When it comes to assessing the impact of how well patients (and arguably centres) in a randomised clinical trial follow the protocol, the standard method for some decades has been via the identification of the ‘Per Protocol’ ...