Background and purpose: Canada is a high-income country with universal healthcare. In international comparisons, its overall level of access to radiotherapy appears sufficient. However, challenges exist due to Canada's large geographic area and small population density. The association between access and cancer outcomes nationally has not yet been described. Materials and methods: We quantified geographic accessibility for 2012 using the linear distance from each Canadian health region centroid to the nearest radiotherapy center. We used geospatial analytic techniques to detect clusters of age-standardized all-cancer mortality-to-incidence ratios (MIRs) across health regions, from 2010–2012. Global ordinary least squares (OLS) and geographically-weighted regression (GWR) were conducted to examine relationships between distance and MIR, adjusting for sociodemographic factors. Results: Median distance from health region centroid to nearest radiotherapy center was 101.73 km (range 1.14–2095.12). One cluster of worse outcomes (MIR range 0.45–0.88) involved most of northern Canada, with a second cluster of better outcomes (MIR range 0.40–0.41) in southern British Columbia. In both regression models, regions with longer distance to radiotherapy center (ß = 0.0001), increased smoking (ß = 0.002), and poorer food security (ß = –0.003) were significantly associated with worse outcomes (OLS R2 = 0.70, GWR R2 = 0.74). Distance remained independently associated with MIR for lung and colorectal cancer subgroups, but not breast and prostate. Conclusions: A clear north–south discordance in cancer outcomes exists in Canada, with poorer outcomes in the north, while radiotherapy centers are concentrated along the south. Increased distance to radiotherapy, along with other sociodemographic and health-system factors, are associated with poorer cancer outcomes. Our study could be replicated, particularly in other high-income countries, to help identify national patterns and regional disparities in access and outcomes.
- Mortality-to-incidence ratios