Approaches to sample size calculation for clinical trials in small populations

Frank Miller

The choice of sample size for clinical trials is critical for balancing higher precision in inference versus costs and feasibility. Sample size should therefore be chosen in a well-informed way. This is even more important if the target population for the investigation is difficult to study, e.g. a rare disease population where not enough patients exist to conduct a trial of traditional size, or a setting where possibilities for inclusion in trials are restricted like paediatric populations.

We review different approaches to sample size and discuss their application in case studies where patient populations with rare diseases are investigated. The approaches we are considering are

  1. the traditional sample size calculation based on power to show a statistically significant effect versus control,  
  2. sample size calculation based on assurance where uncertainty about assumed treatment effects are modelled,
  3. optimal sample size based on a decision theoretic approach.

The talk is based on joint work together with Simon Day, Siew Wan Hee, Jason Madan, Martin Posch, Nigel Stallard, Mårten Vågerö and Sarah Zohar for the InSPiRe project, a project which has received funding from the European Union's Seventh Framework Programme under grant agreement no 602144.