Articles

On the FDA's move to withdraw a drug approval after Phase 3 data were re-adjudicated post-unblinding — and why the statistician is often the last line of defense between commercial pressure and the integrity of the data.

Defining the effect you actually want to measure. Estimands, intercurrent events, and the five strategies for handling them in everyday language, with worked examples and a figure for every strategy.

The FDA asked how AI should be used to optimize early-phase clinical trials. We proposed a focused use case: reviewing the statistics in Bayesian dose-escalation and dose-expansion protocols, with the human reviewer in charge of every call.

How we write Statistical Analysis Plans for trials that borrow strength from data outside the randomized comparison, anchored in regulatory precedent and locked before any outcome data are examined.