A Result Is Not a Plan
Evidence earns the next step through a patient-controlled genomic action airlock.
Genetic information can sharpen a preventive decision, but a report does not arrive as a care plan. It arrives as a signal with a source, population, method, result class, uncertainty, family context, privacy consequence, and evidence boundary. Every one of those features can change what the result means and whether the next responsible step is action, review, no genomic change, or no test at all.
The operational risk is category drift. A health-system team may call family-history assessment, targeted diagnostic testing, carrier screening, pharmacogenetic testing, tumor testing, direct-to-consumer output, polygenic risk estimation, and population sequencing the same thing: genetic screening. They answer different questions and do not share one eligibility rule, consent process, interpretation method, evidence standard, or downstream action.
The Genomic Action Airlock is a decision-control model for preventive care. It forces each signal through six gates: purpose, test quality, interpretation, actionability, person and preference, and delivery. It produces a versioned Genomic Airlock Decision Record that names the question, test provenance, limits, evidence, result taxonomy, patient choice, ordinary prevention, action or no-change outcome, owner, due date, and re-review trigger.
A genetic result is not a care plan. It is a signal waiting to earn an action.
A safe airlock can end with no test, a deferred decision, clinical confirmation, a guideline-supported preventive option, no genomic change, a family communication option, or a bounded re-review. It never treats more testing as the default, a positive result as destiny, a negative result as no risk, a variant of uncertain significance as a reason to change care, or raw vendor output as authority for an autonomous order.
This framework is an editorial operating model, not a USPSTF, CDC, NCI, ClinGen, FDA, CMS, NIH, NHGRI, FTC, laboratory, payer, legal, or clinical standard. It does not establish testing eligibility, medical necessity, coverage, consent validity, laboratory quality, variant classification, treatment, privacy compliance, or a duty to recontact. Apply the current requirements and qualified judgment governing the exact patient, test, laboratory, condition, evidence, setting, payer, and jurisdiction.
The sections below build the airlock from one preventive decision through test category, provenance, interpretation, actionability, consent, delivery, family choice, reclassification, equity, data lifecycle, and a ninety-day pilot.
Start with a preventive decision, not a testing menu.
Write the clinical question before naming a test. A useful question identifies the person or family context, condition or outcome of concern, evidence-supported eligibility pathway, and plausible decisions that could follow. If no responsible downstream choice can be named, more information may create anxiety, incidental work, privacy exposure, or false reassurance without improving care.
Separate screening from diagnosis and risk assessment. A family-history tool may identify who should receive counseling. Counseling may lead to no test, a narrower test, a different affected relative being tested first, or a deliberate decision not to know. A laboratory result may then alter a preventive option, but only after the later gates are passed.
The current USPSTF final recommendation on BRCA-related cancer risk assessment, genetic counseling, and testing was published August 20, 2019 and is being updated. It describes a targeted pathway for women with relevant personal or family history or ancestry factors, not universal genetic screening and not a standing authorization for every multigene panel. The Task Force is an independent panel, and its recommendation is not an HHS position.
For each proposed pathway, list every realistic exit before launch: not eligible, eligible but declines, counseling first, different person tested first, test ordered, test not completed, result pending, result requires confirmation, result supports an option, result changes nothing, or result enters re-review. This prevents the program from treating completion as the only successful outcome.
Stop category drift before it changes the evidence.
Name the test category, specimen, intended use, and result recipient in plain language. Germline testing can address inherited variation. Tumor or somatic testing can characterize acquired changes in cancer tissue. Carrier screening, newborn screening, diagnostic testing, predictive testing, pharmacogenetic testing, direct-to-consumer testing, and research testing occupy different decision paths. One label cannot safely govern them all.
CDC’s May 15, 2024 genetic-testing overview describes multiple test types and emphasizes that results can be complex. It explains that direct-to-consumer tests do not necessarily examine every disease-associated change and should not stand alone for medical decision-making. The page is patient education, not a local eligibility rule or validation of a product.
Do not move evidence across categories without review. Evidence for a targeted pathogenic variant in a clinically ascertained family does not automatically validate a consumer assay, a different variant, a broad panel, a polygenic score, another ancestry group, an asymptomatic population, or a medication action. The result may look genetically adjacent while the clinical question is materially different.
Maintain a category register with allowed question, disallowed use, approved source, required professional review, confirmation rule, consent elements, EHR destination, family implications, retention policy, and review date. A new vendor, test panel, gene, algorithm, laboratory, specimen, population, or claim must reopen the airlock rather than inherit approval silently.
Make the proposed information earn its place.
Purpose is narrower than interest. State the health outcome or decision the information could influence, the time horizon, who can act, and the alternative path without genetic testing. A prevention program should be able to explain why this question is appropriate now and what the patient would still receive if the genomic path stops.
Use pretest counseling or another appropriately qualified process to surface the possible result classes, uncertainty, incidental or secondary findings where relevant, family implications, emotional consequences, privacy questions, cost, and right not to know. CDC’s genetic-counseling page, dated January 31, 2025, is useful education about the role of counseling; it is not a universal testing-eligibility or staffing rule.
A health system should not offer a pathway merely because it can acquire the specimen or return the report. Capacity includes counseling, interpretation, primary and specialty follow-up, laboratory clarification, medication review where applicable, family communication support, language access, financial navigation, and a route for revised classifications.
Preserve provenance from patient to report.
Test quality begins before analysis. Confirm patient identity, specimen type, collection conditions, chain of custody as applicable, ordering question, consent scope, laboratory, assay, genes or variants covered, regions not covered, analytic method, reference build, classification framework, report version, and whether confirmation is required. A result separated from provenance cannot safely carry its original meaning.
CMS administers the Clinical Laboratory Improvement Amendments program for laboratory testing on human specimens in the United States, with its overview modified May 14, 2026. CLIA status addresses laboratory quality requirements within its scope. It does not by itself establish clinical validity, clinical utility, patient fit, payer coverage, informed consent, or the wisdom of a preventive action.
FDA’s direct-to-consumer testing page explains that oversight and the variants included can differ across tests. A positive result does not mean a person will develop a condition, and a negative result does not eliminate genetic or non-genetic risk. Material consumer findings require an appropriate clinical confirmation and interpretation route before care changes.
Build a quarantine route for incomplete provenance. Do not discard the information, but do not let it activate an order, medication change, surveillance change, family alert, or problem-list label. Assign the work needed to obtain the original report, contact the laboratory, confirm the finding, or document why the signal cannot be resolved.
The quality gate should record limitations in language the downstream clinician and patient can use. A technically accurate report can still be mismatched to the clinical question, incomplete for the relevant variant type, uncertain in the population of interest, or unable to distinguish an inherited from an acquired finding without another specimen or analysis.
Translate the result class before discussing action.
Interpretation joins the result to the clinical question, personal and family history, inheritance pattern, test limitations, current evidence, and relevant population. The same words can have different implications across genes, conditions, laboratories, tissues, and testing contexts. Preserve the laboratory’s exact classification and add the qualified clinical interpretation separately.
For inherited-cancer testing, NCI’s fact sheet reviewed April 18, 2024 describes positive, negative, and uncertain results, the role of counseling and informed consent, possible family implications, and the possibility that variant interpretations change. It is scoped to inherited-cancer testing and is not a classification rule for every genomic use.
Use the words true negative only when the family finding and assay scope make that conclusion appropriate. A person can receive a negative panel and still have elevated risk from family history, variants not detected by the method, genes not tested, multifactorial causes, environment, age, or an explanation not yet known. Preserve the exact residual-risk statement.
For a VUS, separate what is known from what is being watched. The care plan should show that no change is based solely on the uncertain variant, identify any action supported independently by personal or family history, and name a bounded re-review process. Do not ask the patient to repeatedly search for reclassification without a defined reason and owner.
FDA’s table of pharmacogenetic associations states that inclusion does not mean FDA recommends genetic testing before prescribing unless a test is specified in labeling. The table’s latest displayed update log is October 26, 2022. Treat it as one evidence input, not patient-specific medication advice, and route decisions through the current label, indication, prescriber, alternatives, and clinical context.
Make actionability explicit, conditional, and current.
Actionability is not a property of a variant alone. It depends on the condition, evidence, magnitude and type of risk, age and health context, available intervention, likely benefit and burden, timing, contraindications, competing risks, access, patient preference, and who can deliver the next step. A finding can be valid yet produce no genomic care-plan change.
Use professional evidence resources as inputs with their dates and scope. ClinGen’s Actionability Working Group curates evidence about gene-condition pairs and interventions. The live resource, checked August 3, 2026, does not create an autonomous order, replace patient-specific judgment, determine coverage, or guarantee that an intervention is appropriate or available.
Distinguish a recommendation to offer from a requirement to perform. A patient may reasonably decline testing, decline a preventive option, choose later review, or prefer ordinary risk-based care after informed discussion. The record should preserve that choice without labeling it nonadherence or allowing the result to trigger action elsewhere.
Version the action basis. A pathway can change when a guideline, classification, laboratory, gene-condition relationship, label, intervention, age threshold, family finding, health state, coverage rule, or patient preference changes. Do not silently apply today’s result through yesterday’s action logic.
Put consent, preference, and the right not to know inside the workflow.
Consent is a continuing decision process, not a signature detached from the pathway. Before testing, explain the question, alternatives, possible result classes, test limits, uncertain or secondary findings where relevant, family implications, cost, data access, research or vendor uses when applicable, retention, and what the health system can and cannot promise about future reinterpretation.
Ask what the patient wants to know, what they do not want to know, who may receive results, how they prefer to discuss family implications, and whether they want another person present. Use accessible language, interpretation, communication supports, time, and teach-back. Do not infer permission from portal access, family presence, proxy status, or a general consent to treatment.
NHGRI’s genetic-discrimination resource, updated January 6, 2022, explains protections and important gaps in the Genetic Information Nondiscrimination Act. Federal employment protections generally address employers with fifteen or more employees. GINA does not cover life, disability, or long-term-care insurance, and context can differ for military members and under state law. Give a current, jurisdiction-appropriate explanation rather than saying genetic information is simply protected.
Record a safe no-test route. A person who declines testing should still receive ordinary prevention, family-history assessment, symptom evaluation, and referrals supported independently of the genetic test. Declining one information pathway must not become abandonment of preventive care.
Close the loop in language, action, ownership, and time.
Delivery is complete only when the correct patient receives an appropriately interpreted result, understands what it does and does not mean, knows what changes and what stays the same, and has a named next step with an owner and due date. Portal release, an inbox acknowledgment, or a mailed report is not evidence that this loop closed.
Match urgency and communication method to the result, consequence, patient preference, accessibility need, and local rule. Verify identity and preferred language, manage proxy access, document unsuccessful contact, and define escalation without exposing genetic information to unauthorized recipients. Use teach-back for meaning, uncertainty, action or no change, baseline prevention, family option, privacy, and re-review limits.
A decision record remains open while the plan depends on an unaccepted referral, unconfirmed consumer result, uninterpreted report, unsigned recommendation, unavailable service, or patient question. Track the actual endpoint, including a documented choice not to proceed, rather than counting an order or message as preventive care delivered.
Give every result an explicit care-plan lane.
One report can support more than one route, but every route needs its own basis and owner. A pathogenic finding may prompt a current discussion, a later surveillance review, and an optional family communication. A negative or uncertain result may produce no genomic change while ordinary prevention continues. Do not force all signals into action now.
No genomic change is a real clinical outcome, not an empty field. Record why the result does not support a change, what care continues independently, what would reopen review, and who answers questions. This protects against both silent overreaction and the false conclusion that nothing matters.
The EHR should show the lane without converting it into a permanent identity. Keep exact report language and provenance available, distinguish active from historical logic, and retire superseded alerts or actions when evidence, classification, patient context, or preference changes.
Keep baseline prevention visible through every genomic outcome.
Genetic information adds to care; it does not erase age, symptoms, examination, exposures, family history, prior disease, or evidence-based preventive recommendations. Put the ordinary care path beside the genomic lane so a pending, declined, negative, uninformative, or uncertain result cannot suspend screening, risk-factor management, diagnostic evaluation, or urgent response that is supported independently.
For every result, answer two separate questions: what changes because of this result, and what remains indicated without it? A negative direct-to-consumer test does not exclude disease risk. A negative clinical panel may leave an unexplained family pattern. A positive result may add an option without determining whether that option fits this patient now.
Audit for genomic overshadowing. Look for delayed symptom workup, canceled ordinary screening, discontinued risk counseling, or lower follow-up after a reassuring-sounding result. Also look for unnecessary procedures, surveillance, medication changes, or anxiety-driven repeat testing after an unsupported or uncertain signal.
Support a family option without forcing disclosure.
A result can have implications for biological relatives, but family benefit does not erase patient control, privacy, or a relative’s right not to know. Explain what information may be relevant, which relatives might consider counseling, what the finding does and does not establish, and how the patient can share it if they choose. Apply current legal and ethical guidance to difficult exceptions rather than improvising.
CDC’s cascade-testing resource, dated May 15, 2024, describes a process in which relatives may consider testing for a known family finding. It also recognizes that relatives may not want testing or information. The care team should not contact or disclose to relatives without appropriate authorization merely because the information could be useful.
Offer a patient-controlled family letter or secure summary that preserves the exact gene and variant, laboratory, report date, classification, condition context, confirmation status, counseling route, and limits. Avoid copying the patient’s unrelated diagnoses or the relative’s assumed risk. Record that the option was offered, not whether the patient fulfilled an organizational outreach target.
Turn reinterpretation into a governed queue, not an indefinite promise.
Variant classifications, gene-condition evidence, laboratory methods, guidelines, medications, patient health, and family history can change. That does not mean a health system can promise continuous surveillance of every result forever. State the current re-review policy before testing: which results enter a queue, what triggers review, how long the pathway operates, who owns it, how patients update contact information, and what cannot be guaranteed.
Re-review is not the same as automatic recontact, and an amended report is not automatically an action. Pass any change back through provenance, interpretation, actionability, preference, and delivery. Prevent duplicate alerts, contradictory chart entries, or a revised label that survives after the underlying classification is superseded.
Measure access and validity before promising equitable precision.
Equity cannot be inferred from offering the same portal invitation or test panel. Measure who is identified, referred, counseled, tested, reaches interpretation, receives a signed plan or no-change record, completes follow-up, pays out of pocket, and enters re-review. Stratify relevant language, geography, disability, payer, and validation-population factors with appropriate privacy and statistical caution.
An NIH research report dated February 20, 2024 describes work to improve genetic risk scores across diverse populations and highlights performance and access concerns. It is a research report, not a clinical guideline or permission to deploy a polygenic risk score as preventive care. Verify the exact model, outcome, population, calibration, comparator, downstream action, and current evidence.
Do not use race as a genetic result or assume ancestry labels solve validation gaps. Ask whether performance differs in the population served and what happens when uncertainty is higher. Provide language access, counseling routes, confirmation, follow-up capacity, and financial navigation before calling broader testing an equity strategy.
Govern genetic data from invitation through deletion or transfer.
Map the complete data path: eligibility query, invitation, family history, consent, order, specimen, laboratory, vendor, raw and interpreted data, report, EHR, portal, proxy, analytics, research, family letter, amendment, export, retention, deletion, and vendor exit. For each stage, identify purpose, minimum data, authorized role, security control, retention basis, correction route, and patient explanation.
The FTC’s September 7, 2023 final order involving 1Health.io addressed alleged failures involving DNA-data privacy and security representations. It is an enforcement precedent, not a universal legal rule. Still, it shows why consent language, actual data practices, vendor claims, deletion promises, security, and secondary use must agree in operation.
Review current applicable FDA status or policy, CLIA certification, state requirements, intended use, analytical limits, privacy and security obligations, and current legal guidance for the exact pathway. Avoid broad claims that one approval, certification, notice, or consent document establishes the entire program’s validity or compliance.
Prove one closed inherited-cancer referral lane before expanding.
In days one through thirty, choose one existing guideline-supported inherited-cancer referral lane. Map referral to counseling, order, report, interpretation, delivery, plan, and follow-up. Approve the question brief, eligibility version, laboratory checklist, result taxonomy, escalation rule, family letter, privacy explanation, and re-review register. Name genetics, primary-care, laboratory, pharmacy, privacy, equity, operations, and patient-advisor roles.
In days thirty-one through sixty, run twenty-five to forty consecutive referrals through a weekly Airlock huddle. Every case receives an explicit outcome: test not indicated, test deferred, pending, actionable option after professional review, no genomic change, family option, or dated re-review. Clinically confirm material consumer findings before change, and block any automated medication start, stop, substitution, or dose.
In days sixty-one through ninety, audit wrong-patient or wrong-result events, unsupported actions, VUS-only changes, confirmation, baseline prevention, teach-back, privacy, failed contact, closed follow-up, turnaround, out-of-pocket burden, and completion by language, geography, payer, disability, and relevant validation population. Continue, modify, or stop. Expand only when every result class has a safe owner and closed loop.
Retire the shortcuts that turn signals into unsafe plans.
Stop ordering broad panels without a decision question, treating population sequencing as automatically preventive, moving evidence between test categories, calling CLIA certification proof of utility, accepting screenshots as provenance, treating a positive result as destiny, calling a negative result no risk, or changing care solely because of a VUS.
Stop allowing raw consumer output to enter the problem list, a medication rule to fire without prescriber review, portal delivery to count as understanding, family benefit to authorize disclosure, generic GINA language to hide insurance gaps, ordinary prevention to disappear, or a recontact promise to remain without an owner, scope, trigger, and end.
Stop measuring specimens, panels, variants, referrals, or alerts as prevention. Measure the exact question, provenance, interpretation, patient choice, signed action or no-change record, closed follow-up, preserved baseline care, privacy, family option, re-review, access, burden, and adverse events.
Conclusion: Let evidence and patient choice earn every genomic action.
Genetic screening can guide preventive care when the health system begins with one answerable decision, preserves test provenance, interprets the exact result class, verifies current actionability, respects the person’s choices and privacy, and closes delivery with a named owner. The same airlock must be able to produce no test, no genomic change, or later review without calling those outcomes failures.
The durable unit is not a variant in a chart. It is a versioned decision record connecting signal, evidence, preference, ordinary prevention, action or no action, family option, delivery, and re-review. When any link changes, reopen the gates. When the signal cannot pass them, hold it safely.
Sources and further reading
- U.S. Preventive Services Task Force: BRCA-Related Cancer Risk Assessment, Genetic Counseling, and Genetic Testing. Final recommendation published August 20, 2019 and currently being updated; a targeted pathway, not universal genetic screening.
- CDC: Genetic Counseling. Updated January 31, 2025; patient and clinician education, not a universal eligibility or staffing rule.
- CDC: Genetic Testing. Updated May 15, 2024; overview of test types and limits, including why direct-to-consumer output should not stand alone for medical decisions.
- National Cancer Institute: Genetic Testing for Inherited Cancer Risk. Reviewed April 18, 2024; inherited-cancer scope, result taxonomy, counseling, consent, family implications, and reclassification context.
- CDC: About Cascade Testing. May 15, 2024; family testing context that preserves voluntary choice and the right not to know.
- Clinical Genome Resource: Actionability Working Group. Live resource checked August 3, 2026; professional evidence input, not an autonomous order or coverage rule.
- U.S. Food and Drug Administration: Direct-to-Consumer Tests. Live resource checked August 3, 2026; oversight, variant-set, interpretation, and confirmation limitations.
- U.S. Food and Drug Administration: Table of Pharmacogenetic Associations. Latest displayed update log October 26, 2022; inclusion is not a recommendation to test or patient-specific medication advice.
- Centers for Medicare & Medicaid Services: Clinical Laboratory Improvement Amendments. Modified May 14, 2026; laboratory quality oversight does not itself establish clinical utility, coverage, or patient fit.
- National Institutes of Health: Researchers Optimize Genetic Tests for Diverse Populations. February 20, 2024 research report; performance and access context, not a clinical guideline for polygenic risk scores.
- National Human Genome Research Institute: Genetic Discrimination. Updated January 6, 2022; GINA protections and gaps, including life, disability, and long-term-care insurance.
- Federal Trade Commission: Final Order Involving 1Health.io and DNA Data. September 7, 2023 enforcement precedent; not a universal privacy or security rule.
