Evaluating Bias From Birth-cohort Effects in the Age-based Cox Proportional Hazards Model