CODECHECK certificate 2024-024

CODECHECK certificate for the article "Evaluating individualized treatment effect predictions: A model-based perspective on discrimination and calibration assessment".

Summary: The three R scripts comprising this project were fairly straightforward to execute. I had some issues getting the simulation script (sim.R) running. It kept throwing an error but the specifics were masked by the usage of doParallel – you only get the error about “cannot open connection” but the problem was actually that the results folder did not exist. Once this was created, the script executed fine. The results folder is not necessary if users are only interested in producing the plots and tables in replicate.R, because by default that script loads in pre-computed simulated data. The full simulation had to be run overnight (~6 hours) on a standard laptop.

The figures are not saved automatically. Because each figure is constructed within a loop, I could not easily add in a ggsave() to save the figures using specific dimensions and DPI. The tables are generated automatically but the script appears to only generate the corresponding LaTeX syntax and print it to the Console. To actually do the codecheck, I copied the replicate.R code into a .qmd file and displayed the tables using flextable.