What is a pilot study?

NCCIH quotes the Dictionary of Epidemiology in defining a pilot study as “a small-scale test of the methods and procedures to be used on a larger scale.” It stresses that the goal is not to test an intervention’s effect but to assess whether the approach is feasible and acceptable: the question is “Can I do this?”, not “Does this work?” In the framework of Eldridge and colleagues, pilot studies form a subset of feasibility studies, in which the future study, or part of it, is run on a smaller scale.

NCCIH’s example questions include whether the target population can be recruited, whether participants stay, and whether the intervention can be delivered as intended. NCCIH recommends quantitative benchmarks set in advance and a sample size based on participant flow, budget and feasibility goals rather than a power calculation. Randomized pilot trials have a 26-item CONSORT extension (Eldridge and colleagues), which does not recommend formal hypothesis testing for effectiveness.

NCCIH holds that an effect size estimated in a pilot study is unstable and gives no useful basis for power calculations; it recommends basing that calculation on a clinically meaningful difference. Thabane and colleagues likewise warn that estimating treatment effects from pilot studies needs caution. Eldridge and colleagues observe that the term is used inconsistently: sometimes “pilot” means a feasibility study, sometimes a small trial of efficacy.