Should Feasibility Study Progression Criteria Use a Green-Amber-Red System Instead of Pass-Fail?

Should Feasibility Decisions Be Classified as Green, Amber, or Red Rather Than Simply Pass or Fail?

You have established progression criteria for a feasibility study and are preparing to evaluate whether the proposed definitive research should proceed.

Some findings may be straightforward. Recruitment meets the required rate, the intervention can be delivered, and the primary outcome is collected successfully.

Other findings may be less decisive.

Suppose recruitment reaches 18 participants per week when the definitive study requires 20. Most intervention activities are completed, but staff report that additional training is necessary. Outcome completeness is slightly below the intended threshold, although the causes appear remediable.

Should these results be classified as failures?

A binary pass-fail framework may be too restrictive when preliminary findings identify problems that can be corrected. A green-amber-red system offers an alternative by distinguishing favorable results from those requiring modification and those suggesting substantial infeasibility.

But is the traffic-light approach methodologically preferable, and how can researchers prevent its categories from becoming arbitrary labels?

Use Green-Amber-Red Criteria When Feasibility Decisions Require More Than Two Responses

Using Traffic-Light Progression Criteria Responsibly

What Is a Green-Amber-Red Progression System?

A green-amber-red progression system classifies feasibility findings into categories associated with different potential actions.

It is sometimes called a traffic-light system because its categories resemble familiar traffic signals.

In feasibility research, the approach is commonly used to structure decisions about whether a definitive study should proceed.

These interpretations are common conventions rather than universally standardized statistical definitions.

Researchers should state what each category means for their particular study.

How Is a Traffic-Light System Different From Pass-Fail?

A binary framework distinguishes outcomes that meet a criterion from those that do not.

A traffic-light framework introduces an intermediate category to recognize that some feasibility problems may be remediable.

Neither approach is automatically more rigorous.

A well-justified binary rule may be preferable to an arbitrary traffic-light classification.

Why Is an Amber Category Useful?

The amber category recognizes that preliminary feasibility research often identifies problems that can be addressed before definitive evaluation.

For example, a recruitment process may fall below the required rate but reveal that additional suitable institutions are available.

The study may therefore be feasible after modifying its recruitment arrangements.

Without an intermediate category, the result might be classified simply as failure.

However, amber should not become a convenient label for every unfavorable finding.

Researchers should identify the proposed corrective action, its plausibility, and whether further evidence is required to demonstrate that it works.

Does Green Mean the Definitive Study Is Guaranteed to Succeed?

No. Green indicates that the observed feasibility evidence meets the specified criterion under the conditions examined.

For example, a pilot may achieve the recruitment speed required by the definitive timeline.

However, enrollment may change as the accessible population becomes depleted or additional sites introduce different conditions.

Similarly, high intervention fidelity in a small pilot does not guarantee equally consistent delivery at scale.

Green should therefore indicate support for progression, not certainty about future performance.

Does Red Automatically Mean the Research Should Be Abandoned?

Not necessarily. Red may indicate that the current design cannot proceed without substantial changes.

For example, a trial may be unable to recruit enough eligible participants under its original inclusion criteria and recruitment arrangements.

Researchers may need to reconsider the population, design, number of sites, or research question.

However, the resulting redesign may introduce new feasibility questions.

A red classification can therefore support a decision to stop the current design while leaving open the possibility of a different investigation.

The category should describe the implications of the evidence, not pronounce the broader research topic unworthy of study.

How Should Green-Amber-Red Thresholds Be Established?

Thresholds should follow the requirements of the proposed definitive study.

For example, recruitment categories should be linked to the sample size, recruitment period, and realistic enrollment capacity.

If a study requires 20 participants per week, the researchers may define green as performance compatible with that requirement.

Amber could represent a lower range for which a credible recruitment improvement strategy exists.

Red could represent recruitment performance unlikely to meet the definitive requirement even after feasible modifications.

However, the exact boundaries require justification.

Researchers should not adopt identical percentages across unrelated feasibility outcomes merely for convenience.

Should Thresholds Be Based on Universal Percentages?

No. A threshold that is appropriate for one study may be unsuitable for another.

For example, 80% intervention adherence may be sufficient when missed activities are supplementary but inadequate when essential intervention components are omitted.

Similarly, 85% primary outcome completeness may have different consequences depending on sample size, missingness patterns, and the planned analysis.

Researchers should derive thresholds from the requirements and risks of the definitive study.

Where empirical evidence is limited, the assumptions underlying the chosen boundaries should be made explicit.

Should Green, Amber, and Red Be Defined Before Data Collection?

Generally, yes, when the categories will guide progression decisions.

Prespecification helps reduce the risk of assigning favorable categories retrospectively.

For example, researchers should not redefine an observed recruitment rate as green simply because it falls below the originally intended threshold.

However, unexpected circumstances may justify revisiting a criterion.

Such changes should be documented, explained, and distinguished from the original decision framework.

Where relevant, protocol amendments or oversight approval may be necessary.

Can the Categories Be Defined Using Confidence Intervals?

Statistical uncertainty can be incorporated into progression decisions, including through confidence intervals or other justified approaches.

For example, a pilot may observe an 85% follow-up completion rate based on a small number of participants.

The estimate may be too imprecise to determine confidently whether the definitive study can achieve the required completeness.

Researchers may therefore consider the range of plausible underlying values rather than classify the point estimate alone.

However, there is no universal confidence interval rule that automatically determines green, amber, or red.

The approach should reflect the feasibility question and consequences of the decision.

What Happens When an Estimate Falls Near a Category Boundary?

A point estimate close to a threshold should be interpreted cautiously.

For example, a recruitment rate of 19.8 participants per week may be practically similar to a requirement of 20, particularly when the estimate is uncertain.

Classifying one value as amber and the other as green without considering uncertainty may create a misleading impression of precision.

Researchers should examine whether the observed difference is consequential and whether the proposed recruitment process can realistically achieve the definitive requirement.

The categories should guide judgment rather than substitute for it.

Should Traffic-Light Criteria Be Applied to Recruitment?

Recruitment is a common candidate because the definitive study may have a clear sample requirement and enrollment timeline.

For example, green could indicate that the required enrollment speed is plausibly achievable.

Amber could indicate that additional recruitment channels or sites are necessary.

Red could indicate that the current recruitment strategy is insufficient and no credible modification has been identified.

However, recruitment speed should be distinguished from eligibility yield and enrollment among eligible participants.

The appropriate indicator depends on the recruitment uncertainty.

See recruitment rate as a feasibility outcome.

Should Traffic-Light Criteria Be Applied to Retention?

They may be useful when continued participation is essential to the definitive study.

For example, green could indicate follow-up retention compatible with the intended design.

Amber could indicate that additional reminders or scheduling modifications are needed.

Red could indicate substantial attrition that threatens the proposed analysis.

However, retention and primary outcome completeness should not be treated as identical.

Participants may remain enrolled while failing to provide usable outcome measurements.

Researchers should select the indicator that directly addresses the definitive study's requirements.

How Should Intervention Adherence Be Classified?

Adherence categories should reflect completion of the intervention's essential activities.

For example, an educational intervention may require students to review feedback and submit revised work.

Green might indicate that the required engagement is achievable under the tested conditions.

Amber might indicate that scheduling or instructions need modification.

Red might indicate that the intervention demands are incompatible with ordinary participation.

However, researchers should distinguish participant adherence from intervention delivery fidelity.

Low adherence may result from activities not being delivered or made accessible.

Can Acceptability Be Classified Using Traffic Lights?

Yes, but care is necessary because acceptability is multidimensional.

For example, participants may rate an intervention favorably while raising serious concerns about privacy or workload.

An aggregate mean score may conceal those concerns.

Researchers should consider whether particular findings require modification regardless of the overall rating.

Qualitative evidence may be especially useful for explaining why an intervention is acceptable or problematic.

Color categories should not reduce complex participant experiences to a simplistic popularity measure.

Can Resource Use Be Classified Using Traffic Lights?

Resource criteria may be appropriate when the definitive study has explicit staffing, funding, or infrastructure constraints.

For example, green could indicate that resource requirements are compatible with available capacity.

Amber could indicate that additional support or operational changes are needed.

Red could indicate that the proposed design cannot be funded or implemented under current conditions.

However, researchers should distinguish essential resource requirements from avoidable inefficiencies.

Eliminating necessary outcome assessments or participant safeguards merely to achieve a green budget classification would undermine the purpose of feasibility research.

Should Every Feasibility Outcome Receive a Traffic-Light Classification?

No. Some outcomes may be recorded primarily to explain implementation rather than determine progression.

For example, the number of technical support requests may help identify common problems without requiring a separate green-amber-red threshold.

Other outcomes may have clear binary requirements, such as whether a mandatory authorization has been obtained.

Researchers should avoid assigning colors to every variable merely to create a visually symmetrical reporting table.

The classification system should reflect the importance and decision relevance of each outcome.

How Should Multiple Traffic-Light Results Be Combined?

Researchers should establish how individual criteria contribute to the overall progression decision.

For example, a study may receive green classifications for recruitment and retention but red for the integrity of primary outcome assessment.

It would be misleading to conclude that the study is generally green because most indicators are favorable.

Some criteria concern essential methodological conditions that cannot be offset by success elsewhere.

Researchers should distinguish critical requirements from remediable operational issues.

A decision matrix may be useful, but the underlying logic should be explicit.

Should the Overall Decision Be Determined by the Worst Color?

Not automatically.

A single red finding may represent an essential condition that prevents progression, such as unresolved allocation concealment problems.

However, another red finding may reflect a resource constraint that could be addressed through substantial but feasible redesign.

The decision depends on what the criterion represents and whether the problem can be resolved.

Researchers should avoid applying an automatic worst-color rule without considering the consequences and decision framework.

Where particular conditions are non-negotiable, that status should be established in advance.

Can Green-Amber-Red Systems Conceal Important Differences?

Yes. Two studies may both classify recruitment as amber while facing very different problems.

One may need a modest extension of the recruitment period. Another may require a substantial change in the target population.

The same color does not imply equivalent feasibility or comparable corrective actions.

Researchers should therefore report the underlying estimates, observations, and contextual explanations.

Colors are useful summaries only when accompanied by evidence.

Can Traffic-Light Systems Encourage Arbitrary Decisions?

They can, particularly when thresholds are selected without a defensible rationale.

For example, classifying 79% adherence as amber and 80% as green may imply a meaningful distinction that the intervention theory does not support.

Similarly, thresholds selected after observing the data may make unfavorable results appear acceptable.

Researchers should justify the categories and consider uncertainty near boundaries.

The framework should not be treated as a substitute for methodological expertise.

What Is the Role of Qualitative Evidence in Traffic-Light Decisions?

Qualitative evidence can explain the conditions underlying quantitative feasibility findings.

For example, an amber recruitment rate may be associated with administrative delays that can be resolved before the definitive study.

Alternatively, interviews may reveal that potential participants consider the intervention fundamentally inappropriate.

These findings suggest different progression decisions despite similar numerical recruitment results.

Researchers should explain how qualitative evidence informs classification and corrective action.

It should not be used selectively to justify an outcome the research team already prefers.

Should Traffic-Light Decisions Be Made by the Research Team Alone?

Not necessarily. The appropriate decision-making arrangement depends on the study's governance and consequences.

For example, a large randomized trial may involve independent oversight, funders, statisticians, and relevant stakeholder representatives.

Researchers should identify who has responsibility for interpreting feasibility evidence and authorizing progression.

However, stakeholder consultation should not replace methodological or ethical requirements.

The final decision should be documented with its evidential basis and any relevant conditions.

How Does the Approach Apply to Internal Pilot Studies?

Internal pilots form an initial phase of a definitive study, and progression decisions may determine whether the trial continues as planned.

Traffic-light criteria can help distinguish continuation without major changes from modifications or substantial reconsideration.

However, changes to eligibility, intervention procedures, outcome measures, or allocation arrangements may affect the compatibility of internal pilot data with the definitive analysis.

Researchers should establish permitted modifications and data inclusion arrangements prospectively.

The framework should be consistent with the trial protocol and relevant oversight requirements.

How Does the Approach Apply to External Feasibility Studies?

External feasibility studies are conducted separately from the definitive study.

Their findings can support broader design modifications because the preliminary data are not necessarily intended for inclusion in the definitive analysis.

For example, researchers may substantially revise an intervention or recruitment strategy after identifying feasibility problems.

However, extensive changes may create new uncertainties that require additional testing.

A traffic-light decision should therefore identify not only whether progression is supported but what modifications and verification are necessary.

Should Traffic-Light Categories Be Revised After the Study?

Consequential revisions may sometimes be justified when unexpected evidence reveals limitations in the original framework.

However, researchers should distinguish such revisions from attempts to reclassify disappointing findings favorably.

The original thresholds, revised criteria, reasons for change, and timing should be documented.

Where applicable, changes should undergo appropriate protocol or oversight review.

Transparent reporting is more important than preserving the appearance of a uniformly successful feasibility study.

What If Every Indicator Is Amber?

An entirely amber assessment does not automatically justify progression.

It may indicate that several manageable modifications are required, but it could also reveal a pattern of cumulative operational weaknesses.

For example, moderate recruitment delays, additional teacher workload, and incomplete outcome data may each appear remediable separately.

Together, however, they may exceed the research team's available capacity.

Researchers should examine the combined consequences of the findings and whether proposed modifications are realistic.

Amber should mean that a credible pathway to resolution exists, not merely that the researchers remain optimistic.

When Is a Binary Pass-Fail Approach More Appropriate?

Binary criteria may be suitable when a requirement is genuinely non-negotiable.

For example, a study cannot proceed with an activity that lacks mandatory ethical authorization.

Similarly, an essential data access agreement may need to be secured before the research can begin.

These conditions may not benefit from an intermediate category if partial compliance does not permit lawful or ethical implementation.

However, a binary criterion can coexist with traffic-light classifications for other feasibility outcomes.

Researchers do not need to use one decision format for every requirement.

How Should Green-Amber-Red Results Be Reported?

The report should present each feasibility outcome, the prespecified thresholds, observed findings, uncertainty, assigned category, and resulting action.

Researchers should explain the rationale for boundaries and identify any deviations from the original framework.

For amber outcomes, the required modifications and verification procedures should be stated.

For red outcomes, the report should explain whether the problem prevents progression under the current design or requires substantial redesign.

The colors should supplement, not replace, the underlying evidence and methodological reasoning.

Using Green-Amber-Red Criteria in an Educational Intervention Feasibility Study

Illustrative Traffic-Light Decision Matrix

The hypothetical categories have been constructed to demonstrate how a traffic-light framework operates. In actual research, each boundary would require a rationale grounded in the definitive design.

The findings do not justify declaring the study ready without further consideration.

Recruitment and outcome completeness require attention, and the team should determine whether the proposed modifications are likely to address the identified problems.

Moreover, the absence of a red category does not guarantee that the definitive study is feasible. Important issues outside the three indicators, such as ethics, data integrity, or institutional resources, may remain unresolved.

A defensible conclusion would specify the conditions under which progression is supported and whether further feasibility evidence is necessary.

Common Misconceptions About Traffic-Light Progression Criteria

How to Decide Whether to Use Traffic-Light Progression Criteria

Begin by determining whether the feasibility study's findings could reasonably support more than two responses.

If intermediate findings may justify specific modifications, a green-amber-red framework may improve the clarity of progression planning.

The best framework is the one that supports transparent and proportionate decisions about the definitive study.

Traffic-light classifications are useful when they clarify the available actions, not when they merely decorate a feasibility report.

Before Using Green-Amber-Red Progression Criteria, Check:

Frequently Asked Questions About Green-Amber-Red Progression Criteria

Use Traffic-Light Criteria to Clarify Progression, Not Replace Judgment

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.