What standard population to use in my epidemiological study?

Question

I need to standardized my mortality rates. for the period 2000-2022 in one European country. However, I am not use what standard population to use? Is Sagi outdated? In addition, I find this website https://seer.cancer.gov/stdpopulations/world.who.html with WHO 2002-2025 standard population? Should I use ''Rounded to Integers'' column for standardization?

Answer 1

MUST BE RATES

The OP has not clarified whether s/he is referring to “rates” or something else. The OP mentions “rations.” This seems to be a mistake. The tag is for “epidemiology.” The assumption here is that the OP seeks to standardize one or more “rates.”

THE GOAL IS COMPARISON

Rate standardization is used most often to compare one standardized rate with another standardized rate (often a published standardized rate); to do an “internal” comparison between two or more groups or populations (e.g. between countries; between states or provinces within a country; between regions of a country); or to assess time trends when the age structure of the population may have changed to “remove” the “effect” of changes in the age distribution over time.

MUST CLARIFY THE COMPARISON(S) TO BE MADE

The choice of standard for your data should be based on what comparisons of the standardized rate(s) you plan to make.

If you want to compare your standardized rate(s) with a rate (or rates) standardized to the WHO 2000-2025 world standard population, then you should use the WHO 2000-2025 world standard population to standardize your rate(s).

If you want to compare your standardized rate(s) with a rate (or rates) that have been standardized to the Segi “world” standard population, then you should use the Segi “world” standard population to standardize your rate(s).

If you want to do an “internal comparison”—say a comparison of a rate from one part of a country with a rate from another part of the same country when there is concern that there may be differences in the age distribution between the two parts of the country—then it does not matter what standard you use. The WHO 2000-2025 world standard population is fine but so is the Segi “world” population.

There are standard populations that are specific for Europe. If you are in Europe and are doing an “internal” comparison, you might want to use a European standard population. Here is a link to a publication that may help.

https://ourworldindata.org/age-standardization

Edouard Mathieu (2023) - “How does age standardization make health metrics comparable?” Published online at OurWorldInData.org.

AVAILABILITY OF DATA ABOUT STANDARD POPULATIONS

Here is the link to the WHO 2000-2025 world standard population data, cited by the OP.

https://seer.cancer.gov/stdpopulations/world.who.html

The National Cancer Institute has made available data files that can be used for standardizing rates when the goal is to compare standardized rates for the US, for Canada, for Europe, and for the “world.” For “world,” data files are available for both the Segi/Doll “world” standard population and the WHO 2000-2025 world standard population.

https://seer.cancer.gov/stdpopulations/

U.S. Standards (1940, 1950, 1960, 1970, 1980, 1990, 2000)

Canadian Standards (1991, 1996, 2011)

European (Scandinavian 1960) Standard2

European (EU-27 plus EFTA 2011-2030) Standard

World (Segi 1960) Standard

World (WHO 2000-2025) Standard

ROUNDED-TO-INTEGER

The linked table provided by the OP shows data on the WHO 2000-2025 world standard population. In addition to the percentage distribution by age, the table has a column that presents, for each age group, the estimated number of people in a hypothetical population of 1 million people calculated to 14 (!!!) digits. Of course, there cannot be fractional people (except in a hypothetical population) and no one (except maybe a physicist) needs 14 digits. Also shown in the linked table are rounded-to-integer values of the number of people in a hypothetical population of 1 million people for each age group. Fractional people become whole people and the number of whole people adds to (almost) 1 million (actually 999,999).

Yes, if you decide to use the WHO 2000-2025 world standard population to standardize your rate(s), use the rounded-to-integer values.

MECHANICS (AKA ARITHMETIC)

The mechanics of calculating an age-standardized rate using the data in the linked table using the WHO 2000-2025 world standard population are as follows.

For your data, calculate the age-specific rates as rates per million. Multiply the age-specific rates per million by the rounded-to-integer values in the table for each age group. This yields the estimated number of deaths/events in each age group for your data. Sum the estimated deaths/events across all age groups for your data. Divide by 1 million. This is the age standardized event rate (per million) for your data.

If you have data on the event rate standardized to the WHO 2000-2025 world standard population (per million) from somewhere else, you can now compare the age-standardized rate for your data to the data from somewhere else.

If you want to do an internal comparison (e.g., for one region of a country to another), simply repeat the calculations for the other groups to be compared.

Of course, there is software to do this. But arithmetic also works.

BACKGROUND ON WORLD STANDARD POPULATIONS

THE SEGI “WORLD” STANDARD POPULATION

Segi attempted to define a “world” population standard using 1950-1957 data from a select group of countries with high quality data on the age distribution of the population.

Segi M. Cancer mortality for selected sites in 24 countries (1950-57). Department of Public Health, Tohoku University of Medicine, Sendai, Japan. 1960.

The Segi “world” standard population underwent some modifications and was used by Doll et al. in 1966 in their seminal report: Cancer Incidence in Five Continents.

Doll R, Payne P. Waterhouse J. (eds) (1966) Cancer Incidence in Five Continents: A Technical Report, Berlin, Springer-Verlag.

The Segi/Doll “world” standard population became widely used for “world” (or “global”) comparisons of rate data after standardization. Monographs on cancer incidence using data from around the world have been published regularly since the 1966 report by Doll and colleagues with the Agency for International Research on Cancer (IARC) as the publisher. The most recent identified publication in the series is from 2021.

Cancer Incidence in Five Continents Volume XI IARC Scientific Publication No. 166. Edited by Bray F, Colombet M, Mery L, Piñeros M, Znaor A, Zanetti R, Ferlay J. 2021.

All of the IARC monographs on cancer incidence, including the 2021 monograph, use the Segi/Doll 1966 “world” standard population.

In the cancer field, the Segi/Doll “world” standard population is alive and well.

THE WHO 2000-2025 WORLD STANDARD POPULATION

In 2001, Ahmad and co-authors proposed a “new” WHO world standard population.

Ahmad OB, Boschi-Pinto CB, Lopez AD, Murray CJL, Lozano R, Inoue M. Age standardization of rates: a new WHO standard. GPE Discussion Paper Series: No. 31. EIP/GPE/EBD. World Health Organization. 2001.

https://www.researchgate.net/ publication/284696312_Age_Standardization_of_Rates_A_New_WHO_Standard

Ahmad and co-authors provide a brief history of attempts to develop a standard population for “the world.” They present the rationale for their “new” WHO 2000-2025 world standard population. They point out that between the 1950’s and 2000, the percentage distribution by age of the world population changed a lot. A higher percentage of the world population was older. Further, in 2001, basing a “world” population on data from only 46 countries was no longer necessary because high quality data on age-distribution from many more countries were available.

COMPARING THE SEGI/DOLL AND WHO 2000-2025 WORLD STANDARD POPULATIONS

In 2002, Bray and colleagues present the results of an assessment of conclusions about cancer incidence rates comparing countries and regions of the world based on rates that used the Segi/Doll “world” standard compared with rates that used the WHO 2000-2025 world standard population. The concluded that:

"there was nothing to gain by changing the standard population for routine comparisons of cancer data worldwide, other than the inconsequential property that the value of the standardized rate would be closer to the crude rate"

Bray F, Guilloux A, Sankila R, Parkin DM. (2002) Practical implications of imposing a new World standard population. Cancer Causes Control, 13, 175–182

Ahmad and co-authors present some comparisons of rankings of mortality and morbidity for various conditions using the Segi “world” standard population, the WHO 2000-2025 world standard population, and another standard population in use at that time—the “Scandinavian” standard population. They found little difference in the relative rankings of different countries when using the Segi “world” population standard, the WHO 2001-2025 world population standard, and the Scandinavian standard population, although the absolute values of the age-standardized rates could be very different.

NOTE ON USE OF DIRECT (VERSUS INDIRECT) STANDARDIZATION In considering standardization, an important consideration is the method of standardization. The above-referenced 2001 publication by Ahmad and colleagues specifically recommends using DIRECT age standardization. So do a lot of other commentators. The alternative is indirect standardization. Indirect standardization was used commonly in the age of (epidemiologic) dinosaurs. As Ahmad and colleagues (among others) point out:

"with the increasing availability of age-specific rates, the use of direct age standardization has become the predominant technique in most applications of demography and epidemiology"

That was back in 2001. One hardly ever sees an epidemiologic analysis that uses indirect age-standardization.