ON SEQUENTIAL DESIGN BASED ON MARKOV CHAINS FOR SELECTING ONE OF TWO TREATMENTS IN CLINICAL TRIALS WITH DELAYED OBSERVATIONS

A truncated sequential design for selecting the better of two medical treatments is constructed by using a fundamental relation of Markov chains, with delayed observations based on a time lag of response time. The design is practical and applicable to a wide range of sequential clinical trials. The iterative formula of Markov chains is based on the trinomial random walk on a lattice diagram which adopts the unitied and the tied pairs on the double-dichotomous responses. The study is to determine an optimum sequential design based on the evaluation of minimum expected loss.