Intent to treat: TheReductio ad absurdumthat became gospel

Abstract The ‘Intent to Treat’ paradigm for the analysis of a controlled randomized clinical trial is the direct result of applying the Neyman‐Pearson formulation of hypothesis testing. If other formulations are used, the ‘Intent to Treat’ paradigm makes no sense. Criticisms of the Neyman‐Pearson formulation and whether it is applicable to scientific investigations have appeared in the statistical and philosophical literature since it was first proposed. This paper reviews the nature of that criticism and notes why the Neyman‐Pearson formulation, and with it the ‘Intent to Treat’ paradigm, is inappropriate for use in the analysis of clinical trials.