Write feasibility progression criteria before recruitment

Decide what could prevent the definitive trial

A feasibility study is useful when its findings change a concrete design decision. Start by listing the uncertainties that could make a larger study impractical: access to eligible participants, delivery of the intervention, completeness of a key measurement or acceptability of follow-up. Rank them by the consequences of failure. An attractive recruitment total cannot compensate for an intervention that cannot be delivered as intended.

The CONSORT pilot and feasibility extension asks authors to report prespecified progression criteria and the implications of the findings for a future trial. A criterion should identify the measure, denominator, observation window and action. It should also explain who will review contextual evidence when a result lies near a threshold.

Build a decision table

For each important uncertainty, write the calculation in words. Distinguish participants approached from those screened, eligible, consenting and randomised. Name the data source and the person responsible for resolving incomplete records. Avoid a threshold such as “good recruitment” without its time period or denominator.

Create actions for an acceptable result, a potentially remediable result and an unacceptable result. Record the rationale for the boundaries: required recruitment capacity, budget, delivery constraints or stakeholder judgement. Traffic-light colours are a communication device, not a universal scientific standard. Decide whether a single serious problem can stop progression even when other measures are favourable.

Hypothetical example

A fictional pilot needs to learn whether a site can randomise 12 participants per month. Before recruitment, the team specifies that fewer than eight per month triggers a review of referral pathways and available capacity. The pilot averages nine. This does not automatically justify proceeding unchanged: the team investigates whether recruitment depended on temporary staff unavailable to the definitive study.

The example thresholds are invented for illustration. They are not recommended benchmarks for another protocol.

Record the progression decision

Present the observed measures and their uncertainty alongside each original criterion. Explain amendments and distinguish prespecified rules from decisions made after seeing results. Include participant or staff feedback when it explains a delivery problem, while keeping confidential details out of public reports.

Finish with a short decision record: proceed, modify and reassess, or stop; the evidence supporting that choice; unresolved risks; and accountable owners. Use the recruitment monitoring worksheet to connect the decision to an operational response. A feasibility decision is not evidence that the intervention is clinically effective.

Source verification: 6 November 2026. Examples are hypothetical.