Fig. 4: Bias in 1- (upper panel) and 5-year (lower panel) net survival estimates for different simulated scenarios of registration errors, with 5%, 10% or 20% of cases missed at diagnosis, and 70% or 90% of those missed found through trace-back (indicated by blue and yellow markers, respectively). | British Journal of Cancer

Fig. 4: Bias in 1- (upper panel) and 5-year (lower panel) net survival estimates for different simulated scenarios of registration errors, with 5%, 10% or 20% of cases missed at diagnosis, and 70% or 90% of those missed found through trace-back (indicated by blue and yellow markers, respectively).

From: Exploring the impact of cancer registry completeness on international cancer survival differences: a simulation study

Fig. 4

The relative risk (RR) of being missed for those with the prognostic Factor X compared to those without Factor X is 5 or 1.5, and scenarios with unaltered follow-up time (circles) and shortened follow-up time for 30% of those found through trace-back (triangles) are presented. Bias is measured as a percentage point difference, and all registration errors lead to an underestimation of survival.

Back to article page