Six years ago, JSM 2019 hosted a session organized by Joyce Orsini on W. Edwards' Deming's influence on and legacy with the statistics community. I posted a note in this forum entitled Some thoughts on W. Edwards Deming in Statistics | ASA Connect
On Friday Journal of Biopharmaceutical Statistics, published Full article: Quality principles in Phase I dose escalation design. My paper attempts to apply some of Deming's statistics theory ideas discussed in the thread six years ago to statistical clinical trial design. It uses improving phase I dose escalation designs common in oncology as an introductory example. The paper notes that model-based Phase I dose escalation designs have had a rather spotty safety record so far and modifications have often had to be introduced shortly after major new designs were published in order to avoid safety problems that became apparent only after introduction into the clinical. It suggests ways to improve this record by better taking safety concerns into account at the design stage. To accomplish this, it applies key Deming principles, including quality as fitness for use, understanding and designing for user needs, empirical loss functions, statistics as a science rather than as mathematics, a focus on process, etc.
I believe this is the first paper attempting to apply W. Edwards Deming approach to statistical theory in order to improve a concrete aspect of clinical trial design, as distinct from general quality management philosophy or general thoughts about his legacy, that has been published in a medical statistics journal (as distinct from a quality-focused journal) for some time. If there are others I would very much appreciate it if people would let me know.
------------------------------
Jonathan Siegel
Director Clinical Statistics
Bayer US LLC
------------------------------