Return of Research Results

Genomic research generates findings beyond the question being asked, and deciding what to tell participants is one of the genuinely unsettled areas in research ethics. The tension is between respect for participants, who generally say they want their results, and the obligation not to hand people unvalidated information that will drive medical decisions. There is no single correct policy, which is why the protocol must state one in advance.

• Primary results answer the study's own question. Secondary or incidental findings are medically relevant discoveries unrelated to it, most commonly a pathogenic variant in a gene associated with a treatable condition.
• The ACMG secondary findings list governs clinical exome and genome sequencing, where laboratories are expected to report pathogenic variants in a defined set of medically actionable genes unless the patient opts out. Research studies are not bound by this list, but it is the most common reference point when a study designs a return policy.
• Actionability is the usual threshold. Most frameworks return findings that are analytically valid, associated with a serious condition, and clinically actionable, meaning surveillance or intervention exists that changes outcome. Findings that are unvalidated, uncertain, or untreatable are more often withheld, though this is contested.
• The participant's stated preference is not automatically decisive. Participants overwhelmingly say they want everything, including untreatable findings. Study designs vary in how much weight this receives against the risk of returning unvalidated results.
• Anything returned must be clinically confirmed in a CLIA-certified laboratory before it is used for care. A research finding communicated without that step invites action on an unvalidated result.
• The right not to know is real and must be preserved. Consent should allow a participant to decline secondary findings, and that choice should be recorded and honored.
• Duty to recontact is unresolved. Variant classifications change, and a variant of uncertain significance may be reclassified as pathogenic years later. Whether investigators must find and inform former participants is contested; most protocols address it explicitly, and the practical answer usually depends on whether the study retained the ability to contact anyone at all.
• Pediatric participants raise a distinct problem. Returning an adult-onset finding in a child forecloses that child's future choice about whether to know. The general principle is to defer predictive testing for adult-onset conditions until the child can decide, with exceptions where childhood surveillance would change outcome.
• Deceased participants and family members. Findings relevant to relatives, particularly after a participant has died, sit at the boundary of confidentiality obligations and are handled through the protocol rather than improvised.

• A research genome sequencing study identifies a pathogenic BRCA1 variant in a participant enrolled for an unrelated neurologic condition. The protocol returns actionable secondary findings. The counselor confirms the variant clinically, then discloses it with counseling about surveillance and cascade testing for relatives.
• The same study identifies a variant associated with an untreatable adult-onset neurodegenerative condition. The protocol does not return non-actionable findings and the participant was informed of this at consent. The finding is not disclosed, and the counselor can point to a policy the participant agreed to rather than to an improvised decision.
• A 10-year-old enrolled in a research study is found to carry a pathogenic variant for an adult-onset cancer syndrome with no childhood surveillance indication. The team defers disclosure to the child and discusses with parents that the appropriate time is when the child can make the decision, while noting the parents' own risk warrants their testing.
• A participant enrolled eight years earlier had a variant of uncertain significance that has since been reclassified as pathogenic, with surveillance implications. The protocol permitted recontact and current contact information exists. The team reaches her, confirms clinically, and initiates screening.

• Do not return a research result without clinical confirmation. This is the most consequential error in this area.
• Do not leave the return policy to be decided when a finding appears. Deciding under the pressure of a specific actionable variant produces inconsistent and indefensible choices.
• Do not assume every participant wants everything. The right not to know requires an actual opt-out, offered at consent and honored later.
• Do not disclose adult-onset predictive findings in children by default. The child's future autonomy is the reason, and childhood actionability is the exception that overrides it.
• Do not promise recontact you cannot deliver. If the study does not retain identifiers or funding to recontact, saying so at consent is more honest than an aspiration.