Complete the calculation
Population distribution and deaths for country A
and country B.
age in Country A Country B
years
population population
Deaths Number percent death rate Deaths Number percent Death Rate
per 1000 per 1000
15 < 1028 538480 236 103004
44 - 15 1629 728363 388 149964
64 - 45 3839 341956 436 40699
65 ≥ 8358 162094 368 6715
Total 14854 1770893 1428 300382
Question 17
• A. what interpretation would you give to these data?
• b. why does a difference in the age structure of these two
countries make it preferable to standardize the rates before
making further comparisons?
• C. Epidemiologists commonly will age-adjust or standardize
crude rates. In general, what condition must prevail so that
it is necessary to adjust or standardize for age or other
factors?
Types of standardization
• The two most common types of adjustment
are termed
• the direct method and
• the indirect method.
The Direct Method
• To age-adjust using this method, a commonly
available standard population of known age
distribution should be selected, for example, the US
population for 1970.
• Complete the Figure 8 by calculating the expected
deaths for country A (columns a x b) and country B
(columns a x d) based upon the standard population
distribution of the [Link] 1970.
• The age-specific death rates can be obtained from
your calculations in Figure 7
Question 18
• a. To calculate the age-adjusted rates, add the expected deaths for each
state (columns (c) and (e).
• Divide the expected total deaths in each state by the total standard
population of the U.S. The age-adjusted death rate per 1000 of country A =
country B =
• b. What information does the age-adjusted rate convey?
• C. Are adjusted rates real or artificial numbers, i.e., should they be
considered real estimates of actual risk?
• d. what steps are performed when calculating an age-adjusted rate by the
direct method?
The Indirect Method
• It is not possible to use the direct method where adjustment
may be necessary but deaths by age group in the population
are not known; in such cases the indirect method is used.
• The following problem illustrates how to perform the indirect
method of adjustment.
• Calculate the crude death rates and compare the proportional
age distributions to establish that age-adjustment is necessary
• Then calculate the U.S. age-specific death rates per 1000 in
Figure 9.
• Then complete Figure 10 to get the expected number of
deaths in Colombia.
Figure 9. Population distributions and mortality by age
for Colombia. South America, 1974, and the US, 1970
United States Colombia Age in
years
Age – population population
specific
rate per deaths Percent number deaths Percent number
1000 of total of total
86215 8.4 17115336 n.a. 17.5 3780000 <5
16847 20.2 40902509 n.a. 27.5 5932000 5 – 14
157071 40.9 83270951 n.a. 44.9 9714000 15 – 44
166017 30.5 61921362 n.a. 10.1 2169000 45 +
9
192031 100.0 203210158 165812 100.0 21595000 Total
2
Crude death rate Crude death rate
per 1000 = per 1000 =
Figure 10. Calculation of expected deaths in Colombia if
US rates are applied
Expected deaths in Age-specific death Colombian Age in years
Colombia rate per 1000 in the population
U.S.
3780000 <5
5932000 5 – 14
9714000 15 – 44
2169000 45 +
21595000 Total
Question 19
• Calculate the age-adjusted rate by dividing the total expected deaths in Colombia by the
total population of Colombia. The age-adjusted death rate is…………
• Age-adjustment using the indirect method also enables us to calculate the STANDARDIZED
MORTALITY RATIO (SMR) To calculate the SMR simply divide the observed deaths for
Colombia by the expected deaths in that country SMRs are expressed as if the expected
number of deaths was equal 100. Thus to calculate the SMR for Colombia, simply multiply
the ratio of observed to expected deaths by 100.
• b observed deaths in Colombia divided by expected deaths in Colombia x 100
• c. What does an SMR greater than 100 mean?
• d. What does an SMR less than 100 mean?
• e. In what circumstances would it be necessary to use the indirect rather than the direct
method to adjust rates?
• f. What are the steps necessary to calculate an age-adjusted rate using the indirect
method?
• 17a. The total (crude) death rate is higher in co than country A.
• Because the age distribution is different for these states (there is a higher
proportion of older persons in Arizona) the crude rates are not suitable for
comparison. The age-specific rates can be compared but they do not
indicate whether the two states have similar proportions of males and
females in each age group.
• Country B has higher age-specific death rates for those below age 45 and
above age 65.
• The crude death rates give an erroneous impression concerning the risk of
death in the two states because of differences in the age structure of the
population in these states
•
• 17b. Because mortality is correlated with age, the crude death rates in
states (or countries) with a high proportion of older persons will
disproportionately emphasize the older age groups when comparisons are
made with states (or countries) having a lower percentage of old persons.
The problem may also be present if comparisons are to be made for two
populations selected from hospitals, medical records, or other sources in
which the age distributions differ .
• 17c. Age is not the only factor that might need to be adjusted.
Adjustment would also be necessary IF THE TWO POPULATIONS
DIFFER in their proportions of males/females, smokers/nonsmokers,
urban/rural residence severe/non severe illness, etc.
• IF THE CHARACTERISTIC IS KNOWN TO BE ASSOCIATED WITH THE
HEALTH OUTCOME being studied
• In other words, when comparisons are made between populations,
those populations should be as similar as possible with regard to
factors related to the disease, i.e., compare apples to apples rather
than apples to oranges.
• Unfortunately, epidemiologists cannot control the age, sex, place of
residence, etc., of human populations. Adjustment techniques
permit epidemiologists to mathematically eliminate the major
differences in the disease-related characteristics (confounding
variables) of the two populations .
Figure 7. Population distribution and deaths
for country A and country B
age in Arizona Alaska
years
population population
Deaths Number percent death rate Deaths Number percent Death Rate
per 1000 per 1000
< 15 1028 538480 30.4 1.91 236 103004 34.3 2.29
15 - 44 1629 728363 41.1 2.24 388 149964 49.9 2.59
45 - 64 3839 341956 19.3 11.23 436 40699 13.5 10.71
≥ 65 8358 162094 9.2 51.56 368 6715 2.3 54.80
Total 14854 1770893 100.0 1428 300382 100.0
• 18b. It represents the expected number of deaths divided by the total US
population, per 1000 The adjusted rate eliminates the age differences between the
two populations and shows us what would happen to the death rate in these
populations if both groups had the same age structure as the standard population
• 18c. Adjusted rates are artificial numbers that do not represent the actual rates
of death (or other health outcome) for these populations. Adjustment is performed
to eliminate the effect of differences in age (or other important characteristics)
between populations that we wish to compare,
• 18d.
• 1. Select a standard population of known age distribution
• 2. Calculate age-specific death rates for two populations to be compared
• 3. Calculate the expected number of deaths in each of the two populations by
multiplying the age-specific rates by the number of persons in that age group for the
standard population .
• 4. Add the expected number of deaths calculated for each of the two populations
and divide by the population of the standard population
Figure 9. population distribution and mortality by age
.for Colombia , south America 1974, and the U.S. , 1970
• 19a. The age-adjusted death rate for Colombia is 97,993/21,595,000 x
1000 3 4.5 per 1000, This is the death rate Colombia would be expected to
have if the age distribution of Colombia was similar to the U.S.
• 19b. SMR observed/expected deaths x 100165,812/97,993 x 100 169.2
• 19c. An SMR greater than 100 means that the standardized ratio has an
excess of mortality compared to the amount expected. In the above
example Colombia had a 69.2 % excess mortality ( 169.2 / 100 ) compared
to the U.S.
• 19d. An SMR below 100 means that the observed mortality is below
that expected among Mormons and non-Mormons was expressed in
relation to unity 1.0).The observed/expected death rates could have been
modified in order to present them as SMRs. Mormons would have an SMR
below 100 .
• 19e. The indirect method of adjustment is useful in two situations:
• first when the age-specific death rates are known for only one of the populations
to be compared;
• second, when the number of persons in an age group is very small, so that the
death rates might show large fluctuations due to the inclusion or exclusion of a
few deaths. Thus, the indirect method might also be useful in mortality studies
where a limited number of observations or cases is available for study.
• 19f.
• 1. Apply the age-specific death rates for a standard population to the number of
persons in each age group of another population to deter-mine the expected
number of deaths for each age group.
• 2. Add the expected number of deaths to obtain the total.
• 3. Divide the total expected deaths by the total population to obtain the
• 4. To convert to an SMR divide the observed by the expected number adjusted
death rate. of deaths and multiply by 100