The association between socioeconomic status (SES) and relative survival of rectal cancer is little investigated. We hypothesized that the impact on risk of death by SES would be much smaller when differences in background mortality (comorbidity, lifestyle factors) were taken into account, i.e. in modelling relative survival of rectal cancer.
Individual data on civil status, education, and income were linked to the Swedish Rectal Cancer Registry 1995-2005 (n = 16,713). Specific life tables by socioeconomic group were used to calculate relative survival, and modelling included age, sex, stage, time period, and SES. The same covariates were applied in a Cox regression based on absolute survival.
Stage distribution was associated with civil status, education, and income (p < 0.001). In spite of modelling based on relative survival, an increased risk of death was found for all other patients compared with those who were married, as well as for all other patients compared with those with the highest income. The pattern was fundamentally the same as in a Cox regression model, only the point estimates were slightly reduced using the relative approach. In stage-specific modelling of relative survival, income was of particular importance in stage III; the hazard ratio (HR) for lowest versus the highest income was 1.37 [95 % confidence interval (CI) 1.15-1.64]. There were also significant differences by income among patients who had a major surgical resection (stage IV excluded).
Large and clinically relevant socioeconomic inequalities remained in stage-adjusted analyses of relative survival, also in a setting of universal healthcare and no screening program operating.
2014. Vol. 38, no 12, 3265-3275 p.