Here we highlight some similarities and differences between the commonly used approach to intervention development / evaluation and MOST. 

Standard approach. The sequence of events may begin with a series of pilot tests (e.g. Sussman, Dent, Burton, Stacy, & Flay, 1995).  In this proposal we use Vogt's (1993) definition of a pilot study as "a preliminary test or study to try out procedures and discover problems before the main study begins. It is a research project's 'dress rehearsal.'" (p. 172).  Note that according to this definition, pilot studies are aimed at feasibility and implementation issues rather than optimization, and may or may not be randomized experiments.  Next, the intervention is subjected to a confirmatory trial.  Further refinement of the intervention takes place after the confirmatory trial, based on analyses of data collected in the course of the trial, in the form of intermediate outcomes, covariates, and the like.  Any conclusions drawn from these data are for the most part based on secondary analyses rather than randomized experimentation.  The newly refined intervention may be tested in an additional confirmatory trial, setting up a cycle of confirmatory intervention trial/exploratory analysis/program revision/confirmatory intervention trial. 

MOST. Before the screening phase of MOST is begun, there may be pilot testing, conducted in a manner essentially similar to the standard approach.  Pilot testing is important in order to arrive at the exact versions of intervention components and dosages that are to be tested in the screening and refining phases.  However, pilot testing in no way replaces the screening and refining phases of MOST, which have an objective that is fundamentally different from pilot testing.  The screening and refining phases consist of randomized, controlled tests aimed squarely at optimization of intervention components and dosages.  Only once an optimized intervention has been identified based on the screening and refining phases, and even then only if the optimized intervention is sufficiently promising, is a confirmatory trial undertaken.  Although future improvements are not ruled out, the objective of MOST is to provide a definitive answer about the efficacy and/or effectiveness of an intervention, without the need for additional confirmatory trials after the confirming phase.  In other words, the aim is for any significant refinement of the intervention to take place before the confirmatory trial.