Introductory Biostatistics

of this, an adjusted rate is artificial; it does not reflect data from an actual
population. The numerical values of the adjusted rates depend in large part on
the choice of the standard population. They have real meaning only as relative
comparisons.
The advantage of using the U.S. population as the standard is that we can
adjust death rates of many states and compare them with each other. Any
population could be selected and used as a standard. In Example 1.13 it does
not mean that there were only 1 million people in the United States in 1970; it
only presents theage distributionof 1 million U.S. residents for that year. If all
we want to do is to compare Florida with Alaska, we could choose either state
as the standard and adjust the death rate of the other; this practice would save
half the labor. For example, if we choose Alaska as the standard population,
the adjusted death rate for Florida is calculated as shown in Table 1.9. The new
adjusted rate,

ð 1590 Þð 100 ; 000 Þ

407 ; 000

¼ 390 :7 per 100; 000

is not the same as that obtained using the 1970 U.S. population as the standard
(it was 770.6), but it also shows that after age adjustment, the death rate in
Florida (390.7 per 100,000) is somewhat lower than that of Alaska (396.8 per
100,000; there is no need for adjustment here because we use Alaska’s popula-
tion as the standard population).

1.3 RATIOS

In many cases, such as disease prevalence and disease incidence, proportions
and rates are defined very similarly, and the termsproportionsandratesmay
even be used interchangeably.Ratiois a completely di¤erent term; it is a com-
putation of the form

TABLE 1.9

Florida

Age
Group

Alaska Population

(Used as Standard) Rate/100,000

Expected
Number of Deaths
0–4 40,000 375.3 150
5–19 128,000 60.3 77
20–44 172,000 190.5 328
45–64 58,000 1101.5 639
65 þ 9,000 4397.9 396

Total 407,000 1590

18 DESCRIPTIVE METHODS FOR CATEGORICAL DATA