AAPS webinars and a short course take on the myths and misconceptions of using statistics.
“I use statistics because I have to for compliance reasons; it doesn’t help, it just adds unnecessary work”
“Sure, using design of experiments (DOE) would be nice, but it’s too expensive and takes too long.”
“Statistical analyses don’t work with complex processes. Solutions will be found using science, not statistics. Statistical results don’t always make sense scientifically.”
“If you use statistics, you can make the data say whatever you want.”
The American Association of Pharmaceutical Scientists (AAPS) and SynoloStats LLC hope to dispel the myths quoted above and other myths and misconceptions about statistics in 2019 with a webinar series and an on-site 1.5 day short course at AAPS PharmSci 360. In the March AAPS Newsmagazine cover article, Tara Scherder and Katherine Giacoletti of SynoloStats LLC described how using advanced statistical modeling can benefit patients and pharmaceutical manufacturers by reducing sampling-associated costs while assuring quality. That article began an initiative to increase statistical awareness and capabilities among AAPS members.
Starting in June, AAPS will host five webinars (one per month) on statistics for chemistry, manufacturing, and controls activities followed by a short course at AAPS PharmSci 360, Beyond Compliance: Essential Statistical Thinking for Business and Patient Benefit. Although each webinar and short course is an independent educational opportunity, scientists who participate in one or more webinars in addition to the short course will benefit the most by building on conceptual knowledge gained from the webinars with deeper learning and hands-on skills in the short course.
Webinar topics were chosen based on the results of a recent survey of AAPS members.
Each webinar topic will include real examples from the pharmaceutical industry, and no software or previous knowledge of statistics is needed to benefit from these sessions.
The short course at AAPS PharmSci 360 in November will dive deeper into these topics, primarily DOE and exploratory data analysis, and include additional topics such as common mistakes in data analysis. Hands-on practice will develop skills that can be immediately applied on the job—and an appreciation of the benefits of statistical thinking and the motivation to continue learning.
The webinar series and the short course will emphasize statistical thinking, rather than statistical tools. The practice of statistics is not a compliance exercise or a set of mathematical tools. It is a way of thinking and a set of methodologies that enable deeper understanding and proactive risk management. Maximizing the benefit of using statistics requires an understanding of fundamental statistical concepts and applying these concepts within the context of the data, the analytical or manufacturing process, and the business problem and process. For instance, you may have heard the aphorism, “All models are wrong, but some are useful.” A “model” can mean anything from the interpretation of X ± 3s, to a linear regression of stability data, to a complex systems model of a process. The ability to examine data for signs that a model may not be useful—or even worse, may be wrong—is an invaluable skill for any scientist applying statistical methods. Learning how to efficiently explore data and apply statistical thinking before embarking on potentially useless—or worse, misleading—statistical calculations can be key to optimizing data analysis and related activities and avoiding fruitless work.
Webinar Dates and Topics
Fundamentals of DOE for Pharmaceutical Scientists
Friday, June 28, 12:30–2:00 pm
Improve Problem Solving Using Exploratory Data Analysis
Thursday, July 18, 12:30–2:00 pm
Statistics for Optimization of Analytical Methods across the Lifecycle
Thursday, August 15, 12:30–2:00 pm
Fit-for-Purpose Statistics for Stability Studies
Thursday, September 12, 12:30–2:00 pm
Statistical Methods for Evaluations of Comparability
Thursday, October 10, 12:30–2:00 pm
Tara Scherder and Katherine Giacoletti of SynoloStats LLC have more than 40 years of combined experience as applied statisticians in the pharmaceutical industry. They are active in industry initiatives to develop and implement best practices for the integration of statistics into pharmaceutical quality systems. In addition to their statistical expertise, they leverage extensive knowledge of manufacturing and analytical processes, business acumen, and an interactive teaching style to create an engaging and productive learning environment.