Abstract Background Rare and uncommon diseases are difficult to study in clinical trials due to limited recruitment. If the incidence of the disease is very low, international collaboration can only solve the problem to a certain extent. A consequence is a disproportionately high number of deaths from rare diseases, due to unclear knowledge of the best way to treat patients suffering from these diseases. Hypothesis testing using the conventional Type I error in conjunction with the number of patients who can realistically be enrolled for a rare disease, would cause the trial to be severely underpowered.
Methods Our proposed method recognises these pragmatic limitations and suggests a new testing procedure, wherein conclusion of efficacy of one arm is grounded in robust evidence of non-inferiority in the endpoint of interest, and reasonable evidence of superiority, over the other arm.
Results Simulations were conducted to illustrate the gains in statistical power compared with conventional hypothesis testing in several statistical settings as well as the example of clinical trials for Merkel cell carcinoma, a rare skin tumour.
Conclusions Our proposed analysis method enables conducting clinical trials for rare diseases, potentially leading to better standard of care for patients suffering from rare diseases