What Is an Unanticipated Problem, and How Is It Different From an Adverse Event?

Is an unanticipated problem just an unexpected adverse event?

A participant experiences an unexpected medical reaction during a study. That sounds like an unanticipated problem. But what if identifiable research data are stolen and participants are exposed to a new risk of social harm, even though nobody has yet experienced an adverse medical event? What if a participant accidentally receives ten times the protocol-specified dose but experiences no detectable harm?

These examples expose an important distinction in research oversight. An adverse event describes an unfavorable medical occurrence. An unanticipated problem involving risks to subjects or others is a broader concept concerned with previously unrecognized or greater research-related risk.

Under the U.S. Office for Human Research Protections (OHRP) framework, the categories overlap, but they are not interchangeable. Most adverse events are not unanticipated problems, and some unanticipated problems are not adverse events at all.

An unanticipated problem is about newly recognized research risk

The two concepts answer different safety questions

An adverse event begins with what happened to a participant

An adverse event in research generally begins with an unfavorable medical occurrence. OHRP's broad definition includes an abnormal sign, symptom, or disease temporally associated with research participation, whether or not it is considered related to participation. Physical and psychological harms can fall within this concept.

The classification therefore does not by itself tell you whether the event was expected, whether the research caused it, or whether it reveals a new risk.

An unanticipated problem begins with what the incident tells us about risk

OHRP uses a three-part framework. An incident, experience, or outcome is generally considered an unanticipated problem when it meets all three of the following conditions:

This framework is specific to OHRP's interpretation of the HHS requirements for unanticipated problems. Researchers working under other regulators, jurisdictions, institutions, sponsors, or ethics committees should use the terminology and criteria applicable to their studies rather than assuming that every system uses identical definitions.

“Unexpected” has a more precise meaning than “we did not expect this participant to experience it”

An adverse event may be unexpected when its nature, severity, or frequency is inconsistent with the risks described in relevant protocol documents or other appropriate sources of information, and inconsistent with what would be expected from the participant's underlying condition and risk profile.

That means a known complication does not become unexpected merely because the investigator hoped it would not occur. Conversely, a known type of event may become important if it occurs with unexpectedly greater severity or frequency. OHRP notes that determining an unexpectedly high frequency often requires analysis across participants rather than inspection of one isolated case.

“Possibly related” does not require proof of causation

The second criterion asks whether the incident is related or possibly related to participation. OHRP defines possibly related as a reasonable possibility that the incident, experience, or outcome may have been caused by the research procedures.

This is deliberately different from requiring certainty. Relatedness often falls along a continuum, and investigators may need to consider timing, plausible mechanisms, alternative causes, participant characteristics, and other available evidence. The more detailed question of who determines research relatedness therefore matters when safety events are evaluated.

The third criterion is greater risk, not necessarily actual harm

This is the distinction that makes the concept much broader than adverse events. An unanticipated problem can reveal a new or greater risk of physical, psychological, economic, or social harm even if the feared harm never actually occurs.

OHRP provides a particularly clear example: a pharmacy error causes a participant to receive ten times the protocol-specified dose of an experimental agent. Careful observation detects no adverse effect. Because the error was unexpected, research-related, and exposed the participant to greater physical risk than previously recognized, OHRP treats it as an unanticipated problem even though no adverse event resulted.

The relationship is an overlap, not a hierarchy

It may be tempting to think of an unanticipated problem as simply a more serious adverse event. That model is incorrect.

OHRP summarizes this relationship directly: the vast majority of adverse events are not unanticipated problems, a smaller subset of adverse events are unanticipated problems, and some unanticipated problems do not involve adverse events at all.

A serious adverse event is not automatically an unanticipated problem

A serious adverse event is classified according to outcomes such as death, immediate risk of death, hospitalization, significant disability, or other specified seriousness criteria. Seriousness does not automatically establish unexpectedness or relatedness.

A serious event may therefore be expected, unrelated to the research, or both. In such cases, it does not automatically satisfy OHRP's three criteria for an unanticipated problem. Conversely, an event need not be serious to constitute an unanticipated problem if it is unexpected, related or possibly related, and indicates greater risk than previously recognized.

A non-serious adverse event can still reveal an important new risk

OHRP gives an example from behavioral research in which a college student experiences intense sadness and depressed mood after survey questions trigger memories of childhood abuse. The reaction resolves without intervention after a few hours and is not classified as serious. However, because the reaction was unexpected, related to the research, and revealed a previously unrecognized psychological risk, OHRP treats it as an unanticipated problem.

This distinction is particularly important when responding to unexpected participant distress. The absence of hospitalization or lasting injury does not necessarily mean that an incident has no implications for the study's risk assessment.

Unanticipated problems can involve risks to people other than participants

The full term is unanticipated problems involving risks to subjects or others. That final phrase matters.

OHRP's framework can encompass incidents that expose other people to newly recognized research-related risks. Its examples include circumstances involving contaminated investigational material that could create risks not only for participants but potentially for others, such as their sexual partners.

A data breach can be an unanticipated problem without being an adverse event

This is closely analogous to an example used by OHRP involving theft of an unencrypted laptop containing identifiable sensitive research data. The regulatory concern is the newly created risk, not merely whether harm has already materialized.

Common mistakes when distinguishing unanticipated problems from adverse events

Evaluate the risk revealed by the incident, not just the harm that occurred

When something goes wrong in a study, do not stop after deciding whether it qualifies as an AE or SAE. Ask whether the incident changes what is known about the risks of the research.

A genuine unanticipated problem may have implications beyond filing a report. OHRP notes that such events commonly warrant consideration of protocol changes, changes to informed consent, or other corrective measures. When a safety signal has broader implications, researchers may need to determine whether the participant-safety problem requires changes to the study.

Before deciding whether an incident is an unanticipated problem

Frequently asked questions about unanticipated problems

An adverse event describes an occurrence; an unanticipated problem identifies a newly important risk

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.