Adjusting relative survival estimates for cancer mortality in the general population.

BACKGROUND In theory, expected survival probabilities used in the derivation of relative survival ratios (RSR) are determined from a control group free of the cancer under study. In practice, expected survival is typically estimated from general population life tables--which include people previously diagnosed with cancer--potentially leading to an overestimation of relative survival. DATA AND METHODS Data are from the Canadian Cancer Registry with mortality follow-up through record linkage to the Canadian Vital Statistics Death Database. Period method RSRs for 2006-to-2008 were derived using general population life tables adjusted for cancer mortality and then compared with estimates derived using corresponding unadjusted life tables. RESULTS For all cancers combined, the use of general population life tables to derive expected survival probabilities overestimated RSRs by 0.6 (1-year), 2.4 (5-year) and 4.6 (10-year) percentage units. Biases in 5-year survival were highest among males (3.0) and among people aged 75 to 99 at diagnosis (4.1). The bias was negligible for most individual cancers; biases were highest for prostate cancer, followed by colorectal and female breast cancer. INTERPRETATION Canadian estimates of relative survival for all cancers combined calculated using general life tables warrant adjustment for cancer mortality. Consideration of adjustment for cancer mortality is recommended for estimates of colorectal, female breast and especially prostate cancer.