Posts

A stratified log-rank test does not control alpha under a marginal null hypothesis, and that's ok

Recently, I’ve become interested in the robustness of standard statistical methods for RCTs under model misspecification. It seems to …

Multiple imputation without a specialist R package

If you’re interested in doing multiple imputation in R, it’s best to use a specialist package. There are many good options out there, …

Landmark/Milestone analysis under a Royston-Parmar flexible parametric survival model using the R package flexsurv

The aim of this post is to demonstrate a landmark/milestone analysis of RCT time-to-event data with a Royston-Parmar flexible …

That'll do: simulated annealing instead of maths

This is a post for future me. Given my steadily declining maths skills, I can see myself needing to use the generalized simulated …

Adjusting for covariates under non-proportional hazards

I’ve written a lot recently about non-proportional hazards in immuno-oncology. One aspect that I have unfortunately overlooked is …

A Bayesian approach to non-proportional hazards

In this blogpost I wanted to explore a Bayesian approach to non-proportional hazards. Take this data set as an example (the data is …

Shiny app for enhancing published Kaplan-Meier plots

Extract patient-level data from a Kaplan-Meier plot

I made a video with step-by-step instructions for extracting (approximate) patient-level data from a Kaplan-Meier curve. It’s a useful …

Is there still a place for significance testing in clinical trials?

I’ve spent a lot of time thinking about hypothesis tests in clinical trials recently. Periodically, it’s good to question the …