Dynamic treatment regimens in small n, sequential, multiple assignment, randomized trials: An application in focal...

Area of Science:

• Nephrology
• Clinical Trial Design
• Biostatistics

Background:

• Focal segmental glomerulosclerosis (FSGS) is a rare kidney disease with limited effective treatments.
• Rare diseases face challenges in clinical trials due to small sample sizes and recruitment difficulties.
• A sequential, multiple assignment, randomized trial (SMART) offers an efficient design for rare disease research.

Purpose of the Study:

• To review and expand the small sample, sequential, multiple assignment, randomized trial (snSMART) design for studying FSGS treatments.
• To adapt the snSMART design to include a standard of care alongside novel therapies.
• To present statistical models for comparing novel therapies against the standard of care and analyzing dynamic treatment regimens (DTRs).

Main Methods:

• The proposed snSMART design involves multi-stage randomization, including re-randomization for non-responders.
• Bayesian joint stage models are used for efficient information sharing across trial stages.
• Both Bayesian and frequentist models are presented for comparing treatments and analyzing DTRs, with a novel sample size calculation method for the frequentist approach.

Main Results:

• The study demonstrates how to estimate and compare tailored treatment sequences (DTRs) using the modified snSMART design.
• Comparison of novel therapies against the standard of care is facilitated through presented Bayesian and frequentist models.
• A method for sample size calculation in the frequentist snSMART model with Dunnett's correction is proposed.

Conclusions:

• The modified snSMART design provides an efficient framework for evaluating treatments in rare diseases like FSGS.
• The proposed statistical methods enable robust comparison of novel therapies and analysis of DTRs.
• This work contributes a practical sample size calculation method for implementing snSMART in clinical practice.