Index: Stata tutorials

In this tutorial, we examine sex difference in survival for patients diagnosed with melanoma and illustrate how to estimate the causal effect of sex on patient survival using regression standardisation (also known and G-Computation). We discuss various approaches to estimating adjusted survival curves and illustrate the meansurv and stpm2_standsurv post-estimation commands to stpm2 (for fitting flexible parametric models in Stata).

Age-standardised net survival using non-parametric methods (Ederer II and Pohar Perme) and ICSS weights.

Estimate conditional (on surviving some time) relative survival using several approaches. We estimate from both life tables and based on a flexible parametric model.

This exercise illustrates an approach to modelling relative survival with missing covariate data using multiple imputation.

Estimation of relative/net survival using four difference approaches (Ederer I, Ederer II, Hakulinen, Pohar Perme)

An illustration of how one can both approximate and exactly replicate estimates from a Cox model using Poisson regression

In this tutorial, we use a model-based approach to estimate all-cause survival that is age-standardised to the International Cancer Survival Standard (ICSS). In a separate tutorial, we accomplished this by fitting a separate model for each age group and then taking the weighted average of the age-specific estimates to get the age-standardised estimates. In this tutorial we apply the weights at an individual level, which precludes the need to explicitly estimate survival within each age group.