ABSTRACT
The current healthcare system in the United States has been criticized as being overly complex, expensive, and inefficient, and pragmatic clinical trials have been considered a plausible solution that would support a transition to a system in which continual learning is possible. The internal trend of intervention effect as well as natural disease progression should be accounted for in both design and analysis. Finally, the stepwise crossover to intervention in SW trials presents great potential for implementing various adaptive design methods. The above sample size formulas are obtained under the missing completely at random assumption. Adopting this assumption leads to simplified mathematical derivation and helps provide a universally applicable sample size solution.
