Introduction We wanted to investigate any differences in breast cancer survival for Aboriginal women compared with non- Aboriginal women in New South Wales, Australia (NSW).

Methods All women diagnosed with primary invasive breast cancer in NSW in 1994–2002 (n=31 525) were probabilistically linked with all NSW hospital admissions. 296 women identified as Aboriginal at any hospital admission. Proportional hazards models were used to investigate the time to first surgery and to analyse cause-specific survival adjusting for differences in comorbidities, diagnosis age, disease spread, diagnosis year, rural residence and socioeconomic status.

Results Aboriginal women in NSW had similar times to their first surgery as non-Aboriginal women (HR 0.74, 95% CI 0.51 to 1.06) after adjusting for age, year of diagnosis, disease spread, socioeconomic status, rural residence and comorbidities. However, Aboriginal women were 53% more likely to die of their breast cancer (HR 1.53, 95% CI 1.19 to 1.97), after adjusting for whether they had surgery, diagnosis age, disease spread, diagnosis year, socioeconomic status, rural residence and comorbidities.

Conclusions Differences in breast cancer survival in NSW for Aboriginal and non-Aboriginal women were not totally explained by demographic, disease or access to surgical treatment differences. Residual differences in breast cancer survival maybe due to differences in uptake of adjuvant therapies including chemotherapy and radiotherapy, unmeasured comorbidities or cultural barriers to accessing optimal cancer care. Overall breast cancer treatment and outcomes were better than for other cancers and we will investigating the reasons for this.