Julia O'Neill on statistical thinking and drug development
Jan 8, 2019 5:14 AM
In medicine developmJulia O'Neill knows the worth of statistical thinkingent, statistical methods are crucial for achieving consistency. Julia O'Neill, Founder & Principal Consultant at Direxa Consulting LLC., has used statistical thinking consistently during her 30-year career in chemical engineering, biotech and pharma, and has implemented successful strategies involving statistical methods to achieve better results.
"Figuring out how to produce drugs consistently and with commercially viable yields is a natural problem for statistics to solve," she says.
On Feb. 7, Julia O'Neill will share her experience and show how statistical models in combination with Quality by Design can achieve impressive results in biopharma development. Don't miss out on this complimentary seminar in Copenhagen, which will also be streamed online.
Julia was kind enough to answer a few questions prior to the event:
You learned from the best: George Box and Stu Hunter – two very famous statisticians – were professors at the University of Wisconsin. How did they influence your career?
Actually, I learned from all three authors of the Box, Hunter & Hunter classic text Statistics for Experimenters. My first class in statistics at the University of Wisconsin was with Bill Hunter, who inspired me to move from the chemical engineering graduate program to statistics. Later, I had a class with George Box on Statistical Methods for Quality Improvement, and attended the weekly “Beer & Statistics” sessions at George Box’s house a number of times during my years in Madison. After graduation when I worked as a statistical consultant in the research division of Rohm and Haas (a specialty chemicals company), Stu Hunter came in to lecture, and I was one of the lucky few to fill a teaching assistant role for his lectures for several years. All three were tremendous inspirations and influences on my approach to statistics.
The main lesson I took from them was to never lose sight of the scientific context of the statistical question. Many of their most important contributions to advancing the practice of statistics resulted from the need to solve an important problem. The key step in developing a useful statistical solution is to carefully frame the problem first.
You bridge statistics with chemical engineering in your work. What’s the advantage of using statistical methods, such as Quality by Design, in drug development, for example?
It’s hard for me to think of statistics as an advantage, because frankly I can’t imagine how a drug could be developed without integrating statistics into the development strategy. Even though many revolutionary drugs were initially discovered serendipitously, figuring out how to produce them consistently and with commercially viable yields is a natural problem for statistics to solve.
Tell us about the case study you will show at the seminar in Copenhagen. What do you hope attendees will learn from it?
It’s easy to become overwhelmed by all the statistical methods that are readily available in JMP or other software. I’ll be sharing my strategy for applying statistics selectively throughout the process of developing and commercializing a medicine. The focus will be on the purpose behind each statistical test or analysis, and a simple roadmap for knowing when statistics will allow you to accelerate results or conserve experimental resources.