03 · What You Need to Know
The Difference Between Eliminating Risk and Minimizing Risk
Zero Risk Is Usually Not the Ethical Standard
The Belmont Report recognizes the central tension directly: learning what is harmful or beneficial may itself require exposing people to some possibility of harm. It therefore asks when pursuing potential benefits or valuable knowledge justifies research risk rather than demanding that all research be completely risk-free.
The 2024 Declaration of Helsinki similarly acknowledges that most interventions in medical practice and medical research involve risks and burdens. Its requirement is careful assessment, minimization, management, and justification, not the impossible proposition that nothing adverse can ever occur.
Some Risks Exist Only Because of Poor Design
The first category is conceptually easy: if a risk contributes nothing necessary to answering the research question or protecting participants, there is little reason to expose participants to it.
Suppose a researcher collects participants' full names, exact addresses, and employer details even though none is needed for analysis, follow-up, or data linkage. Those identifiers create confidentiality exposure without corresponding scientific necessity. Removing them can reduce risk without weakening the study.
Likewise, an additional invasive procedure performed only because the resulting data “might be useful someday” deserves scrutiny if it contributes little to the research objective.
Risk Minimization Is Part of Sound Research Design
Under the U.S. Common Rule, IRBs must determine that risks to participants are minimized by using procedures consistent with sound research design that do not unnecessarily expose participants to risk and, whenever appropriate, by using procedures already being performed for diagnostic or treatment purposes.
The Belmont Report similarly states that risks should be reduced to those necessary to achieve the research objective and that researchers should consider alternative procedures.
This means risk minimization is not something added after the methods section has been finalized. It is part of designing the methods properly.
Ask Whether the Risk Is Necessary Before Asking Whether It Is Acceptable
These questions sound similar but should occur in sequence.
Is the risk necessary?
Could the research question be answered adequately without creating this exposure or with a safer procedure?
Is the residual risk acceptable?
After reasonable minimization, is the remaining probability and magnitude of harm ethically justified?
Jumping directly to the second question can normalize risks that never needed to exist. A favorable risk-benefit assessment should therefore be based on residual risk after appropriate minimization, not on an unnecessarily hazardous initial design.
Some Risk May Be Intrinsic to the Research Question
Consider research evaluating a new intervention whose adverse effects are not yet fully characterized. Some uncertainty is precisely why the research is needed. Removing every uncertainty would require already knowing the answer the study is designed to establish.
Similarly, a qualitative study of traumatic experiences may require discussing difficult events. Removing every question capable of producing emotional discomfort might also remove the information necessary to answer the research question.
In these situations, the presence of risk does not end the ethical analysis. Researchers should ask whether the exposure is scientifically necessary, whether a less risky approach could still answer the question, what safeguards are feasible, and whether the residual risk is justified.
Minimization Does Not Always Mean Choosing the Safest Possible Procedure
The safest imaginable protocol may be scientifically useless. A study that collects no sensitive information, administers no intervention, and asks participants to do nothing would indeed remove many risks, but it would not answer many research questions either.
The relevant goal is a scientifically adequate design that avoids unnecessary exposure. Risk reduction should therefore be considered alongside methodological validity and research value.
This does not create permission to favor convenience or scientific ambition over participant protection. Rather, it recognizes that an ethically acceptable protocol must be both sufficiently informative and appropriately protective.
Risk Can Often Be Reduced Without Being Eliminated
Risk minimization can occur through many parts of a protocol.
| Source of risk |
Possible minimization strategy |
Risk that may remain |
| Sensitive identifiable data |
Collect fewer identifiers, restrict access, separate identifiers from research data, and strengthen security |
Residual possibility of unauthorized disclosure or re-identification |
| Emotionally difficult interview |
Remove unnecessary intrusive questions, train interviewers, permit skipping and stopping, and establish appropriate response procedures |
Some participants may still experience distress |
| Research blood draw |
Use trained personnel, appropriate technique, smaller or fewer samples where scientifically adequate, or clinically collected specimens when suitable |
Pain, bruising, fainting, or other complications may remain possible |
| Experimental intervention |
Use appropriate eligibility criteria, monitoring, dose selection, stopping rules, and clinical management procedures |
Known and unknown adverse effects may remain possible |
The remaining possibility is residual risk. Its existence does not prove that minimization failed.
Do Not Reduce One Risk by Creating a Worse One
Risk minimization sometimes involves trade-offs. A safeguard can introduce a different burden or risk.
For example, collecting emergency contact information may improve safety in some intervention studies while increasing the amount of identifiable information stored. Requiring frequent in-person monitoring may detect adverse effects earlier but substantially increase participant burden. Excluding participants thought to be at greater risk may protect them but could also produce unjustified exclusion or limit evidence relevant to populations who need it.
The 2024 Declaration of Helsinki specifically calls for attention to how benefits, risks, and burdens are distributed and recognizes that research occurs amid structural inequities.
Good minimization therefore asks about the whole protocol rather than mechanically reducing one variable.
Informed Consent Does Not Replace Risk Minimization
A participant's willingness to accept risk is ethically important, but it does not make unnecessary exposure acceptable.
Consent addresses whether a person can make an informed and voluntary decision about participation. Risk minimization addresses whether researchers should expose that person to the risk in the first place. These are different ethical questions.
Watch Out
“Participants will be informed of this risk and may choose whether to participate” is not a sufficient response when the protocol could reasonably remove or reduce the risk without compromising the research objective.
Compensation Does Not Replace Risk Minimization Either
Increasing payment may make participation more attractive, but it does not make a hazardous procedure safer. Nor should researchers treat compensation as a mechanism for purchasing permission to impose otherwise unjustifiable risk.
The question of whether compensation can make unacceptable research risk ethical should therefore be kept separate from the technical task of reducing the risk itself.
Minimal Risk Does Not Mean No Further Minimization Is Needed
A protocol can fall within the applicable minimal-risk threshold and still contain avoidable exposure.
Suppose a survey qualifies as minimal risk but collects unnecessary identifiable information. Its overall classification does not create a reason to retain those identifiers. If removing them preserves the research objective and materially improves protection, minimization remains relevant.
Risk categories are thresholds used for ethical and regulatory purposes. They are not targets researchers should try to approach as closely as possible.
Residual Risk Still Needs Justification
After reasonable minimization, researchers must examine what remains. Under the U.S. Common Rule, risks must be reasonable in relation to anticipated benefits to participants, if any, and the importance of the knowledge reasonably expected to result.
The Declaration of Helsinki requires that the importance of the research objective outweigh risks and burdens and that researchers be confident those risks and burdens have been adequately assessed and can be satisfactorily managed.
Residual risk therefore needs justification, not merely acknowledgment.
Some Residual Risks May Still Be Too Great
“Cannot be eliminated” does not mean “must be accepted.” A procedure might be necessary for answering a research question and still expose participants to a level of harm that cannot ethically be justified.
This becomes particularly important when asking how much risk important knowledge can justify. Scientific importance matters, but it does not create an unlimited license to expose individuals to harm.
Minimization Continues During the Study
Researchers may discover better ways to protect participants after recruitment begins. New safety information may emerge, an adverse event may reveal an underestimated risk, or monitoring may show that a procedure is more burdensome than anticipated.
The Declaration of Helsinki requires risks and burdens to be continuously monitored, assessed, and documented.
When a new risk becomes apparent during research, the appropriate response may include additional safeguards, protocol modification, renewed consent, pausing particular procedures, or reconsidering whether the study should continue.
04 · A Practical Example
What Risk Minimization Looks Like Without Pretending Risk Can Reach Zero
Hypothetical Example
Interviewing Survivors About a Traumatic Experience
A qualitative study examines how adults experienced institutional support following a traumatic event. Answering the research question requires discussing aspects of that experience, creating a possibility of emotional distress.
Ask what can be removed
The researchers eliminate questions seeking graphic details that are not necessary for understanding participants' experiences of institutional support.
Reduce what remains
Interviewers receive appropriate training, participants may skip questions or stop, interviews are conducted privately, and a response plan is established for significant distress.
Recognize residual risk
Even with these safeguards, some participants may become upset when discussing relevant experiences.
Assess justification
The researchers consider the probability and magnitude of that remaining distress, the importance and methodological necessity of the questions, the safeguards, participant population, and the scientific and social value expected from the research.
Eliminating every possibility of distress might require eliminating the substantive interview itself. Ethical design instead removes unnecessary intrusion, reduces foreseeable harm, and then asks whether the residual risk can be justified.
