The impact of additional life-table variables on excess mortality estimates

Area of Science:

• Epidemiology
• Biostatistics
• Population Health

Background:

• Regression-based relative survival models are standard in population-based cancer studies to assess excess mortality.
• These models typically use life tables to adjust for general population mortality.
• Discrepancies arise when life table demographics do not match study cohort demographics, potentially causing bias.

Purpose of the Study:

• To evaluate the impact of additional life table variables on excess mortality estimates.
• To quantify the extent and direction of bias introduced by missing demographic data in life tables.
• To assess how bias affects the estimation of covariate effects in excess mortality models.

Main Methods:

• A simulation approach was employed with various plausible scenarios.
• The study simulated the impact of life table variable omissions on excess mortality.
• A population-based colorectal cancer analysis was used to demonstrate the bias.

Main Results:

• The absence of stratification variables in life tables introduces measurement bias in excess mortality estimates.
• Bias affects not only the covariate of interest but also other covariates in the model.
• The study quantified the extent and direction of these biases.

Conclusions:

• Incomplete life tables can lead to significant measurement bias in relative survival analyses.
• Researchers must consider the demographic completeness of life tables to ensure accurate excess mortality estimation.
• This bias was confirmed in a real-world colorectal cancer dataset.