Scientific Inference in Clinical Trials with R: example repository

한국어 설명은 README.ko.md 에 있습니다.

The companion repository of Kyun-Seop Bae, Scientific Inference in Clinical Trials with R (임상시험에서의 과학적 추론 with R). Its purpose is to let a reader remake every number and every figure in the book; the text of the book is not here.

All you need is R. No commercial software, no license.

The code here is the English edition's: the comments in the snippets, the truncation mark in the frozen output and the labels in the figures are all English. The book was written in Korean first and the Korean edition prints the same code with Korean comments; a reader of either edition can use this repository, because the file names, the chapter numbering, the numbers in the console output and the data do not change with the language. Only one set of code is kept here, so that there is one place to look.

Getting started

git clone https://github.com/AMC-CPT/CTDA
cd CTDA
Rscript R/build-all.R

That remakes the console output in output/ and the figures in figures/. They are committed already, so they can simply be read without running anything. To follow one chapter, open that chapter's files in R/snippets/ and run them yourself.

R/build.R is what does the freezing and R/build-all.R drives it. The difference is the ch17-* snippets (chapter 20 of the book), whose many nonlinear fits need one R process per group; the comment at the head of R/build-all.R says why. To remake a single chapter, run R/build.R.

Structure

R/snippets/<name>.R   the R code printed in the book. 401 files
R/build.R             runs the snippets in order and makes output/ and figures/
R/build-all.R         the same, with the ch17-* snippets run one group per R process
R/_freeze.R           its helper (fixes the seed, shares the session, tidies the output)
output/<name>.txt     frozen console output, as printed in the book. 339 files
figures/<name>.pdf    frozen figures. 91 files
data/                 the example data
pyfig/                the matplotlib scripts of the Ch 1-2 diagrams, and FIGURE-PROVENANCE.md

The number in a snippet's name is not the chapter number of the printed book. The names keep the numbering of the manuscript's files, which did not change when chapters were added and moved, and the book refers to the snippets by these names. The chapters of the book that each group of files belongs to:

Files	Chapter of the book
qdm-*	1, 2
ch02-*	3
ch03-*	4
ch04-*	5
ch05-*	6
ch06-*	7
ch07-*	8
ch08-*	6, 7, 11, 18 (the sections of a former chapter, now spread out)
ch09-*	10
ch10-*	11
ch11-*	12
ch12-*	13
ch13-*	14
ch14-*	15
ch15-*	16
ch16-*, ch16b-*	17
ss-*	18
ctsim-*	19
ch17-*	20
be-*	21
appA-* to appE-*	Appendices A to E

Chapters 9, 22 and 23 print their code in the text itself and chapter 24 has none, so there are no files for them here. The comments of R/build.R and R/build-all.R number chapters the same way as the files: their chapter 17 is the ch17-* files, chapter 20 of the book.

The scripts in pyfig/ are the ones the Korean edition uses; they carry its labels and translate them through pyfig/labels_en.py as each label is drawn, which is why their comments are still Korean. CTDA_LANG selects the language and defaults to English here, so running them reproduces the figures in figures/.

Why the code and its output cannot drift apart

The book prints R/snippets/<name>.R as a whole file, and takes the output it shows from output/<name>.txt, which was produced by running that same file. Code and output come from one source, so nothing is retyped and nothing can fall out of step. R/build.R is that freezing step, and R/_freeze.R manages the random seed and the session.

The snippets run in order, as though one R session ran through them all: a later snippet uses the objects an earlier one made. So running one file on its own can fail with an object-not-found error. When that happens, run the chapter's earlier snippets first; the order in R/build.R is that order.

R packages

Most of it runs on base R; some chapters need these.

install.packages(c("sasLM", "MASS", "survival", "nlme", "lattice", "psych", "nnet",
                   "car", "gmodels", "PropCIs", "KMsurv", "ggplot2", "ggsurvfit",
                   "dobson", "rgl", "mathr", "flexsurv", "numDeriv", "wnl", "nmw",
                   "deSolve", "LBI", "vcdExtra", "meta"))

mathr is on CRAN as well; its development version is at https://github.com/ksbae/mathr.

Where the data come from

Everything in data/ is either the author's own or public statistics. It used to live at http://r.acr.kr/ and was bundled into the repository so that nothing depends on a remote host.

File	Contents
2017KoBody.csv	2017 Korean National Fitness Survey (original: http://r.acr.kr/2017KoBody.csv)
LogP-F.csv	LogP and oral bioavailability for 45 drugs, compiled by the author from public sources
Paxlovid.csv, pax0.csv	survival analysis examples
ExTTE.csv, TTE.R	a time-to-event example and its helper functions
Surv-Trt.csv, Surv-Quartile.csv	survival analysis examples
bax.csv	a dose--response example (change in systolic blood pressure, baseline, eGFR)
Dose-ContResp.csv, Dose-Resp-Logistic.csv, AUC-Resp-Logistic.csv, DoseRes.csv, Pop-SigEmax.csv	dose- and exposure--response examples

The provenance and copyright handling of the Chapter 1-2 diagrams is recorded figure by figure in pyfig/FIGURE-PROVENANCE.md. Not one third-party image is used.

Solutions to the exercises

They are not in the book and they are not here. Instructors may request them from ksbae@acr.kr.

Copyright and license

The copyright in the text of the book belongs to the author and the publisher, and the text is not here. The code, the data and the frozen output and figures in this repository are published under the GNU General Public License v3.0 or later (LICENSE). Running them and changing them is free, and a modified version distributed to others carries the source under the same terms.

Copyright (C) 2026 Kyun-Seop Bae. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. It is distributed WITHOUT ANY WARRANTY; see LICENSE for details.

The other books in the series

1 Scientific Computation with R	https://github.com/AMC-CPT/SciCompR
3 Pharmacokinetics with R	https://github.com/AMC-CPT/PKwR
4 Pharmacometrics with NONMEM and R	https://github.com/AMC-CPT/PMx
5 Essentials of Clinical Drug Development (online appendix)	https://github.com/AMC-CPT/CDD