I was wondering whether is really necessary to improve the number of patients from a sample size calculation (usually we improve in 20% thinking about dropouts) when an ITT analysis is planned.

Similar questions and discussions