Aims: The independent contribution that area and individual-level factors have on observed geographical variations in stage at diagnosis and survival for colorectal cancer in Australia have not been analysed. We explored the association between area disadvantage and geographic remoteness with stage at diagnosis and survival after invasive colorectal cancer while controlling for individual-level characteristics.
Methods: Multilevel models were used to analyze variations in stage at diagnosis for all 18,561 patients aged 20-79 years diagnosed with colorectal cancer in Queensland from 1996 to 2007, and notified to the population-based Queensland Cancer Registry. Survival was examined using multilevel logistic regression and Markov chain Monte Carlo simulations for 17,065 surgically treated cases diagnosed from 1996-2006 and followed up through 2008.
Results: Independently of individual-level variables, patients living in inner (OR 1.09, 95% CI 1.01-1.19) and outer regional areas (OR 1.11, 95% CI 1.01-1.22) were more likely to be diagnosed with advanced disease (p =0.045) although no variations with area disadvantage (p =0.885) were evident. After multivariate adjustment those from more disadvantaged areas had significantly (p = 0.002) worse survival (Mortality odds Ratio (MOR) =1.14, 1.16, 1.21, 1.18 for Quintiles 4, 3, 2 and 1 respectively) than least disadvantaged area. Patients from outer regional (MOR=1.24, 95% CI 1.11-1.38) or remote areas (MOR 1.32, 95% CI 1.12-1.56) also had poorer survival (p < 0.001). Important individual-level predictors of both advanced disease and poorer survival were being Indigenous, female, unmarried or in blue-collar occupations. Patients having less differentiated tumors, advanced disease, comorbidities (except diabetes) or treatment in public hospitals had poorer survival.
Conclusions: Both the risk of being diagnosed with advanced disease and survival after a diagnosis of invasive colorectal cancer depends on where a patient lives, independently of their individual characteristics. The underlying reasons for these inequalities need to be identified.