Survival from breast cancer: an analysis of Australian data by surgeon case load, treatment centre location, and health insurance status.

Objective. Early invasive breast cancer data from the Australian National Breast Cancer Audit were used to compare case fatality by surgeon case load, treatment centre location and health insurance status. Method. Deaths were traced to 31 December 2007, for cancers diagnosed in 1998-2005. Risk of breast cancer death was compared using Cox proportional hazards regression. Results. When adjustment was made for age and clinical risk factors: (i) the relative risk of breast cancer death (95% confidence limit) was lower when surgeons' annual case loads exceeded 20 cases, at 0.87 (0.76, 0.995) for 21-100 cases and 0.83 (0.72, 0.97) for higher case loads. These relative risks were not statistically significant when also adjusting for treatment centre location (P=0.15); and (ii) compared with major city centres, inner regional centres had a relative risk of 1.32 (1.18, 1.48), but the risk was not elevated for more remote sites at 0.95 (0.74, 1.22). Risk of death was not related to private insurance status. Conclusion. Higher breast cancer mortality in patients treated in inner regional than major city centres and in those treated by surgeons with lower case loads requires further study. What is known about the topic? Studies in some countries show an association of poorer outcomes with lower case load and lack of private health insurance. What does this paper add? Lower survivals apply in contemporary Australian environments where annual case loads are 20 or fewer and for patients treated in inner regional compared with major city centres. Poorer survivals for patients without private health insurance status are not statistically significant after adjusting for tumour size and other risk factors. What are the implications for practitioners? Additional research is needed to determine why survivals are lower in Australian settings where case loads are low and when treatment is provided in inner regional centres. Meanwhile, it would be appropriate to target these settings in quality improvement programs. [ABSTRACT FROM AUTHOR]