Before the updated data from recons

Before the updated data from reconsent was available, the Applicant initially addressed the early dropouts with a simulation study assuming the neratinib early dropouts behaved as if they were on placebo. Neratinib early dropout patients were assigned “updated” iDFS times via resampling from the placebo arm. Balance in the stratification factors was maintained by matching patients in each group by these factors prior to resampling. After the updated data from reconsent became available, FDA conducted a similar simulation as a sensitivity analysis. In the FDA’s simulation, the remaining 80 neratinib early dropout patients were assigned “updated” iDFS times via resampling from the 50 neratinib patients with real updated iDFS times. Results from both simulations are summarized in Table 19.