Showing posts with label statistics. Show all posts
Showing posts with label statistics. Show all posts

Friday, November 13, 2009

GSK's Peterson Has Just What Your Design Space Needs

In case you missed it, we recently posted an article on multivariate predictive distribution by GlaxoSmithKline's John Peterson, director of the research statistics unit in the company's drug development sciences department. (The article will also appear in the Nov/Dec issue of Pharmaceutical Manufacturing magazine.) When we first approached Peterson--a highly accomplished statistician, among other things--about writing an article, it was with an eye towards explaining Quality by Design and the concept of a design space from a statistical viewpoint, but for readers who are non-statisticians. Peterson willingly assented and, if you read the article, I hope you agree that he has not only done as planned, but has also put forth some pretty heady concepts that statisticians involved in drug industry Quality by Design projects will benefit from tremendously.

The article seeks to encourage us to, as he says, "understand randomness and think stochastically." It's a testament to the power of statistics to help manufacturers understand and control their processes, in the name of making drug development smarter.

Great appreciation to Dr. Peterson for sharing his vision for us. I hope that you get as much from the article as I, a non-statistician, have, and we'd love to hear your comments.

Are You Innovating and Imaginating All By Yourself?

One of the many people I spoke with this past week at the AAPS show in Los Angeles was Ted Grasela, CEO of Cognigen Corp., which advises manufacturers on model-based drug development. Grasela delivered a presentation on the importance of modeling and simulation to development, and on strategies that pharmacometricians and statistical experts can use to influence drug development in a positive way. It's time for modeling and simulation to take its rightful and significant place in drug development decisionmaking, Grasela said.

Most importantly, statistical modeling can be a kind of glue that binds the entire drug development process and the people involved. “People are innovating and imaginating all over the place," Grasela says, "but often these islands of innovation that never get linked with one another.”

Here's my full writeup of Grasela's presentation. I stopped by the Cognigen booth afterwards for a chat with Grasela and Cognigen colleagues, and came to appreciate their ideas of bringing more science to the development process. Here's the Cognigen home page for those interested.