Research Versus Clinical Genetic Testing

A sequencing result generated in a research laboratory and one generated in a clinical laboratory can describe the same variant and mean entirely different things for patient care. The distinction is regulatory, not technical, and it governs whether a result may be placed in the medical record, acted on clinically, or used to counsel relatives. Families rarely understand the difference, and explaining it is part of consent.

• CLIA certification is the dividing line. The Clinical Laboratory Improvement Amendments govern laboratories that report results for clinical use. A research laboratory is generally not CLIA certified for the assay it is running, so its results are not validated for patient care and may not be used to direct management.
• A research result must be clinically confirmed before it is acted upon. The standard practice is to re-test on a clinical sample in a CLIA-certified laboratory. Until that happens, the finding is a hypothesis, not a diagnosis, and it does not belong in the chart as one.
• Analytic rigor differs in both directions. Research pipelines are sometimes more sensitive than clinical ones and may detect findings a clinical test would miss. They also lack the validation, quality control, and reporting standards that make a clinical result defensible. Sensitive is not the same as reliable.
• Variant interpretation standards differ. A clinical laboratory classifies variants against ACMG and AMP criteria and issues a report with defined categories. A research group may report a candidate gene with far weaker evidence, and that candidate may be reclassified or abandoned.
• Return of results is not automatic. Whether a research study returns individual findings at all is a design decision made in the protocol and disclosed in consent. Many studies return nothing, which surprises participants who assumed enrollment was a route to an answer.
• The reverse boundary also matters: clinical data used for research. Leftover clinical samples and clinical sequencing data are frequently valuable for research, and using them requires consent or an IRB waiver, plus attention to identifiability.
• Ordering practicalities differ. Research testing is typically free to the family and slow, sometimes taking years. Clinical testing is billed and returns on a defined timeline. Families making a decision between them need both facts.
• Documentation. A research result should not be entered into the medical record as a clinical diagnosis. Where it is documented, it should be labeled as a research finding pending clinical confirmation.

• A child with an undiagnosed condition is enrolled in a research sequencing study, which identifies a candidate variant in a gene with two prior published cases. The counselor explains that this is a lead rather than a diagnosis, arranges clinical confirmation in a CLIA laboratory, and defers cascade testing of relatives and any management change until that returns.
• A family asks whether the research study can replace the clinical panel their insurance has approved. The counselor lays out the tradeoff plainly: the panel returns in three weeks with a usable clinical report, and the study is free but may take two years and may return nothing.
• A research group contacts a family five years after enrollment because a gene of uncertain significance has since been established as disease-causing. The protocol permitted recontact and the family had consented to it. Clinical confirmation is arranged, and the diagnosis is finally made.
• A clinical exome is negative. The counselor discusses reanalysis and research enrollment as distinct next steps, noting that data reanalysis by the clinical laboratory has a defined yield and may be simpler than a new research protocol.

• Do not counsel relatives off an unconfirmed research result. Cascade testing built on a finding that is later withdrawn causes real harm, including reproductive decisions made on a variant that turns out to mean nothing.
• Do not let a research report enter the chart as a clinical diagnosis. Once it is there, downstream clinicians will act on it.
• Do not promise that enrollment will yield an answer. Most participants in most studies receive nothing, and this is the single most common misunderstanding at consent.
• Do not assume research results will be returned at all. Check the protocol before implying otherwise.
• Do not treat a clinical negative and a research negative as equivalent. They tested different things to different standards.