Are there any potential solutions to the lack of statistical power of clinical trials in rare diseases?
The lack of power of rare disease clinical trials is not a fatality. A non-significant clinical result can have two main causes: – the product is not efficacious enough – the number of subjects is too small In order to know if the lack of significance is due to lack of product efficacy or the lack of subject numbers, it is firstly possible to use statistical techniques particularly adapted to small samples. Most of these statistical tests belong to the non-parametric test family and are regularly used because there is no requirement for a normal distribution (usually observed in large samples). We can also use simulation techniques such as re-sampling techniques which are based on random generators, in order to clone the master sample in several thousand derived samples. These techniques, derived from Monte-Carlo techniques, are frequently used in operational research and in financial modelling and can be associated with non-parametric tests to establish potential significant results w
