Temporal trends in diabetes mortality among American Indians and Hispanics in New Mexico: birth cohort and period effects.

Rates of diabetes mortality are disproportionately high among ethnic minorities in the United States. To describe ethnic trends and cohort effects in diabetes mortality in New Mexico, the authors examined the trends in mortality rates for non-Hispanic whites. Hispanics, and American Indians in the state during the period 1958-1994. Age-specific rates were examined graphically to qualitatively describe the contribution of calendar period and birth cohort effects to changes in the rates. The authors also fit age-period-cohort models to these data. Age-adjusted diabetes mortality rates for American Indians and Hispanics surpassed rates for non-Hispanic whites for all but the earliest two time periods. In the 1993-1994 period, the age-adjusted mortality rate for American Indians was 3.8 times higher for men and 5.6 times higher for women than for their non-Hispanic white counterparts. Rates for American Indian men and women increased sharply over the 37-year period, by 565% and 1,105%, respectively. Mortality rates increased among Hispanics over the period of study but less rapidly than did rates among American Indians. Graphical analyses of age-specific rates were consistent with birth cohort effects among both American Indians and Hispanics and also with a period effect among American Indians. Results from age-period-cohort models indicate a birth cohort effect starting with the 1912 cohort in American Indians and the 1902 cohort in Hispanics. A period effect was present during the 1960s in American Indians. American Indians have experienced an epidemic rise in diabetes mortality in New Mexico; if current trends continue, diabetes may become the leading cause of mortality among American Indians in the state.