Comparison of differentpopulations: standardization, direct

and indirect standardisation, SMR.

AgeGender

Education

Place of residence

Economic activityFamily status

Population’s mortality

Factors influencing a population’s mortality

Standardization

• „…is a set of techniques used to remove as far as possible the effectsof differences in age or other confounding variables when comparingtwo or more variables.”

Methods of standardization

• There are two methods of standardisation commonly used in epidemiological studies, and these are characterized by whether the standard used is a population distribution (direct method) or a set of specific rates (indirect method). Both direct and indirect standardisation involves the calculation of numbers of expected events (e.g. deaths), which are compared to the number of observed events.

Why standardize?Why standardize?Why standardize?Why standardize?

�Examining crude rates alone can be misleading if underlying populations are different (age-specific rates are better)

But

�Cumbersome to compare age-specific rates especially when doing large number of comparisons

Slide from: Nam Bains

Concept of directstandardization

Example 1

Step 1 – Calculate age-specific mortality

• Calculate the age-specific mortality rates for each age group in each population

How many age categories?How many age categories?How many age categories?How many age categories?

Lots (detailed age groups)

• better control of the effect of any differences in age distributions but,

• lots of strata means there might not be enough events (larger variance)

Fewer (broad groups)

• will produce less precise adjustment

• broad groups (i.e., 65+) will not be sensitive to changes in age-specific rates within that group

Other considerations

• availability of data (i.e., CCHS)

Slide from: Nam Bains

Step 1 - Calculation

Step 1 - Results

-

2 ,0 0 0

4 ,0 0 0

6 ,0 0 0

8 ,0 0 0

10 ,0 0 0

12 ,0 0 0

-

2 ,0 0 0

4 ,0 0 0

6 ,0 0 0

8 ,0 0 0

10 ,0 0 0

12 ,0 0 0

USA 1940 Canada 1991 World “Segi”

USA 2000 European WHO World

Step 2 – Choose a standard populationStandard populations are "artificial populations" with fictitious age structures, that are used in age standardization

as uniform basis for the calculation of comparable measures for the respective reference population(s).

Choice of standard population: considerationsChoice of standard population: considerationsChoice of standard population: considerationsChoice of standard population: considerations

�When several different populations are being compared, a ‘pooled’ standard minimizes the variance of the adjusted rates

�In examining trends, an appropriate standard is one that reflects the average structure of the population over the time period

�The standard should be similar to the population of interest

�It should not change frequently (all historic data would need to be recomputed)

�It should be used consistently to ensure comparability of rates

Choi, 1999. Am J Epi

Step 2 – Choose a standard population

Step 3 – Project Swedish and Mexicanmortality rates onto the standard population

Step 3 - Calculate

Mexico Sweden

St. Pop Mortality

Expected

deaths St. Pop.

Mortalit

y

Expected

deaths

0-29 51000 1,7‰ 86,7 51000 0,4‰ 20,4

30-59 37000 4,1‰ 151,7 37000 2,4‰ 88,8

60+ 12000 39,6‰ 475,2 12000 43,2‰ 518,4

713,6 627,6

Step 4: Calculate standardized mortality ratesfor Mexico and Sweden

• Mexico= 713,6/100 000*1000 = 7,136

• Sweden= 627,6/100 000*1000 = 6,276

Step 4 – Calculate Comparative MortalityRatio (CMR)

• CMR: Standardized mortality rate of Mexico/Standardized mortalityrate of Sweden

• CMR=1,13

• Mortality in Mexico is 13% higher than in Sweden.

Concept of indirectstandardization

When do we use indirect standardization?

• The indirect method of standardisation is commonly used when age-specific rates are unavailable. For example if we did not know the age specific mortality rates for country B (see next slide).

• In this method, instead of taking one population structure as standard and applying sets of rates to it to estimate expected events, a set of rates from a standard population (country A, see next slide) is applied to each of the populations being compared to calculate standardized morbidity/mortality ratios.

• Indirect standardization may be also used if we want to compare a small population (e.g. province, city, workers in a given factory) to alarger population (e.g. country). See Example 3.

When do we use indirect standardization?

• The indirect method of standardisation is commonly used when age-specific rates are unavailable. For example if we did not know the age specific mortality rates for country B (see next slide).

• In this method, instead of taking one population structure as standard and applying sets of rates to it to estimate expected events, a set of rates from a standard population (country A, see next slide) is applied to each of the populations being compared to calculate standardized morbidity/mortality ratios.

• Indirect standardization may be also used if we want to compare a small population (e.g. province, city, workers in a given factory) to alarger population (e.g. country). See Example 3.

Example 2

• In Country B 15 300 people died. We would like to compare thenumber of deaths with Country a. Unfortunately we don’t know thenumber of deaths in Country B. What do we do?

Step 1 - Calculate how many deaths would be expected in Country B if it had the same age-specific mortality rates as Country A

Step 2 – Calculate standardized mortality ratio (SMR)

• SMR = Observed number of deaths (O) X 100Expected number of deaths (E)

• Mortality in country B is 60% higher than the number we would expect if Country B had the same mortality experience as Country A.

Example 3

• Calculate how the number of deaths in a fictitious chemical factory inMexico relates to the number of deaths in the whole Mexicanpopulation.

Practice

20 years after the American National Health And Nutrition Examination Survey –

NHANES, 1971-75 Gu et al was trying to find out if there was a difference in mortality

between 1971 and 1993 of those claimed themselves diabetic in 1971 compared to the healthy

population. The following table shows some of their results:

Male Diabetic Non-diabetic

Population Number of death Population Number of death

25-44 years 454 10 34461 154

45-64 years 1222 60 28412 706

65-74 years 1484 157 18189 1371

The standard population of 1990:

Age-group Population

25-44 years 325,000

45-64 years 186,000

65-74 years 73,000

1. Calculate the standardized mortality of the diabetic population (per thousand)!

2. Calculate the relative mortality risk of the diabetic compared to the non-diabetic

population!

A study examined the prevalence of diabetes in two villages (A and B). The result is shown

by the table:

A village B village

Age group Population No. of diabetic Population No. of diabetic

15-39 4200 42 500 20

40-59 3000 450 600 240

60+ 1200 300 900 540

Total 8400 792 2000 800

Calculate the prevalence of diabetes in both villages!

Prevalence A:

Prevalence B:

Standardize the data using the following standard population and calculate prevalence again.

Age gorup Population

15-39 6500

40-59 5500

60+ 3000

Standardized prevalence A: Standardized prevalence B:

A study examined if visiting disco regularly can be an exposition factor for drug-usage. The

result is shown by the table:

Non-visiting population Regular disco visitors

Age group Population Have ever tried a

drug

Population Have ever tried a

drug

15-20 25000 525 7750 1248

21-25 35000 1190 12250 2217

26-30 10000 300 2000 216

31-35 10000 200 2000 200

Total 80000 2215 24000 3881

Calculate the prevalence of drug usage in both population! Prevalence among those not attending disco:

Prevalence among those visiting a disco regularly:

Standardize the data using the following standard population and calculate prevalence again.

Age group Population

15-20 71000

21-25 76000

26-30 86000

31-35 88000

Standardized prevalence among those not attending disco:

Standardized prevalence among those visiting a disco regularly:

The following table presents the mortality rate of two villages (A and B).

A village B village

Age group Population No. of death Age group No. of death

18-35 20000 40 12000 36

36-65 40000 300 30000 300

66+ 24000 1200 20000 800

Total 84000 1540 62000 1136

Calculate the crude mortality in both villages!

Mortality A: Mortality B:

Standardize the data using the following standard population and calculate mortality again.

Age group Population

18-35 65000

36-65 55000

66+ 30000

Standardized Mortality A:

Standardized Mortality B:

(2)(1)

22,95355,56555-64

8,21268,68745-54

2,86860,83835-44

1,59437,03025-34

1,3837,98920-24

Standard Death Rates per

1,000,000 (All Causes of Death)

Number of Farmers and

Farm Managers

(Census, 1951)Age

Group

Example: SMR for Male Farmers, England and

Wales, 1951

Total observed deaths per year among farmers: 1,464

SMR = ?