01 · The Question
If We Already Know What We Should Do, What Would Another Study Add?
Some research problems remain serious even after evidence points clearly toward a course of action. Perhaps an intervention should be implemented, an unsafe practice should be stopped, an accessibility barrier should be corrected, or a policy should be changed. There will always be some uncertainty, but the available evidence may already be strong enough for a responsible decision.
Should researchers continue studying the problem anyway?
Sometimes. Additional evidence may refine implementation, identify differential effects, reveal important risks, or improve confidence in a high-stakes decision. But if the remaining uncertainty is unlikely to change what should be done, another study may have little decision value. In that situation, research can become a way of postponing action rather than improving it.
03 · What You Need to Know
Research Should Reduce Uncertainty That Actually Matters
Decision-Relevant Uncertainty Is More Important Than Uncertainty in General
Almost every empirical conclusion contains uncertainty. Estimates have confidence intervals, findings may vary across contexts, future conditions can change, and no finite evidence base eliminates every conceivable doubt.
If the standard for action were complete certainty, many decisions could be postponed indefinitely.
The more useful question is whether the uncertainty that remains is decision-relevant. Could obtaining better evidence reasonably lead a decision-maker to choose differently, modify the action substantially, target a different population, alter the timing, or manage an important risk differently?
Residual uncertainty
Everything that remains unknown or imperfectly estimated after considering existing evidence.
Decision-relevant uncertainty
Uncertainty substantial enough that reducing it could plausibly change an important choice or how that choice should be implemented.
A Serious Problem Does Not Automatically Create a Need for Another Study
The importance of a problem establishes a reason to care about its consequences. It does not establish that producing more evidence is the best next response.
This distinction is particularly important because research has opportunity costs. WHO's ethics guidance on research priority setting emphasizes that research resources, including time, infrastructure, personnel, and funding, are scarce. Investing them in one project means they cannot be used for another.
If additional evidence is unlikely to alter the relevant decision, those resources may have greater value elsewhere.
This is another way of distinguishing social importance from the scientific value of another study.
The Decision Can Be Clear Even When the Evidence Is Not Perfect
Decision-making under uncertainty is normal. A policymaker, institution, practitioner, or individual may need to choose before every uncertainty has been resolved.
What counts as sufficient evidence depends partly on the decision. A low-cost and easily reversible action with substantial expected benefit may reasonably be adopted with more uncertainty than an expensive, irreversible, or high-risk intervention.
The consequences of being wrong matter, as do the consequences of delaying a beneficial action.
Ask Whether More Information Could Change the Choice
One useful way to think about additional research is through the value of information. In decision analysis, information has value when reducing uncertainty can improve the decision and its expected outcomes. PCORI's methodology work, for example, discusses value-of-information analysis as a way of balancing research costs against the potential usefulness of additional certainty.
You do not need a formal economic model to apply the basic reasoning to an ordinary research problem.
| Question |
Why It Matters |
| What decision must be made? |
Research should be connected to an identifiable uncertainty or knowledge need rather than "more information" in the abstract. |
| What does current evidence support? |
This establishes the starting point before asking what additional evidence might contribute. |
| What uncertainty remains? |
Not all uncertainty is equally consequential. |
| Could plausible new evidence change the decision? |
If not, the decision value of another study may be low. |
| Could evidence change how the decision is implemented? |
The main value of research may concern targeting, adaptation, safety, equity, or delivery rather than the basic choice. |
| What happens while we wait? |
Delaying action can itself produce costs or harms. |
| What does the research cost? |
Participants, time, money, personnel, and alternative research opportunities all matter. |
More Research Can Be Valuable Even When the Basic Decision Will Not Change
A decision can be clear at one level while important uncertainties remain at another.
Suppose an institution already knows that it should implement a particular accessibility standard. Research may still help determine the most effective implementation strategy, identify where compliance fails, compare lower-cost approaches, or examine whether some groups remain underserved after implementation.
The decision "implement the standard" remains unchanged. Research improves how that decision is carried out.
High-Stakes and Irreversible Decisions May Justify Greater Certainty
The amount of uncertainty worth reducing depends partly on what is at stake. If a decision is difficult to reverse, exposes people to substantial risk, requires enormous resources, or could create serious unintended consequences, additional evidence may be valuable even when one option currently appears preferable.
In clinical research, this concern is especially explicit. Ethical guidance recognizes the importance of genuine uncertainty when comparing interventions, particularly because participants should not knowingly be deprived of an established effective intervention without adequate justification.
The broader lesson is that the threshold for "enough evidence" should be sensitive to the consequences of error rather than treated as universal.
Reversible Decisions Can Sometimes Be Made While Learning Continues
Not every choice requires research to finish before action begins. When an intervention is reversible, monitorable, and reasonably safe, decision-makers may sometimes act on current evidence while collecting additional data.
This creates a different relationship between research and action. Instead of "research first, decide later," the process may involve implementation accompanied by evaluation and adjustment.
That approach is not appropriate in every context, particularly where experimentation or data collection raises ethical or safety concerns. But it highlights a useful principle: the choice between research and action is not always binary.
The Costs of Waiting Belong in the Decision
Researchers often focus on the risk of acting with insufficient evidence while paying less attention to the risk of waiting.
If a harmful condition continues while another marginal study is conducted, delay has consequences. WHO Europe has explicitly cautioned in environment and health priority-setting work that uncertainty should not unnecessarily hamper or delay policy action.
This does not justify careless decisions. It means that the expected benefit of additional certainty should be compared with the consequences of postponing a decision that current evidence already supports.
Watch Out
"More research is needed" should not function as a default conclusion whenever uncertainty remains. Specify what remains uncertain, what evidence would resolve it, and what decision could change because of that evidence.
A Clear Decision Does Not Mean Every Stakeholder Will Agree
Another complication is disagreement. People may reject a decision despite agreeing about the evidence because they assign different values to costs, risks, fairness, autonomy, or competing priorities.
Additional empirical research may not resolve a disagreement that is fundamentally normative. If two stakeholders interpret the same evidence differently because they value outcomes differently, producing a more precise estimate may leave the disagreement intact.
This is why research priority setting involves value judgments as well as technical evidence. WHO's ethics guidance explicitly treats questions about whose interests count and how potential research benefits should be distributed as ethical components of priority setting.
The Research Question Should Follow the Remaining Decision Need
Once a broad decision is sufficiently clear, the research agenda should move rather than simply repeat itself.
Initial uncertainty Should we adopt Intervention X?
Evidence accumulates Existing evidence supports adoption strongly enough for the decision.
Decision Adopt Intervention X.
Possible remaining uncertainty Which implementation strategy works best, which populations need adaptation, what unintended effects require monitoring, and how sustainable is the intervention?
New research focus Study the uncertainties that could improve implementation rather than repeatedly asking whether Intervention X should be adopted at all.
This is closely related to what happens to the research agenda once the existence of a problem is already established. Knowledge should move the question forward.
06 · What This Means for You
Ask Whether Your Study Could Actually Change Something That Is Still Undecided
If you are proposing research around a problem for which the broad decision appears clear, identify precisely what additional evidence would contribute.
A useful test is to state the decision, describe the remaining uncertainty, and then imagine plausible results from your study. Would any credible result lead to a different decision or substantially alter how that decision should be implemented?
A simple decision framework
If plausible study results could reverse the preferred decision
Additional research may have substantial decision value.
If the basic decision will remain the same but implementation choices are uncertain
Redirect research toward the uncertainties that could improve implementation.
If remaining uncertainty concerns serious risks or irreversible consequences
Greater certainty may be justified before proceeding.
If the decision is low-risk, reversible, and monitorable
Consider whether action with ongoing evaluation is preferable to delaying action for marginal additional evidence.
If no plausible result would change the decision or its implementation
Question whether the proposed study has enough decision value to warrant priority.
If the likely solution is already known, this test can help distinguish useful research about remaining uncertainties from repeated investigation of a decision that evidence already supports.
07 · A Quick Checklist
Before Conducting More Research on an Already Clear Decision
Before proposing another study, check:
State the actual decision that the evidence is intended to inform.
Review the quality, consistency, relevance, and limitations of the evidence already available.
Identify the specific uncertainty that remains rather than saying only that "more research is needed."
Ask whether any plausible result from the proposed study could change the decision materially.
Consider whether additional evidence could improve implementation, targeting, safety, sustainability, or equity even if the basic decision will not change.
Assess the consequences of making the wrong decision and whether the action is reversible.
Assess the consequences of delaying action while further research is conducted.
Compare the expected value of additional information with the costs, participant burden, time, and alternative uses of research resources.
