Misclassification of American Indian race in state cancer data among non-federally recognized Indians in North Carolina

Abstract

Background: Equitable distribution of health care resources relies on accurate morbidity and mortality data, classified by race. Accurate classification is a problem, particularly for non-federally recognized American Indians (AI) receiving care outside of the Indian Health Service.

Methods: We identified counties in North Carolina that had the majority of AIs belonging to the 7 state-recognized, non-federally recognized tribes. We collaborated with the tribe in each county and compared the incident cases of cancer in the North Carolina Central Cancer Registry (NCCCR) to the tribal rolls. Data were analyzed to calculate what percent of names on both lists were not correctly identified as AI in the NCCCR. We corrected the NCCCR classification and calculated the percentage misclassified, then recalculated the cancer incidence rates for 4 major cancers (prostate, female breast, lung, and colorectal). We compared the recalculated rate to the original rate.

Results: There were 626 AIs on the tribal rolls; 112 (17.9%) were not identified as AI on the NCCCR list. Comparing 1996-2000 age-adjusted cancer incidence rates before and after reclassification, the increase in rates were prostate 41%, female breast 18%, lung 10%, and 11% for colorectal cancers. There was less than a 2% increase in cancer rates for the combined 4 sites for Blacks and Whites, before and after reclassification, and 19% for AIs.

Conclusions: The study estimated 18% misclassification of non-federally recognized AIs in cancer registration in North Carolina, and determined an underestimation of cancer rates in the population. The underestimation of cancer burden among AIs in North Carolina may affect resources allocated for prevention, screening, and treatment programs, as well as funding for research.