Transcript

Applications of statistics in chemistry, manufacturing and controls (CMC)

ANNE MILLEY: After reading the abstract of the talk you gave at last year's Fall Technical Conference on saving lives with statistics, I wish I could have heard it. The challenges facing life sciences from drug development to manufacturing—you've already talked about it a little bit. But anything else to elaborate?

JULIA O'NEILL: Yeah. What got me started putting together a talk like that was a couple of years ago, I was invited by some of the statisticians at the FDA to give them a full-day course on applications of statistics in chemistry, manufacturing and controls. That's CMC—that is the jargon in pharmaceuticals. And it was such a great request because these are extremely well-qualified statisticians. I mean, if you needed to prove a theorem or do something theoretical, they would be far better than I to do it.

But they have worked most of their careers exclusively in clinical trials—it's common—which is a very focused area of statistics. Now, with the FDA's guidance in 2011, and further guidances coming out on quality metrics, there is such a push towards more use of statistics throughout development and manufacturing and ongoing product lifecycle management, that these statisticians are now being asked to review those CMC sections of submissions for approval for new products. And what they asked me for is they said, you know, we get the statistics. We don't know the context. We're not sure what's important. When we review these, what should we be looking for? So they said give us the big picture, the concepts, what's important—you can use formulas if you want.

It was so much fun to put that together because there are many places throughout development and manufacturing where statistics is used. Things like stability studies, specification setting, control strategy, design of experiments for developing control strategy. So I had a chance to put together this course and give it in 2016, and that was just a lot of fun. Based on that, I've been writing up parts of that, and I'm working on a book, actually. I wish it was going a little faster.

ANNE MILLEY: Well, good. Then more people can get exposed to those good ideas and things that they probably really need to know, too.

JULIA O'NEILL: Right. Actually, recently, there've been a couple of good references that have come out, but more aimed towards statisticians, and, therefore, they're typically written in chapters. So each chapter has an author and a slightly different voice. And the audience, I think, is intended more for statisticians. What I have in mind is a more accessible introduction for scientists, engineers, and statisticians, especially clinical statisticians, who want to get into the application of statistics in CMC, but need some context.

ANNE MILLEY: Got it. Very good. Well, that's just helping to spread the value of statistics to more parts, where there are data and where better things can be done, better decisions can be made. We look forward to promoting that book when it comes out.

JULIA O'NEILL: Well, great.