2000 US Standard Population vs. Standard Million
With the increasing availability of single year of age population data, single-age standardizing populations were needed in order to be able to age-adjust by single ages as well as by any arbitrary age groups (e.g. <18, 18+). Since single year of age standards have not been previously published, we obtained the original single-age population projections from the U.S. Census Bureau's Population Projections Program P25-1130 series1 that were used by the National Center for Health Statistics (NCHS) in developing the currently available 2000 US standard millions. These single-age projections are unpublished data provided by the Census Bureau that expand upon the age groups included in the published population projections (PDF). From the single-age populations, we created single-age standard millions following NCHS's methodology. As expected, when we summed the single ages in the 18 or 19 age groups, the standards did not match the NCHS published standards. Rather than trying to adjust the new single-age standards to match previously published numbers, we decided to use the raw population numbers that were used by NCHS for the standards. The new standards sum to 274,633,642 rather than forcing them to add to 1,000,000.
With the ability in SEER*Stat to redefine the age-adjustment categories by collapsing age groups, you can age-adjust to the 18 and 19 age group levels from a database with single age populations. We feel that it is very important for statistics produced in this manner match those same statistics produced from a database with populations by 18 or 19 age groups. Therefore, in addition to the existing standard millions, we now provide new standards for 18 and 19 age groups which were created by summing the appropriate single-age populations. Starting with the November 2004 SEER submission of incidence and SEER's presentation of US Mortality Data (diagnoses/deaths through 2002), new standard population is utilized.
To investigate the impact of the new proportions for age-adjusting, rates were created for 1975-2001 SEER 9 incidence data using both the existing 19-age-group 2000 US standard million and the new 19-age-group 2000 US standard population. These rates were calculated by cancer site (105 site groupings), SEER registry (10 groupings, including total), year of diagnosis (27 individual and five ranges), age at diagnosis (48 groupings), race (all, white, black), and sex (total, male, female). The age-adjusted rates were compared with two levels of precision (one decimal place -- our standard for reporting -- and four decimal places). With one decimal place of precision, 11,868 out of 8,769,600 (0.14%) differed. With four decimal places of precision, 1,964,337 of the 8,769,600 (22.40%) differed. When the comparisons were limited to rates based on 25 or more cases, the percentages increased as follows: for one decimal place, 6,294 out of 1,928,787 (0.33%) differed, and for four decimal places, 1,004,258 out of 1,928,787 (52.07%) differed. Of these 1,928,787 rates based on 25 or more cases, the largest absolute difference in two rates was 0.0778 and the largest percent difference was 0.27%. For specific comparisons, see the Incidence Rates Comparison Table.
1 Day, Jennifer Cheeseman, Population Projections of the United States by Age, Sex, Race, and Hispanic Origin: 1995 to 2050, U.S. Bureau of the Census, Current Population Reports, P25-1130, U.S. Government Printing Office, Washington, DC, 1996. (https://www.census.gov/prod/1/pop/p25-1130/p251130.pdf)
2 Anderson RN, Rosenberg HM. Age Standardization of Death Rates: Implementation of the Year 2000 Standard. National vital statistics reports: vol 47 no. 3. Hyattsville, Maryland: National Center for Health Statistics. 1998.