03 · What You Need to Know
Developing Defensible Progression Criteria for Feasibility Research
What Are Feasibility Study Progression Criteria?
Progression criteria are prespecified conditions or decision principles used to evaluate whether findings from preliminary feasibility research support moving toward a definitive study.
They connect observed feasibility evidence to the requirements of the proposed research.
For example, a definitive trial may require recruiting 240 participants within 12 weeks. A feasibility study could assess whether the proposed recruitment process is capable of achieving approximately 20 enrollments per week.
The observed recruitment rate becomes meaningful because it can be compared with an explicit operational requirement.
Progression criteria should therefore explain what level of performance is needed and why that level matters.
They are not simply targets chosen to make a feasibility study appear successful.
How Are Progression Criteria Different From Feasibility Outcomes?
A feasibility outcome is an observation or measurement collected during preliminary research.
A progression criterion specifies how that evidence will contribute to a decision about the definitive study.
Feasibility outcome
What is observed or measured, such as recruitment speed, follow-up completeness, intervention adherence, or staff workload.
Progression criterion
The condition or decision rule used to judge whether the observed feasibility evidence supports progression, modification, further investigation, or discontinuation.
For example, observing that 85% of participants provide usable primary outcome data is a feasibility result.
Whether 85% is sufficient depends on the definitive analysis, the consequences of missingness, and the progression criterion established for the study.
The distinction prevents researchers from treating an observed percentage as inherently satisfactory.
Why Should Progression Criteria Be Established in Advance?
Prespecification can reduce the risk of interpreting feasibility findings according to whether they support a preferred decision.
For example, researchers who strongly favor proceeding with an intervention may be tempted to redefine acceptable adherence after observing disappointing completion rates.
Establishing the intended decision framework before examining the results makes such changes more visible.
However, prespecification does not eliminate the need for judgment.
Unexpected circumstances may reveal that an original threshold was poorly justified or that an important feasibility problem was overlooked.
Researchers should document consequential departures from the original criteria and explain their rationale rather than silently revising them.
Which Feasibility Outcomes Should Have Progression Criteria?
Progression criteria should prioritize outcomes that could materially affect whether the definitive study can answer its research question.
Not every recorded feasibility variable requires a numerical threshold.
| Feasibility domain |
Possible progression question |
Basis for justification |
| Recruitment |
Can the required sample be enrolled within the available period? |
Definitive sample requirement, recruitment capacity, and timeline |
| Retention |
Can sufficient participants remain available for follow-up? |
Follow-up design and consequences of attrition |
| Primary outcome completeness |
Can the study obtain enough usable outcome information? |
Planned analysis, precision, and missing-data risks |
| Intervention adherence |
Can participants complete the essential intervention activities? |
Core components and intended intervention exposure |
| Delivery fidelity |
Can implementers deliver essential components adequately? |
Intervention protocol and implementation requirements |
| Acceptability |
Are important burdens or concerns sufficiently addressed? |
Participant experiences, ethical considerations, and implementation context |
| Resources |
Can the study operate within sustainable resource constraints? |
Available funding, staffing, infrastructure, and time |
| Randomization |
Can allocation be implemented with appropriate safeguards? |
Protocol integrity and allocation concealment |
The appropriate domains depend on the study. A noninterventional qualitative investigation, for example, would not ordinarily require criteria concerning intervention fidelity or randomization.
How Do You Choose the Most Important Progression Criteria?
Begin with the requirements of the definitive research rather than the measurements most convenient to collect.
Ask which failures would prevent the study from producing meaningful evidence.
For example, a randomized trial may recruit sufficient participants but remain methodologically compromised if allocation concealment cannot be maintained.
Similarly, favorable intervention adherence cannot compensate for a primary outcome that cannot be measured appropriately.
Criteria should reflect the consequences of failure and the uncertainty surrounding each procedure.
The selection process is closely related to identifying the feasibility questions that must be answered before a definitive study.
Should Progression Criteria Always Be Numerical?
No. Numerical criteria are useful when outcomes have clear operational definitions and meaningful thresholds.
Recruitment speed, retention proportions, and assessment completeness may lend themselves to quantitative criteria.
However, other concerns require qualitative or procedural judgments.
For example, a feasibility study may reveal that participants consider a proposed data collection procedure intrusive or inconsistent with their expectations of confidentiality.
The seriousness of that concern may not be captured adequately by the percentage of participants expressing dissatisfaction.
Some progression conditions may therefore concern whether specific ethical, organizational, or methodological problems have been resolved.
How Should Recruitment Progression Criteria Be Developed?
Recruitment criteria should be linked to the number of participants required by the definitive study and the time available for enrollment.
For example, if the proposed study requires 240 participants over 12 weeks, the average required recruitment speed is 20 participants per week.
The feasibility study can examine whether the proposed recruitment process plausibly supports this requirement.
However, recruitment may vary across sites and time periods. A short pilot rate should not be extrapolated mechanically.
Additional sites, seasonal effects, and recruitment capacity should be considered.
See recruitment rate as a feasibility outcome.
How Should Retention Criteria Be Established?
Retention criteria should reflect the need for continued participation and the consequences of attrition.
For example, a longitudinal study may require participants to remain available through several follow-up assessments.
The research team should determine what retention level is compatible with its planned sample, analysis, and resources.
However, retention should not be confused with availability of usable primary outcome data.
A participant may remain enrolled without completing the required measurement.
Researchers may therefore need separate criteria for retention and outcome completeness.
There is no universal retention threshold suitable for every design.
How Should Primary Outcome Completeness Influence Progression?
Primary outcome completeness may be particularly consequential because the definitive study's main analysis depends on obtaining usable outcome information.
For example, a study may retain 95% of participants but obtain usable primary outcome records from only 80%.
The latter figure may be more directly relevant to the planned analysis.
Progression criteria should consider the number of observations required, potential loss of precision, and risks associated with missing outcomes.
However, the amount of missingness alone does not determine whether bias will occur.
Researchers should also investigate missingness patterns and possible causes.
See data completeness as a feasibility outcome.
How Should Intervention Adherence Criteria Be Defined?
Adherence criteria should correspond to the intervention activities that participants are expected to undertake.
For example, an intervention may require students to complete weekly feedback and revision activities.
The researchers should determine which activities are essential and what level of participation would permit meaningful evaluation.
However, an arbitrary threshold such as 80% completion is not universally defensible.
Some activities may be essential to the intervention's mechanism, while others may be supplementary.
Researchers should also determine whether incomplete adherence reflects participant behavior or failure to deliver the activities.
How Should Intervention Fidelity Influence Progression?
Delivery fidelity concerns whether the intervention is implemented according to its intended design.
Progression criteria may specify whether essential components must be delivered consistently or whether implementers must demonstrate particular competencies.
For example, an AI-supported feedback intervention may require teacher review before generated feedback reaches students.
If that review cannot be implemented reliably, the intervention may not be ready for definitive evaluation.
However, fidelity should not be reduced automatically to a single completion percentage.
Quality of delivery and the distinction between core components and permissible adaptations may be important.
Can Participant Acceptability Be a Progression Criterion?
Yes, when participant experiences could affect the practicality or ethical suitability of the proposed study.
For example, students may find an intervention educationally useful but regard its monitoring procedures as intrusive.
A progression criterion might require consequential concerns to be addressed before proceeding.
However, favorable average acceptability ratings do not necessarily establish that all important problems have been resolved.
Researchers should consider differences between participant groups and the nature of particular concerns.
For further discussion, see participant acceptability as a feasibility outcome.
Should Cost and Resource Use Have Progression Criteria?
They may, particularly when resource constraints could prevent the definitive study from being completed.
For example, the research team may establish the maximum sustainable staff workload or determine whether required technology and support can be funded.
However, resource thresholds should reflect actual constraints rather than arbitrary spending limits.
Some costly procedures may be necessary for participant safety, measurement quality, or methodological integrity.
Researchers should distinguish efficiency improvements from modifications that compromise essential requirements.
The relationship is examined further in cost and resource use as feasibility outcomes.
What Is a Go/No-Go Criterion?
A go/no-go criterion establishes a condition for deciding whether the proposed study should proceed under the specified design.
For example, a trial may be unable to proceed unless an essential institutional authorization is obtained.
Such a condition may be appropriately treated as non-negotiable.
However, many feasibility findings are not naturally binary.
Recruitment slightly below a target may be improved through additional sites, while substantial implementation difficulties may require redesign.
Researchers should avoid forcing every outcome into a simple pass-or-fail classification when modification or further testing is a legitimate option.
Should Progression Criteria Use a Green-Amber-Red Framework?
A traffic-light framework can help distinguish outcomes supporting progression from those requiring modification or substantial reconsideration.
Green generally indicates favorable feasibility evidence, amber indicates concerns that may be remediable, and red indicates substantial problems with the current design.
However, the categories should be defined for the particular study.
The framework is not a universal statistical classification, and colors do not establish the importance of an outcome.
The methodological implications are examined in whether green-amber-red progression criteria should replace pass-fail decisions.
How Should Statistical Uncertainty Affect Progression Decisions?
Feasibility estimates are often based on relatively small samples and may therefore be imprecise.
For example, observing that 18 of 20 participants complete follow-up gives a retention proportion of 90%.
However, that observation does not establish that the underlying retention probability is precisely 90%.
Confidence intervals may help communicate uncertainty when appropriate.
Researchers should avoid treating a point estimate just above or below a threshold as definitive evidence of feasibility or infeasibility.
The required precision depends on the consequences of the progression decision.
Should Progression Be Based on Statistical Significance?
Generally, progression criteria should not depend on whether a pilot intervention effect achieves conventional statistical significance.
Feasibility studies are commonly designed to investigate whether a definitive study can be conducted, not to provide conclusive evidence about intervention effectiveness.
A nonsignificant preliminary effect does not establish ineffectiveness, while a significant result in a small pilot may be unstable.
Where formal statistical decision rules are used for feasibility outcomes, they should be justified by the feasibility question and relevant uncertainty.
Researchers should distinguish feasibility testing from hypothesis testing concerning the definitive intervention effect.
How Should Multiple Progression Criteria Be Combined?
When several criteria are used, researchers should establish how their findings will contribute to the overall decision.
For example, a study may meet recruitment and retention criteria but fail to demonstrate adequate primary outcome collection.
Simply averaging the indicators would conceal the importance of the measurement problem.
Some criteria may represent essential conditions that cannot be compensated for by favorable performance elsewhere.
Others may identify remediable operational weaknesses.
Researchers should distinguish critical conditions from supporting indicators and avoid treating all percentages as interchangeable.
What If Some Criteria Are Met and Others Are Not?
Mixed findings require interpretation of the underlying problems and their consequences.
For example, recruitment may be adequate while teacher workload exceeds expectations.
The team should determine whether additional resources or revised procedures could address the workload without changing the intervention's essential components.
If the problem is remediable, proceeding with modifications may be defensible.
If the problem undermines an essential requirement and cannot be resolved, the definitive study may need substantial redesign or discontinuation.
Progression should therefore be based on an integrated assessment of the evidence rather than a mechanical count of criteria passed.
Can Qualitative Findings Influence Progression?
Yes. Qualitative evidence may explain why a numerical feasibility target was or was not achieved.
For example, low intervention adherence may reflect unclear instructions, technical barriers, or excessive workload.
These explanations have different implications for progression.
Qualitative findings may also reveal ethical or contextual problems that were not captured by the original quantitative indicators.
However, researchers should document how qualitative evidence informs decisions and avoid selecting isolated quotations merely to justify a preferred conclusion.
Should Progression Criteria Be Different Across Sites?
Site-specific conditions may warrant attention when the definitive study involves multiple institutions.
For example, recruitment capacity may differ between large and small universities.
Researchers may establish common essential standards while examining contextual variation in operational indicators.
However, applying different thresholds retrospectively to classify poorly performing sites as successful can undermine transparency.
Any planned variation should be justified according to the study's design and implementation requirements.
Overall feasibility should not conceal serious problems concentrated at particular sites.
Who Should Establish the Progression Criteria?
Progression criteria should be developed by individuals with relevant methodological, operational, and substantive expertise.
Depending on the study, this may include investigators, statisticians, intervention specialists, participating institutions, and relevant participant representatives.
For example, researchers may understand statistical requirements while teachers provide important information about realistic implementation workload.
Stakeholder input can improve the relevance of criteria, but it does not replace methodological justification.
Researchers should also consider whether independent oversight is appropriate for consequential progression decisions.
Can Progression Criteria Be Revised After the Feasibility Study Begins?
Sometimes, particularly when unexpected circumstances reveal that the original criteria were inappropriate or incomplete.
However, revisions should be distinguished from retrospective attempts to make unfavorable findings appear acceptable.
Researchers should document the original criteria, reasons for change, timing of the revision, and implications for interpretation.
Where applicable, protocol amendments and relevant approvals may be necessary.
The final report should distinguish prespecified decisions from judgments made after examining the evidence.
What Is the Difference Between External and Internal Pilot Progression?
An external pilot is conducted separately from the definitive study, and its findings inform the design and decision to proceed.
An internal pilot forms an initial phase of the definitive study, with prespecified feasibility assessment and progression arrangements.
In an internal pilot, substantial design modifications may affect whether the initial data remain suitable for inclusion in the final analysis.
Consequently, progression planning should consider data compatibility, trial integrity, and the permitted scope of modifications.
The distinction should be established prospectively rather than decided after the preliminary results are known.
What If Progression Requires Substantial Redesign?
The research team should determine whether the revised study remains sufficiently similar to the design that generated the feasibility evidence.
For example, replacing the intervention platform, changing the primary outcome, or substantially altering recruitment criteria may introduce new uncertainties.
Previous feasibility findings may no longer address the most important requirements of the revised design.
Additional testing may therefore be appropriate.
However, not every modification requires repeating the entire feasibility study.
Researchers should identify which changes create consequential new uncertainty and test those components proportionately.
Can Progression Criteria Justify Not Proceeding?
Yes. A feasibility study may demonstrate that the definitive research cannot be conducted responsibly under the proposed conditions.
For example, recruitment may remain insufficient despite realistic modifications, or essential outcome data may be unobtainable.
Such findings do not make the feasibility study unsuccessful.
They provide evidence that can prevent unnecessary expenditure, participant burden, and methodologically compromised research.
The value of feasibility research lies partly in its ability to support a defensible decision not to proceed.
How Should Progression Decisions Be Reported?
The report should present the prespecified criteria, their rationale, observed outcomes, and relevant uncertainty.
Researchers should explain which criteria were met, which were not, and how contextual evidence influenced the final decision.
Where modifications are recommended, the report should identify the changes and whether additional feasibility testing is required.
Departures from prespecified criteria should be disclosed and justified.
The reporting should distinguish evidence about study feasibility from claims about intervention effectiveness.
Watch Out
Do not establish progression thresholds retrospectively simply because the observed results fall below the original expectations. Nor should favorable recruitment or adherence compensate automatically for unresolved ethical concerns, invalid outcome assessment, or failures in essential research procedures. Progression criteria should support defensible decisions rather than manufacture a favorable verdict.
