A simple, efficient method for comparing dose-finding designs described in preprint and implemented in escalation v0.1.8
If you want to send me data in an Excel sheet, laying it out like this will let me start analysing it right away.
Emax is an awesome, flexible non-linear model for estimating dose-response curves. Come learn how to fit it in R.
Empirically, what sample size is used in dose-escalation trials?
Descriptive data and dose-level outcomes from 122 manuscripts reporting results of dose-finding clinical trials in cancer.
Simulation is the popular approach but brute-force enumeration is more accurate and can even be quicker.
During 2019, I was working on simulations using CRM designs in several different trials. I found I would frequently get the designs mixed up:
“We are targeting 20% toxicity in this trial, right? No, that was the other trial…we are targeting 33% here…”
Similarly, once or twice, I got to the stage where I wanted to run simulations only to discover we had not specified some important design aspect, like when the trial should stop.