midas2: Bayesian Drug-Combination Platform Design(MIDAS-2)
Implementation of Bayesian drug-combination platform design. More and more immuno-oncology drug combinations make the traditional two-arm phase II trials inefficient, which stimulate the emerge of platform trials. In the case of multiple trial objectives such as candidates screening and subgroup analysis, we propose an information borrowing drug-combination Bayesian design for platform trials with subgroup exploration. MIDAS-2 consists of one control arm and several experimental agents. We use Bayesian spike and slab prior to identify factors that should be included in regression model and borrow information between combinations in the existence of subgroup interaction. Promising drug combinations are allowed to graduated early to move to next stage and new combination strategies can be added accordingly. Catch-up rule, curtail rule and early stopping rules are also applied to accelerate the trial process.
||MCMCpack, coda, R2jags
||Su Liwen <cpusullivan at 163.com>
Please use the canonical form
to link to this page.