An IRT Expert Walks into a Bar…

An IRT expert walks into a bar....

While this sounds like the beginning of a joke what followed was serious and problematic.

I was at a conference and chatted up this consultant who was doing development work for a small biotech a few years back. The biotech asked if he would create a randomization schedule because the IRT provider wanted to charge $20k to do it. So off he went with Excel in hand to create the list that would determine which patient was assigned which material on which the integrity of their entire study would hinge.

The first thing to know is that Excel's random number generator (RNG) will not stand up to regulatory scrutiny for a registrational trial. This is primarily focused on repeatability, and reproducibility. And while regulators won’t tell you explicitly the software you must or mustn't use, they will tell you what it must be capable of and how it must be controlled. But don't take my word for it, the fine folks at Statistics in Biopharmaceutical Research (Carter et al., 2023) are happy to tell you.

One of my favorite things to say is no $10 solutions to $2 problems. Do not over engineer. The opposite is also true and a $2 solution for your multi-million-dollar clinical trial is a bad bet. 

Carter, K., Scheffold, A. L., Renteria, J., Berger, V. W., Luo, Y. A., Chipman, J. J., & Sverdlov, O. (2023). Regulatory guidance on randomization and the use of randomization tests in clinical trials: A systematic review. Statistics in Biopharmaceutical Research, 16(3), 428–440.