Table 3 Motivating example—feasibility trial for oral protein energy supplements as flavoured drinks to improve nutritional status in children with cystic fibrosis

A feasibility trial is being set up to see whether children aged 2 to 15 years with cystic fibrosis will take oral protein energy supplements as flavoured drinks to improve their nutritional status, compared to receiving dietary advice alone. Children are to be randomised in a 1:1 allocation ratio using a parallel two-arm design. The research team wants to be sure they can meet three feasibility objectives before they go ahead and plan the main trial: reasonable recruitment uptake, high treatment fidelity (i.e. extent to which dietician practitioners comply with the treatment protocol) and adequate retention of children at follow up. The team asks their senior statistician to help them decide on an appropriate methodology including pilot sample size. The statistician suggests a traffic light approach incorporating hypothesis testing of the feasibility outcomes.

Together, the team devise three progression criteria that should be met before the main trial can be considered feasible as follows:

a. At least 35% of the children screened as eligible should be recruited but the trial will not be feasible if recruitment uptake is 20% or less.

b. A high level of treatment fidelity should be maintained with 75% or more children being given the correct treatment plan by the dietician, but if 50% or less children are given the plan as specified in the protocol then the trial is not feasible.

c. 85% or more of the children should be retained in the study at follow up, with 65% or less retention indicating that the main trial is not feasible.

The decision criteria and required sample size around these are detailed through two possible approaches within Table 4 (simple 3-tier approach) and Table 5 (extended 4-tier approach). The statistician is to use the normal approximation method (with continuity correction) for the sample size calculation and analysis.