Home/Insights/Methodology reference/Feasibility study methodology
Feasibility study methodology
It asks whether the definitive study can be run, not whether the intervention works — and reporting an effect estimate as though it answered the second question is the standard error.
Feasibility and pilot are not synonyms
A feasibility study asks whether the elements of a future trial can work together: can participants be recruited at the expected rate, will they accept randomisation, can the intervention be delivered as intended, can outcomes be measured.
A pilot study is a miniature version of the planned trial, run to test the procedures end to end. All pilots are feasibility studies; not all feasibility studies are pilots.
Progression criteria
The defining feature of a well-designed feasibility study is that success criteria are stated before it starts: the recruitment rate, retention rate, adherence and data completeness that would justify proceeding, with thresholds that trigger modification or abandonment.
Without pre-stated criteria, every feasibility study concludes that the trial is feasible, because the judgement is made after seeing the result.
Why it is not powered for efficacy
Feasibility samples are small by design. An estimate of effect from such a sample has a confidence interval wide enough to be consistent with substantial benefit and substantial harm simultaneously.
Reporting a p-value for the clinical outcome invites exactly the wrong inference. The CONSORT extension for pilot and feasibility trials advises against hypothesis testing of effectiveness outcomes, and asks instead for feasibility outcomes with estimates of precision.
A non-significant result in a feasibility study is uninformative about efficacy. So is a significant one.
What to report
Recruitment: approached, eligible, consented, randomised, with reasons for non-participation. Retention and adherence. Intervention fidelity. Completeness of outcome measurement, and any problems with the measures themselves. The parameters needed to size the definitive trial — such as the standard deviation of the outcome or the intracluster correlation.
In practice
Fix, flag or query.
The distinction that keeps an edit defensible: what an editor may change silently, what must be shown to the author, and what only the author can answer.
- State before starting: the progression criteria for proceeding to the full trial.
- Do not: hypothesis-test the clinical outcome.
- Report: the parameters needed to power the definitive study.
Reporting standards
- CONSORT 2010 extension to randomised pilot and feasibility trials
- EQUATOR Network, for pilot and feasibility reporting guidance
Reporting guidelines named here are hosted by the EQUATOR Network. Writing to the checklist for your design before submission is the cheapest available improvement to a methods section.
Related