Many multi-stage clinical trials rely on dichotomous outcomes to determine the next stage treatment. We explore how to allow for a continuous outcome to assign next treatment while maintaining other trial properties.
Small sample, sequential, multiple assignment, randomized trials (snSMARTs) are multistage trial designs to identify the best overall treatment. This design is a modification of sequential, multiple assignment, randomized trials which are used to identify the best sequence of treatments for patients to receive based on their response to prior treatments.
In snSMARTs, binary response/nonresponse outcomes are measured at intermediate and final timepoints. If the patient is responding at the the end of the first stage timepoint, they continue on the same treatment. Otherwise, they are re-randomized to one of the remaining treatments. We expand the snSMART design to allow for continuous outcomes. The probability of staying on the same treatment is a function of the first stage outcome thus eliminating the need for a categorical tailoring variable defining response/nonresponse. This re-randomization scheme allows for trials to continue without requiring a dichotomous variable. This method reduces bias and increases efficiency relative to a dichotomization cut off re-randomization scheme. We also show that patient outcomes are similar using this design relative to a traditional snSMART design.