Choice of Survival Metric and Its Impacts on Cancer Survival Estimates for American Indian and Alaska Native People.

BACKGROUND: Different survival metrics have different applicability to clinical practice and research. We evaluated how choice of survival metric influences assessment of cancer survival among American Indian and Alaska Native (AIAN) people relative to non-Hispanic Whites (NHW). A secondary objective was to present variations in survival among AIAN people by age, sex, stage, and Indian Health Service (IHS) region. METHODS: Five-year survival was calculated using the North American Association of Central Cancer Registries Cancer in North America dataset. We calculated survival among AIAN people, compared with NHW using four approaches: (i) observed (crude) survival, (ii) cause-specific survival, (iii) relative survival using age- and sex-adjusted lifetables, and (iv) relative survival using lifetables additionally adjusted for race, geography, and socioeconomic status. For AIAN people, we evaluated how survival varied by age, stage at diagnosis, and IHS region. RESULTS: Observed survival methods produced the lowest estimates, and-excepting prostate cancer-cause-specific methods produced the highest survival estimates. Survival was lower among AIAN people than NHW for all methods. Among AIAN people, survival was higher among those 20-64 years, females, and tumors diagnosed at local stage. Survival varied by IHS region and cancer sites. CONCLUSIONS: These results support the assertion that using the same methodology to compare survival estimates between racial and ethnic groups is of paramount importance, but that the choice of metric requires careful consideration of study objectives. IMPACT: These findings have the potential to impact choice of survival metric to explore disparities among AIAN people.