Pilot and Feasibility Studies: What They Can Establish Before a Larger Trial

The question before the main question

A team may want to know whether a service improves an outcome but first need to establish whether eligible participants can be recruited, the service can be delivered, and follow-up information can be collected. Those preliminary uncertainties can make a large trial uninformative if they are not understood beforehand.

A feasibility study makes those uncertainties its subject. A pilot may rehearse the planned process, such as enrolling a smaller group and following the intended assessment schedule. The CONSORT extension provides a reporting framework for randomized pilot and feasibility trials conducted before a definitive randomized trial. Its scope is specific; not every project casually called a pilot has that design or purpose.

What could actually be learned?

Consider a proposed program with weekly sessions and a three-month follow-up. Useful preliminary questions might concern how many eligible people can be contacted, how many choose to enroll, whether the sessions fit their circumstances, and whether the follow-up procedure is workable. Each question calls for a different measure and denominator.

Attendance among enrolled participants does not reveal how many eligible people declined because the schedule was unsuitable. A completed questionnaire may not establish that its wording was understood. Looking at the process in stages can reveal a barrier that a single overall completion percentage would conceal. The goal is information that could change the larger study's design, rather than collecting every available metric.

Recruitment is a process, not just a number

Suppose a pilot enrolls 20 people in a month. That result means something different if the team contacted 25 people through an established service than if it contacted 2,000 people through unusually intensive outreach. The number of sites, staff time, eligibility rules, and recruitment route all affect whether the result can inform a larger study.

The pilot's setting may also be unusually supportive. An experienced team or a highly engaged partner can make procedures workable under conditions that will not hold everywhere. This does not invalidate the finding. It identifies what would need to remain true when expanding. An estimate of recruitment capacity is more useful when its supporting conditions are described alongside it.

Why a small effect estimate can mislead

An outcome difference from a small pilot may be imprecise and sensitive to who happened to participate. Choosing whether to proceed solely because that estimate looks favorable can select an unusually optimistic result. The later study may then be planned around an effect that is larger than the underlying effect, or even around a chance difference.

Pilot outcomes can still provide information about measurement procedures, missingness, and the range of observed values. The interpretation depends on the design and purpose. A result labeled statistically significant does not turn a feasibility study into a definitive test, just as a nonsignificant result does not establish that a potentially useful intervention has no effect. The uncertainty belongs in the conclusion.

Progression criteria make the next decision explicit

Before results are known, a team can describe which findings would support proceeding, revising, or stopping. These criteria might concern recruitment, acceptable burden, reliable delivery, or usable follow-up data. They need to connect to requirements of the proposed larger study rather than to convenient thresholds chosen after the pilot ends.

Criteria can inform judgment without mechanically replacing it. A missed recruitment target may be addressable through a feasible change, while an apparently achieved target may conceal a method that cannot be sustained. Clear reporting explains both the criteria and the reasons for the eventual decision. That account is more informative than presenting continuation as proof that everything worked.

Qualitative findings can explain the numbers

Interviews or observations may reveal why a procedure was difficult. A participant might miss a session because of transport, find an instruction ambiguous, or consider a survey intrusive. These are different problems with different possible responses. The numerical pattern alone may not identify which explanation applies.

Participants' accounts should still be interpreted within the sample and method used. A few interviews do not estimate how common an experience is throughout a future population. They can identify mechanisms, overlooked concerns, and changes worth investigating. Combining this information with the operational data can help distinguish a fixable procedure from a more fundamental mismatch.

A useful pilot can lead to a different study

The output may be a revised recruitment route, a shorter assessment, a different timetable, or a decision that the proposed trial is not currently feasible. Those outcomes can prevent a much larger investment in an unworkable design. They do not need to be reframed as evidence of intervention effectiveness to have value.

The report is most useful when it states what was learned, which changes follow, and what remains uncertain. If the pilot answers only whether one experienced team could run the procedure once, that is the boundary of the finding. A clear boundary gives the next study a sounder starting point than a broad claim that the approach has been proven.