Report to the Nation, 19752004/Espey et al.
TABLE 1 U.S. Population Coverage for Data Sets Used in Cancer Incidence Analyses by Race/Ethnicity
AI/AN (CHSDA Counties)
(Table 4) 81.8 82.1 77.9 89.7
AI/AN indicates American Indian and Alaska Native; CHSDA, Indian Health Service Contract Health Service Delivery Area; SEER, Surveillance, Epidemiology, and End Results; NPCR, National Program of Cancer Registries; NAACCR, North American Association of Central Cancer Registries. SEER areas include Connecticut, Hawaii, Iowa, Utah, and New Mexico and the metropolitan areas of San Francisco, Detroit, Atlanta, and Seattle-Puget Sound (ie SEER 9). NPCR and SEER areas with cancer registry data reported by the NAACR as meeting high-quality standards for the specified period. The Wisconsin and Arizona data for the specified period are used for AI/AN as well as AI/AN and non-Hispanic white in CHSDA counties only. Percentage of the U.S. population. Data from 17 states that met high-quality standards for the specified period. a b c d e
All race/ethnic groups White Black API AI/AN AI/AN (CHSDA counties) Hispanic Non-Hispanic
(Table 2) 9.4
(Table 4) 59.3
58.8 56.0 80.6 48.5 20.7e 84.6 55.7
County level population estimates from the U.S. Bureau of the Census, modified by NCI, were used as denominators in the rate calculations.18 Because the 2000 census allowed respondents to identify them- selves as multiracial, CDC’s National Center for Health Statistics collaborated with the Census Bureau to develop bridging methods for single-race esti- mates to describe long-term trends in disease.19 The bridged, single-race annual population estimates were used for the years 1990 through 2004.
All analyses (long-term trends, fixed-interval trends, and average annual rates) were based on inci- dence data for various geographic areas. Incidence data were not available uniformly for all populations, time periods, and geographic areas. Long-term inci- dence trends (1975 through 2004) for all races/ethni- cities combined by sex for all sites combined and the 15 most common cancers were based on SEER inci- dence data covering approximately 10% of the U.S. population, the only source for long-term incidence analyses.20 Fixed-interval trends (1995 through 2004) for all race/ethnic populations combined and for white, black, Asian/Pacific Islander (API), Hispanic, and non-Hispanic populations for all sites combined and for the 15 most common cancers were based on data from 24 cancer registries covering approxi-
mately 59% of the U.S. population. Average-annual (2000 through 2004), sex-specific, and age-adjusted incidence rates for white, black, API, Hispanic, and non-Hispanic populations were based on incidence data from 39 cancer registries, covering approxi- mately 82% of the U.S. population. For the AI/AN population, data for 2 geographic catchment areas are presented for fixed-interval trends and rates. The first includes AI/AN data from all counties in states that were included in the analytic dataset. The sec- ond includes only data from Indian Health Service (IHS) Contract Health Service Delivery Area (CHSDA) counties in the states that were included in the ana- lytic dataset (see ‘‘Data for AI/AN’’ below, for an ex- planation of CHSDA) (Table 1). Because the AI/AN population is smaller proportionally than the other major race/ethnic populations, 6 years of data (1999 through 2004) were used for AI/AN incidence and death rates to increase the precision of the estimates. All registry data included in the analyses met NAACCR’s standards for high-quality cancer inci- dence data. The percent of population coverage for each of the cancer incidence datasets used for the tables in this report are summarized in Table 1, including the percentage of the AI/AN population covered by each.