Sample size calculation based on precision for pilot sequential multiple assignment randomized trial (SMART)

Abstract

The sequential multiple assignment randomized trial (SMART) is a design used to develop dynamic treatment regimes (DTRs). Given that DTRs are generally less well researched, pilot SMART studies are often necessary. One challenge in pilot SMART is to determine the sample size such that it is small yet meaningfully informative for future full-fledged SMART. Here, we develop a precision-based approach, where the calculated sample size confines the marginal mean outcome of a DTR within a prespecified margin of error. The sample size calculations will be presented for two-stage SMARTs, and for various common outcome types.