Number 6—Revised

March 1995
Revised to clarify computations in tables III and V

From the CENTERS FOR DISEASE CONTROL AND PREVENTION/National Center for Health Statistics

Direct Standardization
(Age-Adjusted Death Rates)
Lester R. Curtin, Ph.D. and Richard J. Klein, M.P.H.

Introduction generate more events (deaths) than a smaller group simply

because of its size. Therefore, to compare relative differences
Most population-based mortality objectives and
in mortality among population groups, or for a given
subobjectives in Healthy People 2000 are tracked using
age-adjusted rates from the National Vital Statistics System population group over time, the number of deaths must be
(appendix table I). The exceptions are deaths from related to the ‘‘population at risk’’ of dying to produce death
alcohol-related motor vehicle crashes, all motor vehicle rates. The population of interest may be the entire population
crashes, and work-related injuries (objectives 4.1, 9.3, and of an area or a population subgroup (for example, people in
10.1), which are monitored with crude death rates from other a certain age group).
data systems. In addition, objectives that refer to specific age The simplest death rate is the crude death rate (CDR),
groups are tracked with age-specific rather than age-adjusted defined as the total number of deaths divided by the midyear
rates. population. CDRs are usually expressed as a rate per 1,000
Although the age-adjusted death rate (ADR) is one of or 100,000 population. CDRs for individual age cohorts,
the most frequently used indexes of mortality, there is often called age-specific death rates (ASDRs), are the ratio of the
confusion concerning the basic concepts of its construction, number of deaths in a given age group to the population of
use, and interpretation. Some of the persistent issues include that age group, again usually expressed per 1,000 or 100,000
the appropriateness of the ADR as a summary measure, the population.
validity of comparisons between ADRs, the method of To compare the relative health of population groups or
calculation, and the appropriateness of alternate summary to assess change in mortality over time, two criteria must be
measures. considered. First, rates should relate the number of events to
the population at risk. Second, because many health
outcomes vary by age, the effect of the population’s age
Why use age-adjusted death rates? distribution must be taken into account.
The total number of health events (for example, the Although it does relate the number of events to the
number of deaths) occurring in a population is useful for population, the crude rate does not take into account the age
determining the magnitude of a public health problem. distribution of the population. As such, it is not an
However, the absolute number of deaths is seldom useful for appropriate measure for comparing differences between
comparisons between population groups (for example, population groups or for assessing change in mortality over
comparing males and females) or for comparing trends. time. Because death rates for most diseases generally
Assuming equal risk, a larger population group will tend to increase with age, a population group with a relatively

U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES

Public Health Service
Centers for Disease Control and Prevention
National Center for Health Statistics CENTERS FOR DISEASE CONTROL
AND PREVENTION
Table A. Crude death rate comparison
Community A Community B

Age Deaths Population Rate1 Deaths Population Rate1

0–34 years. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 1,000 20 180 6,000 30

35–64 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 3,000 40 150 3,000 50
65 years and over . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360 6,000 60 70 1,000 70

Total. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500 10,000 50 400 10,000 40

1Per 1,000 population.

Table B. Age-adjusted death rate calculation

Community A Community B

Standard Rate × Rate ×

Age population Rate1 population Rate population

0–34 years. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3,000 20 60 30 90

35–64 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3,000 40 120 50 150
65 years and over . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4,000 60 240 70 280

Total. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10,000 42 420 52 520

1Per 1,000 population.

younger age distribution will tend to have fewer total deaths In contrast, only 10 percent of community B’s population is
from a given disease than a comparably sized population in the oldest age group. Because the death rate is highest in
group with an older age distribution. Similarly, even if the the oldest age group, there are fewer total deaths in
age-specific risks of dying for a group remain unchanged community B.
between two time points, the number of deaths will increase
as the population ages.
Direct standardization
As an alternative to crude rates, ASDRs can be used.
The most comprehensive and reliable method of comparing There are two basic methods of standardization, or
death rates over time or between different population groups age-adjustment; both were introduced in the 19th century.
is to compare individual ASDRs for all age groups of These two methods have become known as the direct and
interest. However, this method often requires an extremely indirect methods. (Indirect standardization is discussed in a
large number of comparisons and tends to overwhelm both later section.) When the direct standardization method is
the investigator and the intended audience. applied to ASDRs, the resultant summary index is called the
Because the crude death rate is not appropriate and ADR. Two assumptions are made when this index is
ASDRs provide too much detailed information, a summary computed for a population: The population’s observed
measure that controls for a population’s age distribution is age-specific rates are assumed to be valid, and the age
needed. A commonly used measure is the ADR (1,2). distribution of the population is assumed to be that of a
Age-adjusted rates were developed in 1841 for the standard, or reference, or population.
analysis of mortality data (3). In the 19th century, mortality Table B illustrates the calculation of the ADR using the
data provided the most useful, and often the only, measure hypothetical data from table A. Specific computational
of the health of a population. About that time, it was formulae for the ADR are given in the technical appendix.
observed that a community could have ASDRs that were To calculate the ADR, the standard population and the
lower than the national average at each age interval, but, age-specific death rate for each age interval are multiplied
because the community’s population was older, the overall and these products are summed. In this example, the total
CDR was higher for the community than for the Nation. for community A is 420. This sum is divided by the total
Table A presents a hypothetical comparison to illustrate standard population (10,000 in this case) to obtain the ADR.
this situation. As with crude rates, the ADR is usually expressed in terms
In each of the two comparably-sized communities in of a rate per 1,000 or per 100,000 population. Thus, the
table A, the ASDRs increase with age; at each age the ADR for community A is 42 deaths per 1,000 population
age-specific rates are higher for community B than for and the ADR for community B is 52 per 1,000. Note that,
community A. Yet, the total (or crude) death rate is lower for although the crude rate for community A was larger than that
community B. This occurs because community A is an older for community B, the ADR for community A is smaller than
population; 60 percent of its citizens are 65 years and over. the ADR for community B. This is consistent with each of

2
the age-specific rates for community A being smaller than ‘‘When not to adjust.’’) Although the magnitude of the
those of community B. ADRs may be greatly affected by the choice of a standard
Because of the method of computation, the age-adjusted population, relative mortality, as measured by trends, race
rate is often interpreted as the hypothetical death rate that ratios, and sex ratios, is generally unaffected (8).
would have occurred if the observed age-specific rates were Despite this, controversy continues over which standard
present in a population whose age distribution is that of the population to use when age adjusting death rates to measure
standard population. It is very important to realize that the temporal changes in cause-specific mortality. Examination of
ADR is an artificial measure whose absolute value has no the issue shows that standard populations generally yield
intrinsic meaning. The ADR is useful for comparison only a small effect on trend comparisons by cause of death
purposes only, not to measure absolute magnitude. (To (9). Thus, any standard population is adequate so long as
compare absolute magnitude, crude rates are used.) It is also comparison populations are not very ‘‘unusual’’ or
important to note that in order to compare two age-adjusted ‘‘abnormal’’ with respect to the population under study (10).
rates, the same standard population must have been used. This means that the age distribution of the standard
population should be somewhat similar to the population of
Selection of a standard population interest.

After the decision to use an ADR is made, the standard

population must be selected. There are two basic types of
Small number issues
standard populations, internal and external. Internal standards One problem with ADR is that rates based on small
are created from the data to be used in the analysis; for numbers of deaths will exhibit a large amount of random
example, the average age distribution of all populations to be variation. (See the technical appendix for more detail.)
compared. The use of an internal standard has certain Therefore, if the number of deaths is small, mortality data
statistical advantages for the ADR (4). However, if an should be aggregated over a number of years, or several
internal standard is used, the results cannot be directly small geographic areas must be combined into larger areas
compared to other studies that use adjusted rates computed before computing the ADR (11). A very rough guideline is
using a different standard population. that there should be at least 25 total deaths over all age
External standards are standard populations drawn from groups.
sources outside the analysis. For example, the National
Center for Health Statistics (NCHS) typically uses a standard
When not to adjust
based on the 1940 United States population (5). This U.S.
standard population is usually given in terms of a ‘‘standard The general consensus of the scientific literature is that,
million’’ in 10-year age groups. The U.S. standard million if it is appropriate to standardize, then the selection of the
population is presented in appendix table I. A specific standard population should not affect relative comparisons.
example of age-adjustment for one of the Healthy People However, standardization is not appropriate when age-
2000 mortality objectives using the standard million and specific death rates in the populations being compared do
10-year ASDR is given in appendix table II. The calculation not have a consistent relationship (12).
of the variances of the age-adjusted rates in table II are For example, evaluating trends in age-adjusted cancer
shown in appendix table III. death rates over time can be difficult because the ASDRs for
NCHS publishes a large number of ADRs based on the younger ages have been decreasing while death rates at older
U.S. standard population. This standard is also used to track ages are increasing. If a relatively young standard population
those mortality objectives in Healthy People 2000 and those is used, the trend in ADR may show a small increase or
Health Status Indicators which are monitored with even a decrease; if a relatively older standard population is
age-adjusted rates. Several States use the same standard used, cancer mortality shows a much larger increase. Thus,
population in their publications. As long as the identical using a more current (i.e., older) population than the 1940
standard is used, ADRs from various national and State standard, such as the 1990 U.S. population, as a standard
publications can be compared. But, if different standard population yields a much greater increase in the cancer
populations are used to compute the ADR, then these ADRs mortality trend curve than does an analysis using the 1940
are not comparable. Thus, there are considerable advantages standard. Under these circumstances, a single summary
to using the U.S. standard when computing State and local measure is likely to be inappropriate for describing trends
ADRs. over time. Here, one should not use ADR, but should look at
In recent years there have been discussions about trends among ASDRs.
whether the 1940 standard should be supplanted by a more This does not mean it is inappropriate to publish ADR
contemporary ‘‘standard’’ that more closely reflects the U.S. for cancer mortality. Within a defined time interval, e.g.,
population’s current (or future) age distribution (6,7). In 1990, geographic or race-sex comparisons may still be
considering this issue, it is important to remember that the appropriate. It can be noted that when age-adjusted rates are
actual magnitude of the ADR is beside the point. The ADR computed using two distinct standards and the comparisons
is an index number used for relative comparisons that should are different, then it is not appropriate to standardize in the
not be affected by the choice of a standard. (See section on first place. Again, only age-specific comparisons may be

3
valid. Kitagawa illustrates this situation with an example of Summary
the mortality of white males living in metropolitan counties This paper describes some of the issues related to the
compared with those residing in nonmetropolitan counties in computation and use of age-adjusted rates. The following
1960 (13). In this case, ASDRs for white males under age 40 points were made:
were lower in metropolitan counties than in nonmetropolitan
+ The age-adjusted rate is an index measure, the magnitude
counties. After age 40, the reverse was true. A summary
of which has no intrinsic value. It should be used for
index, such as the ADR, does not adequately describe the comparison purposes only.
mortality differentials in the two groups. In cases such as + If it is appropriate to use age-adjustment, then the
these, the ADR is an imprecise indicator of mortality; the comparison should not be affected by the selection of a
age-specific comparisons would be a better choice. standard population. Conversely, if the comparison can be
affected by the choice of a standard population, then it is
not appropriate to age-adjust for that comparison.
Indirect standardization
+ The standard population should not be ‘‘abnormal’’ or
Because of concerns with the use of ADR, some ‘‘unnatural’’ when compared to populations under study.
mortality analysts prefer indirect standardized rates. Indirect Considering the amount of published material, there are
standardization is generally thought of as an approximation advantages to using the U.S. standard population.
to direct standardization. That is, when data needed to + Standardization is not a substitute for the examination of
compute a direct measure (e.g., ASDRs) are not available, age-specific rates.
there may still be enough information to compute an While standardization is most often applied to a series
indirectly standardized measure. However, indirect of age-specific death rates, direct or indirect standardization
standardization has intrinsic value and should be considered can also be applied to variables other than age. For example,
on its own merits, not solely as an approximation to direct infant mortality rates can be adjusted for birthweight
standardization (14,15). distribution (16). Age-adjustment can also be used to
monitor other measures of health at the local level, such as
For indirect standardization, a standard set of
incidence or prevalence of disease.
age-specific death rates are assumed to apply to the observed
Throughout the history of the ADR, the utility of the
population. For example, the age-specific U.S. death rates
measure has often come into question. Any summary index,
could be applied to the age-specific local area population.
including direct or indirect standardization, will mask
This technique yields an ‘‘expected’’ number of deaths in a
age-specific differences. Therefore, some authors have
population, assuming the standard set of ASDRs was stressed the importance of comparing individual age-specific
operating in the population. rates rather than attempting to summarize differences among
An indirect adjusted death rate (IADR) can be computed the age-specific rates (17,18). A summary index, however, is
from the expected number of deaths, but the index most more easily compared than an entire table of age-specific
often used is the ratio of the expected to the actual observed rates. Thus, the age-adjusted rate continues to be an integral
number of deaths. This ratio is called the standardized part of the analysis of mortality trends and differentials.
mortality ratio (SMR). The mathematics of indirect Accepting this, the need for a summary index must be
standardization and an example of the calculation of an balanced with recognition of the limitations of summary
SMR are given in the appendix. measures.

4
References 11. Kleinman JC. Age-adjusted mortality indices for small areas:
Applications to health planning. American Journal of Public
1. Shyrock HS, Siegel JS. The methods and materials of demog-
Health, 67:834–40. 1977.
raphy, vol 2. U.S. Bureau of the Census. Washington: U.S.
12. Fleiss JL. Statistical methods for rates and proportions. John
Government Printing Office. 1971.
Wiley and Sons, New York. 1973.
2. Spiegelman M. Introduction to demography. Rev. ed. Cam-
13. Kitagawa EM. Theoretical considerations in the selection of a
bridge, MA. Harvard University Press. 1968.
3. Neison FGP. On a method recently proposed for conducting mortality index and some empirical comparison. Human Biology,
inquiries into the comparative sanatory condition of various 38:293–308. 1966.
districts. Journal of the Royal Statistical Society of London (now 14. Inskip H, Beral V, Fraser P. Methods for age-adjustment of rates.
the Royal Statistical Society), vol 7, pp 40–68. 1844. Statistics in Medicine 2:455–66. 1983.
4. Kalton G. Standardization: A technique to control for extraneous 15. Tukey JW. Statistical mapping: What should and should not be
variables. Applied Statistics, 17:118–36. 1968. plotted. Proceedings of the 1976 workshop on automated cartog-
5. National Center for Health Statistics. Vital Statistics of the raphy and epidemiology. National Center for Health Statistics.
United States, 1989, vol II, mortality, part A. Washington: Public DHEW (PHS) 79–1254. 1979.
Health Service. 1992. 16. Foster JE, Kleinman JC. Adjusting neonatal mortality rates for
6. Johnson R. Proposed new standard population. Proceedings of birthweight. National Center for Health Statistics. Vital Health
the social statistics section, American Statistical Association, pp Stat 2(94). 1982.
176–81. 1990. 17. Woosley TD. Adjusted death rates and other indices of mortality.
7. Feinleib MF, Zarate AO, eds. Reconsidering age adjustment Chapter 4 in Vital Statistics Rates in the United States,
procedures: Workshop proceedings. National Center for Health 1900–[Link]: U.S. Government Printing Office. 1959.
Statistics. Vital Health Stat 4(29). 1992. 18. Elveback LR. Discussion of indexes of mortality and tests of
8. Spiegelman M, Marks HH. Empirical testing of standards for the their statistical significance. Human Biology 38:322–24. 1966.
age adjustment of death rates by the direct method. Human 19. Chiang CL. Standard error of the age-adjusted death rate. U.S.
Biology, 38:280–92. 1966.
Department of Health, Education, and Welfare: Vital Statistics
9. Curtin LR, Maurer J, Rosenberg HM. On the selection of
Special Reports 47:271–85. 1961.
alternative standards for the age-adjusted death rate: Proceed-
20. Keyfitz N. Sampling variance of standardized mortality rates.
ings of the social statistics section, American Statistical Associa-
Human Biology 38:309–17. 1966.
tion, pp 218–23. 1980.
10. Wolfenden HH. On the theoretical and practical considerations
underlying the direct and indirect standardization of death rates.
Population Studies, 16:188–90. 1962.

5
Appendix

Table I. Healthy People 2000 mortality objectives

Objective ICD–9 Objective ICD–9
number Cause of death identifying codes number Cause of death identifying codes

1.1 Coronary heart disease 410–414, 402, 9.3 Motor vehicle crashes E810–E825
429.2 9.3a [Ages 14 and younger]
1.1a [Blacks] 9.3b [Ages 15–24]
9.3c [Ages 70 and older]
2.1 See 1.1 9.3d [American Indians/Alaska Natives]
2.1a See 1.1a 9.3e [Motorcyclists]
2.2 Cancer (all sites) 140–208 9.3f [Pedestrians]

3.1 See 1.1 9.4 Falls and fall-related injuries E880–E888

3.1a See 1.1a 9.4a [Ages 65–84]
3.2 Lung cancer 162.2–162.9 9.4b [Ages 85+]
3.3 Chronic obstructive pulmonary disease 490–496 9.4c [Black males 30–69]

4.1 Alcohol-related motor vehicle crashes E810–E819

9.5 Drowning E830, E832, E910
4.1a [American Indians/Alaska Natives]
9.5a [Ages 0–4]
4.1b [Ages 15–24]
9.5b [Males 15–34]
4.2 Cirrhosis 571
9.5c [Black males]
4.2a [Black males]
4.2b [American Indians/Alaska Natives]
9.6 Residential fires E890–E899
4.3 Drug-related deaths 292, 304,
9.6a [Ages 0–4]
305.2–305.9,
9.6b [Ages 65 and older]
E850–E858,
E950.0–E950.5, 9.6c [Black males]
E962.0, 9.6d [Black females]
E980.0–E980.5
10.1 Work-related injuries1 E800–E999
6.1 Suicides E950–E959 10.1a [Mine workers]
6.1a [Ages 15–19] 10.1b [Construction workers]
6.1b [Males 20–34] 10.1c [Transportation workers]
6.1c [White males 65 and older] 10.1d [Farm workers]
6.1d [American Indian/Alaska Native males]
13.7 Cancer of the oral cavity and pharynx 140–149
7.1 Homicides E960–E969
7.1a [Children 0–3] 14.3 Maternal mortality 630–676
7.1b [Spouses 15–34] 14.3a [Blacks]
7.1c [Black males 15–34]
15.1 See 1.1
7.1d [Hispanic males 15–34]
15.1a See 1.1a
7.1e [Black females 15–34]
15.2 Stroke 430–438
7.1f [American Indians/Alaska Natives]
15.2a [Blacks]

7.2 See 6.1

16.1 See 2.2
7.2a See 6.1a
16.2 See 3.2
7.2b See 6.1b
16.3 Breast cancer in women 174
7.2c See 6.1c
16.4 Cancer of the uterine cervix 180
7.2d See 6.1d
153.0–154.3,
16.5 Colorectal cancer 154.8,159.0
7.3 Firearm injuries E922.0–E922.3,
E922.8–E922.9, 17.9 Diabetes-related deaths1 250
E955.0–E955.4, 17.9a [Blacks]
E965.0–E965.4,
17.9b [American Indians/Alaska Natives]
E970,
E985.0–E985.4
Epidemic-related pneumonia and
Knife injuries E920.3, E956, 20.2 influenza deaths for ages 65+ 480–487
E966, E986, E974

1Healthy People 2000 uses multiple cause-of-death data.

9.1 Unintentional injuries E800–E949
9.1a [American Indians/Alaska Natives]
9.1b [Black males]
9.1c [White males]

6
 This appendix presents examples of the computation of are based on the 1940 U.S. population and are called the
the age-adjusted death rate (ADR), indirect adjusted death ‘‘standard million.’’ As the name implies, the standard
rate (IADR), and the standard mortality ratio (SMR). These million weights sum to one million. The standard million is
examples demonstrate that each standardized index is a shown in table II.
weighted average of the age-specific rates. For the ADR, the The age-adjusted rates shown in most NCHS
weights are determined by the standard population. A publications and those used to track the Healthy People 2000
discussion of the variability of the ADR is also included. objectives are computed using the standard million and
Suppose the data are aggregated into i = 1, 2, ..., I age ASDRs in 10-year age groups. A specific calculation for
groups. Let: stroke mortality (Healthy People 2000 objective 15.2) for
di = the number of deaths in the i-th age interval, and males and females is shown in table III. For illustrative
pi = the population size in the i-th age interval. purposes, the deaths and populations are those of a
The total number of deaths is hypothetical medium-sized State.
d = ∑i di In this example, di = deaths in 10-year age-groups,
m i = 10-year ASDR per 100,000, and wsi = psi / Σi psi
the total population is
(weights on a unit basis).
p = ∑i pi
Age specific death rates (ASDRs) are defined as Indirect standardization
number of deaths for age interval i For direct standardization, the observed ASDRs and a
ASDR =
midyear population for age interval i standard population are used. For indirect standardization,
thus, the observed population and a standard set of ASDRs are
ASDR = mi = the death rate in the i-th age interval. used. Indirect standardized rates are sometimes calculated
and presented, but more often an SMR is presented. The
The age-specific death rate is given by indirect standardized rate and SMR are defined as follows:
m i = di / p i
number of observed deaths
SMR =
In this form, the death rate (m) on a unit basis (i.e., per number of expected deaths
person) will be between 0 and 1. ASDRs are usually or
expressed as a rate per 1,000 or per 100,000 population. For
example, if there are 10 deaths in an age group that has a
SMR =
∑i di
total population of 1,000 persons, the ASDR on a unit basis ∑i msi * pi
is 0.01; per 1,000 it is 10, per 100,000 it is 1,000.
The annual crude death rate is defined as the total where msi are the standard ASDRs on a unit basis. The indirect
number of deaths over all ages divided by the midyear adjusted death rate is then
population. The crude death rate is then IADR = SMR * (crude rate for the standard population)
m = total deaths / total population or
Again, it is usually expressed per 1,000 or per 100,000
Ms * ∑i di
population. IADR =
Algebraically, the direct standardized (or age-adjusted) ∑i msi * pi
rate is a weighted average of the age-specific death rates. To
For the data in table A, the age-specific rates for community A
compute the ADR, the standard population is used to deter-
mine a set of weights. For convenience, let can be used as the standard rates. Then the crude rate and the
psi = population in age group i in the standard population indirect standardized rates are the same for community A (50
and let the standard weights be given by per 1,000); the SMR for community A is 1. The calculation for
the IADR and SMR for community B is shown in table IV.
psi
wsi = Then
∑i psi
400
[NOTE: That in this form 0<wsi<1 and the wsi sum to 1. The SMR = = 1.33 IADR = 1.33 * 50 = 67
300
weights are often expressed as a standard million so that wsi
sum to 1,000,000.] Each index has advantages and disadvantages. Indirect rates
can be used when age-specific numbers of deaths are not
Then the ADR is given by available or when the number of deaths is small. Also, the
ADR = ∑i wsi * mi
indirect standardized rates have smaller variability. However,
indirect rates may not be comparable across areas; they can be
The ASDRs used by NCHS to compute the ADR are used for comparisons of areas only if age and area effects are
rounded to one decimal place. The weights used by NCHS independent.

7
Variability (27.1, 28.7)

The numbers of deaths reported for a community Because the 95-percent confidence intervals do not overlap,
represent complete counts. As such, numbers of deaths and the difference between the ADRs for males and females is
death rates are not subject to sampling error, although they statistically significant at the 0.05 level.
are subject to errors in the registration process. However, Some care has to be exercised when both the rates are low
when used for analytic purposes, such as comparison of rates and the number of deaths is small. In this case, the above
over time or for different areas, the number of events that formula can result in the lower bound being less than zero, and
actually occurred may be considered as one of a large series death rates cannot be negative. One way to avoid this is to use
of possible results that could have arisen under the same log transformations of the rates; another way is to use a
circumstances. The probable range of values may be discrete distribution function. These methods are beyond the
estimated from the actual figures according to certain scope of this report. The NCHS Office of Research and
statistical assumptions. From these assumptions the standard Methodology can provide assistance on computing variances
errors of ASDRs and ADRs can be calculated (19,20). for ADRs based on small frequencies.
When ASDRs are based on sufficiently small numbers, a
The variance of an ASDR is assumed to be determined
simple Poisson approximation may be used to compute the
by a binomial distribution. This assumes that the chance of
variance of the ASDRs, as follows:
dying in an age interval is constant within the age interval
and that everyone has the same chance of dying; this is an
m2i
assumption of homogeneity. Under the homogeneity
di
assumption, the variance of an age-specific rate on a unit
basis is given by where mi is the ASDR on a unit basis for the i-th age group,
mi * (1 – mi ) and di is the corresponding number of deaths
Variance (mi ) = In these cases, the resultant Poisson age-specific variances can be
pi
used in the formulae described in this section to compute the
NOTE: If the rates are per 100,000, then the (1–mi) term variance of the ADR. More information on random variation
becomes (100,000–mi). can be found in the annual vital statistics volumes (5).
The variance of an ADR can be defined as weighted average
of the variances of the ASDRs. Under the assumption that
ASDR are independent, or that the covariance (mi , mj) = 0 for
i not equal to j, then the standardized rates are simply Table II. Standard million age distribution used
weighted averages of the age-specific rates, and the variance is to adjust death rates to the U.S. population
given by in 1940
Variance (ADR) = ∑i w2i * Variance (mi ) Standard
million Unit basis
Age (psi ) (psi / 1,000,000)
An example of the calculation of the variances of the age-
adjusted rates for males and females computed in table III is All ages . . . . . . . . . . . . . 1,000,000 1.00000
shown in table V. In order to obtain meaningful variances, the
number of deaths and the populations for males and females Under 1 year . . . . . . . . . . 15,343 0.015343
used in table V are those of the same hypothetical medium- 1–4 years . . . . . . . . . . . . 64,718 0.064718
5–14 years. . . . . . . . . . . . 170,355 0.170355
sized State used in table III. The variance of an age-adjusted
15–24 years . . . . . . . . . . . 181,677 0.181677
death rate for the entire U.S. population is extremely small.
25–34 years . . . . . . . . . . . 162,066 0.162066
Confidence intervals can be formed using the variances. If 35–44 years . . . . . . . . . . . 139,237 0.139237
the number of deaths is large enough (again, a rough principle 45–54 years . . . . . . . . . . . 117,811 0.117811
is 25 or more) then a 95-percent confidence interval for the 55–64 years . . . . . . . . . . . 80,294 0.080294
age-adjusted rate is formed as: 65–74 years . . . . . . . . . . . 48,426 0.048426
75–84 years . . . . . . . . . . . 17,303 0.017303
(ADR – 1.96 * √var (mi) , ADR + 1.96 * √var (mi) ) 85 years and over . . . . . . . 2,770 0.002770

For the example in table III, the 95-percent confidence interval

for the ADR for strokes for males for the hypothetical State is
[33.0 – (1.96 * 1.05) , 33.0 + (1.96 * 1.05)]
or
(30.9, 35.1)
For females, the 95-percent confidence interval is
[27.9 – (1.96 * 0.80) , 27.9 + (1.96 * 0.80)]
or

8
 Revised Table III
Table III. Age-adjusted death rate calculation for stroke (ICD-9 430–438) for males and females: Hypothetical
medium-sized State
pi mi wsi
Age di (thousands) (per 100,000) (unit basis) mi * wsi

Males

Under 1 year . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 38 2.6 0.015343 0.0398918

1–4 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . – 150 – 0.064718 0.0000000
5–14 years. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 322 0.3 0.170355 0.0511065
15–24 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 344 0.6 0.181677 0.1090062
25–34 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 443 1.8 0.162066 0.2917188
35–44 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 379 5.5 0.139237 0.7658035
45–54 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 256 18.0 0.117811 2.1205980
55–64 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 189 54.5 0.080294 4.3760230
65–74 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 136 186.8 0.048426 9.0459768
75–84 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371 57 650.9 0.017303 11.2625227
85 years and over . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 12 1,766.7 0.002770 4.8937590
Age-adjusted rate = Sum (mi * wsi ) = 33.0
(per 100,000)

Females

Under 1 year . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 36 2.8 0.015343 0.0429604

1–4 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . – 143 – 0.064718 0.0000000
5–14 years. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 309 0.3 0.170355 0.0511065
15–24 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 337 0.6 0.181677 0.1090062
25–34 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 458 1.5 0.162066 0.2430990
35–44 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 401 5.2 0.139237 0.7240324
45–54 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 267 15.4 0.117811 1.8142894
55–64 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 208 39.9 0.080294 3.2037306
65–74 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 178 137.6 0.048426 6.6634176
75–84 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553 100 553.0 0.017303 9.5685590
85 years and over . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 661 34 1,944.1 0.002770 5.3851570
Age-adjusted rate = Sum (mi * wsi ) = 27.8
(per 100,000)

Table IV. Calculation of SMR and indirect adjusted death rate

msi msi
Age (per 1,000) (unit basis) di pi msi * pi

0–34 years. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 0.02 180 6,000 120

35–64 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 0.04 150 3,000 120
65 years and over . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 0.06 70 1,000 60

Total. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 0.05 400 10,000 300

9
 Revised Table V
Table V. Variance calculation for the age-adjusted death rate calculation for stroke (ICD-9 430–438) for males
and females: Hypothetical medium-sized State
pi mi wsi
Age di (thousands) (per 100,000) (unit basis) Var (mi ) w2i* Var(mi )

Males

Under 1 year . . . . . . . . . . . . . . . . . . . . . . 1 38 2.6 0.015343 6.84193 0.00161

1–4 years . . . . . . . . . . . . . . . . . . . . . . . . – 150 – 0.064718 0.00000 0.00000
5–14 years. . . . . . . . . . . . . . . . . . . . . . . . 1 322 0.3 0.170355 0.09317 0.00270
15–24 years . . . . . . . . . . . . . . . . . . . . . . . 2 344 0.6 0.181677 0.17442 0.00576
25–34 years . . . . . . . . . . . . . . . . . . . . . . . 8 443 1.8 0.162066 0.40631 0.01067
35–44 years . . . . . . . . . . . . . . . . . . . . . . . 21 379 5.5 0.139237 1.45111 0.02813
45–54 years . . . . . . . . . . . . . . . . . . . . . . . 46 256 18.0 0.117811 7.02998 0.09757
55–64 years . . . . . . . . . . . . . . . . . . . . . . . 103 189 54.5 0.080294 28.82026 0.18581
65–74 years . . . . . . . . . . . . . . . . . . . . . . . 254 136 186.8 0.048426 137.09637 0.32150
75–84 years . . . . . . . . . . . . . . . . . . . . . . . 371 57 650.9 0.017303 1134.49700 0.33966
85 years and over . . . . . . . . . . . . . . . . . . . 212 12 1,766.7 0.002770 14462.39759 0.11097
Variance of age-adjusted death rate = Sum w21
* Var (mi) = 1.10
Standard error of ADR = Square root of variance = 1.05

Females

Under 1 year . . . . . . . . . . . . . . . . . . . . . . 1 36 2.8 0.015343 7.77756 0.00183

1–4 years . . . . . . . . . . . . . . . . . . . . . . . . – 143 – 0.064718 0.00000 0.00000
5–14 years. . . . . . . . . . . . . . . . . . . . . . . . 1 309 0.3 0.170355 0.09709 0.00282
15–24 years . . . . . . . . . . . . . . . . . . . . . . . 2 337 0.6 0.181677 0.17804 0.00588
25–34 years . . . . . . . . . . . . . . . . . . . . . . . 7 458 1.5 0.162066 0.32751 0.00860
35–44 years . . . . . . . . . . . . . . . . . . . . . . . 21 401 5.2 0.139237 1.29669 0.02514
45–54 years . . . . . . . . . . . . . . . . . . . . . . . 41 267 15.4 0.117811 5.76690 0.08004
55–64 years . . . . . . . . . . . . . . . . . . . . . . . 83 208 39.9 0.080294 19.17504 0.12362
65–74 years . . . . . . . . . . . . . . . . . . . . . . . 245 178 137.6 0.048426 77.19700 0.18103
75–84 years . . . . . . . . . . . . . . . . . . . . . . . 553 100 553.0 0.017303 549.94191 0.16465
85 years and over . . . . . . . . . . . . . . . . . . . 661 34 1,944.1 0.002770 5606.77868 0.04302
Variance of age-adjusted death rate = Sum w21 * Var (mi ) = 0.64
Standard error of ADR = Square root of variance = 0.80

10
DEPARTMENT OF
BULK RATE
HEALTH & HUMAN SERVICES POSTAGE & FEES PAID
Public Health Service PHS/NCHS
Centers for Disease Control and Prevention PERMIT NO. G-281
National Center for Health Statistics
6525 Belcrest Road
Hyattsville, Maryland 20782

OFFICIAL BUSINESS
PENALTY FOR PRIVATE USE, $300

DHHS Publication No. (PHS) 95–1237

4-1476 (3/95)