Enumerating Ethnicity in the U.S. Census: It's Consequence for Hispanic/Non-Hispanic Cancer Rates

Changes between the 1990 and 2000 U.S. Census for enumerating ethnicity appreciably changed the numbers of Americans who identify themselves as ethnic Hispanics. The consequence of this change on estimating disparity in Hispanic/Non-hispanic (H/NH) health events is described using breast, prostate, colorectal and lung cancer incidence across Connecticut in 1994-2000, in relation to 1990 or 2000 population counts of H/NH state residents. SEER data, grouped into 3-year moving averages, were examined according to cancer site, ethnicity and gender. Risk ratios were calculated and t-tests were performed that contrasted H/NH findings. No significant differences in ratios of H/NH rates based on 1990 or 2000 census data were noted for women with breast or colorectal cancers, whereas significant differences in H/NH ratios were noted between analyses based on 1990 or 2000 census data for men with colorectal or prostate cancers. These results suggest that H/NH health disparities are intricate and that the way ethnicity is measured can affect our understanding and complicate longitudinal surveillance. In particular, attention to gender differences in ethnic identity is an important area for ongoing study.