Statistical challenges in the evaluation of treatments for small patient populations.

Abstract

The development of a new treatment typically involves evaluation of its efficacy in a large clinical trial in which patients are randomly assigned either the new treatment or the standard of care. Results from these large randomized clinical trials allow for a definitive and unbiased assessment of the clinical benefit of the new treatment over the standard one. For rare diseases or for small patient subgroups identified within the context of a common disease, it may not be possible to conduct a large randomized trial. In this Review, we discuss alternative clinical study designs and statistical challenges that arise when attempting to assure that study results yield robust conclusions about the safety and effectiveness of a new medical product.