When Should Research Move From Efficacy to Implementation?

When should researchers stop asking only whether an intervention can work and study how it works in practice?

An intervention can perform well in a carefully controlled study and still struggle outside research conditions. Organizations may not adopt it. Practitioners may find it difficult to deliver. Participants may not engage with it. Costs, staffing, workflow, training, or local conditions may make sustained use difficult.

These are not merely logistical inconveniences after the “real” research is finished. They are research questions in their own right.

The challenge is timing. Moving toward implementation too early can devote substantial effort to disseminating an intervention whose benefits remain uncertain. Waiting for perfect evidence, however, can unnecessarily delay learning about whether the intervention can function in routine practice.

Move toward implementation when real-world delivery becomes a consequential uncertainty

Efficacy and implementation ask different questions about the same intervention

Efficacy asks whether an intervention can produce the intended effect

Efficacy research generally examines whether an intervention produces beneficial outcomes under comparatively controlled or favorable conditions. Researchers may closely manage eligibility, delivery, adherence, training, monitoring, and other aspects of the study to obtain a strong test of the intervention itself.

This is scientifically valuable because poor implementation is less likely to obscure whether the intervention can produce the intended effect under the studied conditions.

But successful efficacy does not guarantee successful routine use.

Effectiveness moves closer to ordinary conditions

Effectiveness research generally asks whether an intervention produces useful outcomes under conditions closer to routine practice, often with broader populations and less experimental control than a conventional efficacy study.

Efficacy and effectiveness are better understood as a continuum than as two perfectly separated categories. Studies can contain both explanatory and pragmatic features depending on their aims and design.

The important point is that evidence an intervention can work under favorable conditions does not automatically establish how well it will perform across the people, organizations, resources, and constraints encountered in practice.

Implementation research asks how an intervention becomes routine practice

Implementation research shifts attention toward the processes and strategies involved in adopting, delivering, integrating, scaling, and sustaining an evidence-based intervention or practice in real-world systems.

Proctor and colleagues distinguished implementation outcomes from service and clinical outcomes. Their widely used taxonomy includes acceptability, adoption, appropriateness, feasibility, fidelity, implementation cost, penetration, and sustainability.

The transition becomes important when intervention failure and implementation failure need to be separated

Suppose an intervention produces good outcomes in controlled research but disappointing outcomes after being introduced into routine settings. What went wrong?

One possibility is intervention failure: the intervention itself did not produce the expected effect in the new context. Another is implementation failure: an effective intervention was not adopted, delivered, accessed, or sustained sufficiently well to produce the expected outcomes.

Proctor and colleagues emphasize that distinguishing implementation effectiveness from treatment effectiveness is important because poor implementation can prevent an otherwise effective intervention from producing its intended outcomes.

You do not necessarily have to finish effectiveness research before studying implementation

The traditional research pipeline is often represented as a sequence from efficacy to effectiveness and then implementation. That sequence can be useful, but it is not mandatory in every case.

Curran and colleagues proposed effectiveness-implementation hybrid designs specifically to combine questions that conventional staged research often separates. Later methodological work has continued to develop this approach.

A Type 1 hybrid primarily tests intervention effectiveness while gathering information about implementation. A Type 2 hybrid gives substantial attention to both intervention effectiveness and implementation. A Type 3 hybrid primarily tests an implementation strategy while continuing to collect information about intervention outcomes.

These designs make an important point: moving toward implementation need not mean abandoning questions about whether the intervention itself works. The balance of questions can change gradually as the evidence develops.

Implementation research becomes premature when intervention evidence is too uncertain

There is little value in optimizing large-scale uptake of an intervention when researchers do not yet have a defensible reason to believe that the intervention provides the intended benefit.

This does not mean researchers must demand absolute certainty. Science rarely supplies it. The relevant judgment concerns whether existing evidence is strong enough, the intervention sufficiently plausible and appropriate, and the risks sufficiently understood to justify investing in questions about routine use.

The threshold should also reflect consequences. Interventions that are costly, invasive, difficult to reverse, or potentially harmful may warrant stronger effectiveness evidence before widespread implementation than low-risk practices.

Waiting too long has costs too

The opposite problem is assuming that implementation questions should begin only after a perfectly linear research pipeline has been completed. Curran and colleagues argued that blending effectiveness and implementation research can accelerate learning and produce information that is more useful to decision makers.

Implementation barriers discovered late can be particularly consequential. An intervention might require resources that ordinary settings cannot supply, depend on specialist expertise that is scarce, conflict with existing workflows, or be unacceptable to the people expected to use it.

Studying such issues while effectiveness evidence develops can reveal whether the intervention is realistically implementable and what support might eventually be required.

Implementation is not simply “using the intervention in the real world”

Putting an intervention into practice and observing what happens is not automatically implementation research. Implementation science investigates questions about the process of uptake and the strategies used to improve it.

For example, researchers might compare different training strategies, examine determinants of adoption, measure fidelity, investigate organizational barriers, evaluate implementation costs, or study sustainment after initial support ends.

The implementation strategy itself can therefore become the object of scientific evaluation.

The move toward implementation often follows a broader maturation of the question

When evidence about basic effectiveness has accumulated, researchers may already have begun asking for whom, when, and why an intervention works. Implementation adds another layer: can those benefits be realized reliably when the intervention enters ordinary systems?

This can be particularly important when new studies rarely change the central effectiveness conclusion. Another efficacy study under familiar conditions may then provide less information than research examining the barriers between evidence and routine use.

When a successful educational intervention meets ordinary practice

These questions are related but not interchangeable. A highly effective program that almost nobody adopts will have limited real-world impact. A widely adopted program that does not improve outcomes has a different problem entirely.

Implementation is not simply the final box in a research pipeline

Identify whether the uncertainty now lies in the intervention or in putting it into practice

Before proposing implementation research, examine the evidence supporting the intervention itself. Then identify what prevents the existing evidence from translating confidently into routine use.

Research progression should therefore follow the uncertainty rather than a ceremonial sequence of study types. If the original efficacy question has been answered reasonably well, ask which new question has become consequential. Sometimes that question is no longer whether the intervention can work, but how to make its benefits attainable under the conditions where people actually need it.

Check whether implementation research is becoming the priority

Questions about moving from efficacy toward implementation

Move toward implementation when putting the evidence into practice becomes part of the scientific problem

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.