Keywords: Philosophy of science, epistemology, ethics, methodology, trial design
Despite advances in Bayesian adaptive designs and model-based dose-finding, oncology dose-finding studies have remained conceptually in thrall to 'the' maximum tolerated dose (MTD). This concept ignores the long-recognized heterogeneity in pharmacokinetics and pharmacodynamics (PK/PD) of persons with cancer, and the great diversity in their individual values and goals of care. Under this conceptual yoke, dose-finding studies constitute a choke-point in drug development, where the potential value in new molecules may suffer severe degradation through the hobbling of subsequent ‘efficacy’ trials by inadequate individual-level dosing.
Strangely, Bayesian innovation in dose-finding studies has proceeded apace without issuing a meaningful challenge to the inherently frequentist conception of an MTD as determined by whole-cohort frequencies of dose-limiting toxicities (DLTs). Thus, even as Bayesianism has made progress toward the ethical imperative of *efficient use of data* in such studies, it has neglected to confront the distinct ethical dimension of *individualism*. This seems a great irony, as the dynamic learning model of Bayesianism is equally suited--and indeed equally essential--to solving the latter problem.
In a recent publication (https://f1000research.com/articles/6-112/v1), I developed a concept of Dose Titration Algorithm Tuning (DTAT), under which (I argue) we ought to conceive early-phase studies as most properly yielding a *dosing algorithm*, not a fallacious single 'suggested dose'. The DTAT argument represents just one example of how one may challenge (what philosopher Alfred North Whitehead called) 'Fallacies of Misplaced Concreteness' (FMC) in pharmaceutical biostatistics.
Questions for those around the table will be: (1) WHAT *OTHER* FALLACIES OF MISPLACED CONCRETENESS BURDEN PHARMACEUTICAL TRIAL DESIGN? (2) WHAT NEW METHODOLOGIES DO WE NEED IN ORDER TO CHALLENGE THESE FALLACIES SUCCESSFULLY? Following introductions around the table, I would aim to spend 10 minutes outlining the DTAT argument, primarily as an illustration of how methodological advances may help us to recognize and transcend fallacious notions. The remainder of our luncheon should be devoted to voicing your own conceptual grievances, and brainstorming the methodologic advances needed to redress them. Priority will be given to those courageous enough to bring (what they regard as) open problems lacking clear formulation or obvious solutions. (Example topic: 'equipoise')