Inverse probability weighting and doubly robust standardization in the relative survival framework

Area of Science:

• Epidemiology
• Biostatistics
• Survival Analysis

Background:

• Relative survival is crucial for cancer patient prognosis, accounting for general population mortality.
• Marginal relative survival summarizes cancer patient prognosis.
• Causal effects of exposures on survival can be estimated by comparing relative survival between groups.

Purpose of the Study:

• To extend inverse probability weighting (IPW) and doubly robust standardization methods to the relative survival framework.
• To provide robust tools for estimating average causal effects in cancer survival analysis, especially when relative survival models may be misspecified.

Main Methods:

• Applied regression standardization within a relative survival model framework.
• Developed and extended inverse probability weighting (IPW) and doubly robust standardization methods for relative survival analysis.
• Conducted a simulation study to assess method sensitivity to model misspecification and evaluated standard error estimation.

Main Results:

• Regression standardization provides consistent causal effect estimates if the relative survival model is correctly specified.
• IPW and doubly robust methods offer unbiased or robust estimates even with relative survival model misspecification.
• Simulation results demonstrate the performance and sensitivity of these methods.

Conclusions:

• IPW and doubly robust standardization are valuable additions to causal inference in relative survival analysis.
• These methods enhance the reliability of estimating average causal effects of exposures on cancer patient survival.
• The study provides practical tools and insights for handling model misspecification in survival research.