one way. The methods are described below but aren't necessary to understand the example.

To remove the distorting effect of age on our single summary measure we need to create a synthetic rate through age-adjustment. The following simple example will illustrate:


We have the following information about two communities:

From census data we know that County A has a population composed of one-half young people and one-half old people.
County B has two-thirds young people and one-third old people.

Age-specific death rates per 1,000 population

County Young Old
A 4 16
B 5 20


From this information we can determine the crude death rates (CDR), as follows:

CDR in A = (1/2)(4) + (1/2)(16) = 10 per 1,000

CDR in B = (2/3)(5) + (1/3)(20) = 10 per 1,000

Notice that although the probability of dying is lower for County A in both age groups (4 vs. 5 and 16 vs. 20), the crude death rates are identical. This is because County A has an older population than County B.

To calculate the age-adjusted rate we apply each age-specific rate NOT to the population in the county from which the death rate arose (and certainly NOT to the comparison county) but to a THIRD "Standard Population". (There are any number of "Standard Populations" in use. The US population in 1940 was used to calculate the age-adjusted suicide rates. It doesn't matter much which population is used. It just matters that the age-specific rates are compared with the SAME population.)

For this example, our standard population is composed of one-third young and two-thirds old, arbitrarily chosen. Calculation of the age-adjusted death rate (AADR) is as follows:

AADR in A = (1/3)(4) + (2/3)(16) = 12 per 1,000

AADR in B = (1/3)(5) + (2/3)(20) = 15 per 1,000

We now have a comparison of dying for these communities that is not influenced by the age composition of these populations. Rates are age-adjusted in order to remove the effects of age to better the effect of other factors. However, it is ALWAYS necessary to first examine the overall crude rates because crude rates represent real events in the population. An adjusted rate gives an accurate COMPARISON but does not reveal the underlying raw data.

Two methods to calculate the Crude Death Rate

Method 1.
Number of all deaths Total population = Crude Death Rate

Method one is very straightforward. We only need knowledge of the total number of people who died of a certain cause and the total population. We usually have additional information. We usually know the number of people of who died of a specific cause, their age at death, and the population in the county who are of that age. For convenience, we usually group people by 5-year age groups (birth through age 4, age 5 through age 9, etc.).

Method 2.
The weighted sum of all age-specific death rates:

The proportion of the population birth - 4 age-specific death rate
The proportion of the population 5 - 9 age-specific death rate
The proportion of the population 10 -14 age-specific death rate
The proportion of the population 15 - 19 age-specific death rate
Crude Death Rate

Visit again soon. This explanation will evolve and possibly improve.

Back to Main Data Menu Page


rev. 01 November 1999 05:34 PM