Do people who choose their treatment in a large randomised trial with parallel preference groups differ at baseline from those who agree to random treatment allocation? Results from the protect study

Background Many patients decline participation in randomised trials due to treatment preferences. It is not known whether it is possible to combine evidence from those choosing treatment and those randomised in RCTs with parallel non-randomised preference groups without compromising external validity. This study explored whether people who chose their treatment within an RCT in which recruiters were trained to explore treatment preferences were different at baseline from those who agreed to randomisation.