Survival from breast, colon, lung, ovarian and rectal cancer by geographical remoteness in New South Wales, Australia, 2000-2008

Abstract

Objective: This study aims to compare survival from breast, colon, lung, ovarian and rectal cancer by geographical remoteness in New South Wales (NSW).

Design: Retrospective population-wide registry study.

Setting: NSW, Australia.

Participants: A total of 107 060 NSW residents, who were diagnosed with any of the five cancers between 01 January 2000 and 31 December 2008.

Main outcome measures: Kaplan-Meier survival curves and proportional hazards regression were used to compare survival by geographical remoteness of residence at diagnosis, controlling for gender, age and extent of disease at diagnosis. Remoteness was classified using standard definitions: major city, inner regional (InnReg), outer regional (OutReg) and remote (including very remote).

Results: Significant differences in survival (likelihood of death) were identified in all five cancers: breast (adjusted hazard ratio(HR) = 1.22 (95% confidence interval (CI), 1.001-1.48) in regionalised and HR = 1.30 (1.02-1.64) in metastatic disease for OutReg areas); colon (HR = 1.14 (1.01-1.29) for OutReg areas in metastatic disease); lung (HR range = 1.08-1.35 (1.01-1.48) for most non-metropolitan areas in all stages of disease excepting regionalised); ovarian (HR = 1.32 (1.06-1.65) for OutReg areas in metastatic disease, HR = 1.40 (1.04-1.90) for InnReg areas and HR = 1.68 (1.02-2.77) for OutReg areas in unknown stage of disease) and rectal (HR = 1.37 (1.05-1.78) for OutReg areas in localised and HR = 1.14 (1.002-1.30) for InnReg areas in regionalised disease). Where significant differences were found, major cities tended to show the best survival, whereas OutReg areas tended to show the worst. Although no definitive interpretation could be made regarding remote areas due to small patient numbers, their survival appeared relatively favourable.

Conclusions: Reasons that contribute to the differences observed and the disparate results between cancer types need to be further explored in order to facilitate targeted solutions in reducing survival inequality between NSW regions.

Keywords: Kaplan-Meier estimate; outcome assessment; proportional hazards regression; regional and rural; registries.