NLM logo

Finding and Using Health Statistics

Age-Adjustment

Sometimes, health statistics are used to compare how healthy two different groups of people are, or how healthy a certain group is during two different time periods. Since older people are more likely to get ill, and younger people are more likely to injure themselves, age adjustment can make studies more accurate.

Age is another confounding variable: something that impacts the people being studied but is not related to the health event being studied. To be able to better compare groups while adjusting for age, we use a process called direct standardization.

When we use direct standardization, we assume both groups have the same number of people. Then we calculate the expected number of deaths and death rates in both groups. By doing this, the two populations can be directly compared, independent of the age distribution of each group.

If the age-specific mortality rates (expected deaths) for the population of interest are not known, researchers can use indirect standardization, which requires only the observed deaths and age structure of the population of interest. Direct standardization is generally preferable because the populations of interest are compared to just one standard population, However, indirect standardization can be preferable with small sample sizes in different age groups.

When we use direct standardization, we assume both groups have the same number of people. Then we calculate the expected number of deaths and death rates in both groups. By doing this, the two populations can be directly compared, independent on the age distribution of each group. It is important to remember that age adjusted rates are not the actual rates of death or disease in the population – those are called “crude rates.” Age adjusted rates are only useful for comparisons to other populations.

Expected No. of Deaths in Standardized Populations and Age-Adjusted Death Rates

 

White

Black

Age

Rate per 1,000

Population
(standard)

Expected no. of deaths

Rate per 1,000

Population
(standard)

Expected no. of deaths

<1 yr

13.3

15,000

200

19.7

15,000

295

1-4 yr

0.8

50,000

40

1.7

50,000

85

5-17 yr

0.5

150,000

75

0.6

150,000

90

18-44 yr

2.5

325,000

800

3.7

325,000

1202.5

45-64 yr

18.0

210,000

3,780

25.0

210,000

5,250

>65 yr

70.0

200,000

14,000

77

200,000

15,400

Total
(age-adjusted rate)

24.8

761,000

18,895

29.3

761,000

22,322.5

This table shows the expected death rate and number of deaths in a standardized population, which provides the age-adjusted death rates.  Age-adjusted rates were calculated by dividing the expected number of deaths by the population (standard) and multiplying by 1,000.

"Population Analysis for Policies and Programmes: Comparison of Direct and Indirect Standardization." http://papp.iussp.org/. United Nations Population Fund (UNFPA), London School of Hygiene and Tropical Medicine and University of London, n.d. Web 5/26/2016

1

Previous Section | Next Section