Background: In New Zealand the burden of breast cancer is not shared equally among ethnic groups. Although incidence is similar for indigenous Maori women and Pacific women compared with non-Maori/non-Pacific women, Maori and Pacific women are more likely to die from breast cancer. This suggests much of the breast cancer mortality disparities are due to inequalities in survival rather than incidence. We investigated the role that demographic and tumor factors play in explaining ethnic disparities in breast cancer survival.

Methods: Breast cancer cases notified to the New Zealand Cancer Registry (NZCR) from April 2005 to April 2007 were followed up to April 2009. Maori, Pacific and non-Maori/non-Pacific women were categorized according to ethnicity on the NZCR. Deprivation was analyzed as quintiles of the New Zealand area-based index of socioeconomic position. Relative survival rates were estimated using ethnic-specific life tables. Missing values were imputed and excess mortality modeling was used to estimate the contribution of demographic and tumor factors to ethnic inequalities in survival.

Results: There were 2,968 breast cancer cases (76.5% non-Maori/non-Pacific, 17% Maori, and 6.5% Pacific) included and 433 recorded deaths. Relative survival rates at 4 years were 91.45% (CI 89.68 to 92.93) for non-Maori/non-Pacific, 86.15 (CI 80.33 to 90.36) for Maori, and 79.56% (CI 68.17 to 87.24) for Pacific women. Using non-Maori/non-Pacific as the reference group, the age-adjusted hazard ratio (HR) dropped for Maori from 1.76 (CI 1.22 to 2.48) to 1.43 (CI 0.97 to 2.10) when further adjusted by deprivation. For Pacific the HR dropped from 2.49 (CI 1.57 to 3.94) to 1.94 (CI 1.20 to 3.13). Inequalities persisted after adjustment for subtype variables (ER/PR/HER2), but adjusting for access to care variables (stage/size) eliminated the ethnic inequalities in excess mortality.

Conclusion: Ethnic disparities in breast cancer survival in New Zealand can be attributed to differential access to health care rather than differences in breast cancer subtypes.

Citation Information: Cancer Epidemiol Biomarkers Prev 2010;19(10 Suppl):A93.