MedDeviceGuideMedDeviceGuide
Back

Contrived IVD Specimens: Which Claims Still Require Native Patient Samples?

Learn which IVD claims contrived specimens can carry, which still require native clinical samples, and how FDA rules and EU IVDR Annex XIII separate them.

Ran Chen
Ran Chen
Global MedTech Expert | 10× MedTech Global Access
Published 2026-11-15Last reviewed 2026-11-1519 min read

When an in vitro diagnostic (IVD) validation package supports every positive result with analyte spiked into a negative matrix, the regulatory question is not whether contrived testing is valid. The question is which individual claims in the instructions for use (IFU) that dataset can actually carry, and which still require native patient specimens.

Contrived specimens can carry analytical claims: limit of detection (LoD), inclusivity, precision across the measuring range, analytical interference, cross-reactivity, and carryover. They can carry a limited clinical claim only when the material is built on a real negative clinical matrix, the analyte form is justified, and the result is anchored to native clinical specimens from the intended-use population. They do not, by themselves, establish clinical sensitivity or specificity, performance across each claimed specimen type, in vivo matrix effects, prevalence-dependent predictive values, or patient-level performance across genotypes and symptom stages.

What a contrived specimen can and cannot establish

A contrived specimen is a laboratory-prepared material in which a known amount of analyte or organism is added to a matrix; it is not a native patient specimen (FDA RVP Special Controls Guidance; MedTech Current analysis of FDA November 2026 guidance). That distinction defines the scope of every conclusion drawn from it. A spiking experiment answers a concentration and reactivity question under controlled conditions: given a known target input in a defined medium, does the assay detect it reliably, repeatedly, and without cross-reaction?

What a contrived experiment cannot answer on its own is whether the assay detects the target as it actually presents in patients. Industry technical commentary from Arete Biosciences highlights four recurring gaps when spiked samples stand in for patient material: endogenous matrix interference, analyte heterogeneity (for example, isoforms, glycosylation variants, or bound versus free fractions), target sequence or antigenic diversity, and physical specimen properties such as viscosity, clots, particulates, and volume variability.

Reviewers therefore get better answers when they evaluate contrived evidence at the claim level rather than treating an entire validation package as binary pass or fail. A single 510(k), De Novo, PMA, or IVDR technical file commonly uses contrived materials legitimately in analytical studies—covered in detail in our IVD analytical performance validation guide—while reserving clinical performance claims for patient specimens. Related questions on transferring an existing claim across sample types or interpreting sequence-based variant coverage belong to matrix equivalence bridging and in silico inclusivity boundaries. Here, the focus stays strictly on a single package: which claims survive contrived-only evidence, and what remains open.

Terminology that changes what a dataset supports

Validation plans and vendor protocols routinely blur terms such as contrived, spiked, surrogate, synthetic, banked, and archived. Regulatory texts do not treat those terms as interchangeable, because each choice alters both biological realism and selection bias. Before mapping evidence to an IFU claim, establish exactly which material the study used.

Specimen or Material TermWhat It Actually IsWhat It Controls vs. What It OmitsTypical Regulatory Handling
Native (fresh) clinical specimenUnmodified specimen collected prospectively from a patient in the intended-use population and tested under IFU conditions.Preserves natural analyte form, patient-level matrix variability, and clinical prevalence; harder to standardize concentration.Preferred basis for clinical performance (FDA RVP guidance; Nov 2026 FDA emerging-pathogen guidance) and, for genetic health risk tests, the first-choice material for accuracy studies (21 CFR 866.5950).
Archived clinical specimenPreviously collected native patient specimen stored (often frozen) and tested later.Retains patient matrix and natural analyte, but freeze-thaw cycles and storage time can degrade fragile targets or alter matrix behavior.Permitted to supplement or substitute when fresh collection is impractical, subject to stability and handling justification.
Banked preselected specimenStored clinical specimens chosen specifically for known positive or negative status rather than sequential enrollment.Useful for rare or low-prevalence targets, but selection bias distorts prevalence and spectrum of disease.Reported separately as PPA and NPA rather than predictive values, with a labeling limitation (FDA RVP guidance).
Contrived or spiked specimen (real matrix)Known concentration of organism or analyte added into an individual known-negative native clinical specimen.Controls target concentration while retaining real donor-to-donor matrix variation, but spike form may still differ from endogenous in vivo analyte.Standard for analytical LoD and interference; allowed for clinical/accuracy gaps only in limited or emergency circumstances and reported separately.
Pooled negative matrix (spiked)Multiple known-negative patient specimens mixed together before spiking with target analyte.Reduces donor-to-donor variance by averaging out individual inhibitors, pH extremes, or interferents.Nov 2026 FDA emerging-pathogen guidance accepts individual or pooled natural negative matrix for LoD once specimens are available, but builds contrived clinical samples on individual (not pooled) specimens; 21 CFR 866.5830 treats pooled samples as contrived specimens.
Transport medium or artificial (surrogate) matrixBuffer, viral transport medium (VTM), saline, or stripped/synthetic fluid lacking true patient cellular and protein content.Controls dilution precisely but omits patient-derived nucleases, proteases, blood, mucus, and competitive background.Transport medium alone is not real clinical matrix, and simulated or artificial matrix should not be used in an LoD study (Nov 2026 FDA emerging-pathogen guidance).
Synthetic construct or purified analyteIn vitro transcribed RNA, plasmid DNA, synthetic peptides, or recombinant proteins added to matrix or buffer.Establishes sequence or epitope reactivity, but bypasses extraction lysis, capsid disruption, or native protein complexing.Nov 2026 FDA emergency guidance would likely accept synthetic material when pathogen stocks are unavailable, if it closely mimics the natural target (e.g., synthetic RNA for an RNA virus); recombinant antigen should not be used for LoD.
Human cell line or infected cell materialCultured cells harboring a target genetic variant or infected with a pathogen, diluted into matrix.Captures cellular context better than naked nucleic acid, yet cultured cells lack clinical sample heterogeneity and patient pathology.FDA RVP guidance describes LoD by limiting dilution of infected cell samples; 21 CFR 866.5950 allows human cell line samples only in some limited circumstances, with results presented separately.

The claim-by-claim separation table

To evaluate a contrived-only validation package without conflating analytical performance and clinical performance, map each proposed IFU or technical-file claim against what spiking data demonstrate, what native patient specimens add, and where the governing regulatory text draws the boundary.

IFU or Validation ClaimWhat Contrived Evidence ShowsWhat Native Clinical Evidence AddsRegulatory Anchor
Limit of detection (LoD)Lowest target concentration detected reliably under controlled dilution in real negative clinical matrix.Confirms whether endogenous analyte accessibility and extraction recovery match spiked dilutions near the cutoff.FDA RVP 2009 guidance (LoD for each specimen type, e.g., by limiting dilution of infected cell samples); Nov 2026 FDA emerging-pathogen guidance (real clinical matrix; no artificial matrix in LoD studies).
Inclusivity (strains and variants)Wet-lab reactivity against quantified strains, isolates, or constructs across target lineages.Tests whether circulating patient strains behave comparably in real clinical matrices across viral or bacterial loads.FDA RVP 2009 guidance, Section 6.B.ii (analytical reactivity); EU IVDR Annex XIII Section 1.2.2 with Annex I Section 9.1(a).
Precision (repeatability and reproducibility)Within-run, between-day, and multi-site imprecision across target levels spanning the measuring range.Captures matrix-driven variability near medical decision points that homogenized pools can mask.21 CFR 866.5830, brain trauma assessment tests (at least one unmodified clinical specimen near the medical decision point; contrived specimens may cover the measuring range).
Analytical specificity — interfering substancesAssay resilience when defined concentrations of potential interferents are spiked alongside known target.Exposes in vivo metabolites, medication complexes, and specimen-specific inhibitors that single-compound spiking misses.FDA RVP 2009 guidance explicitly cautions that spiking experiments may not accurately model the in vivo scenario.
Analytical specificity — cross-reactivityAbsence of false positives when high-titer non-target organisms or structurally related analytes are tested.Verifies background co-infections, commensal flora, and endogenous cross-reactants encountered in patients.Nov 2026 FDA guidance (cross-reactivity and interference studies expected when clinical evaluation lacks natural specimens); EU IVDR Annex XIII.
Carryover and cross-contaminationWhether high-positive samples contaminate adjacent negative runs during automated or manual processing.Demonstrates whether physical specimen viscosity or mucoid clumping increases splash or tip-retention risk.FDA RVP 2009 guidance, Section 6.B.vi (alternating high-positive and high-negative samples in carryover and cross-contamination studies).
Clinical sensitivity or PPAShows only agreement against known spiked inputs; does not measure detection across natural disease presentation.Establishes true-positive detection across symptom onset, specimen quality, analyte shedding, and comparator agreement.FDA RVP 2009 guidance (prospective intended-use population; banked preselected results reported separately); 21 CFR 866.5950 by analogy (clinical-specimen and contrived accuracy results presented separately); EU IVDR Annex XIII Section 1.2.3.
Clinical specificity or NPAUnspiked negative matrix runs test reagent background only, often on a limited number of healthy or pooled donors.Measures false positives caused by real clinical co-morbidities, inflammatory states, and non-target flora in the intended population.FDA RVP 2009 guidance (prospective intended-use population); EU IVDR Annex XIII Section 1.2.3.
Claimed specimen type and matrixDemonstrates compatibility only with the specific negative matrix used for spiking, not unstudied matrices.Confirms collection device interaction, extraction efficiency, and inhibitor profile in each claimed sample type (e.g., swab vs. wash vs. plasma).FDA RVP 2009 guidance expects clinical samples from each claimed specimen type and matrix.
Intended-use population and subgroupsProvides no clinical evidence for pediatric, symptomatic vs. asymptomatic, or immune-compromised subgroups.Tests clinical spectrum, co-medications, and disease prevalence differences across target clinical settings.21 CFR 866.5830; FDA RVP 2009 guidance; EU IVDR Annex XIII clinical performance requirements.
Predictive values (PPV and NPV)Cannot be computed from contrived panels because positive rate is set artificially by the operator.Requires prospective prevalence in the intended-use population alongside verified clinical sensitivity and specificity.EU IVDR Annex XIII Section 1.2.3 with Annex I Section 9.1(b) (diagnostic sensitivity, specificity, PPV, NPV); FDA RVP 2009 guidance (banked preselected results calculated as PPA and NPA).
flowchart TD
  A["Start: Single IFU or Validation Claim"] --> B{"Is the claim purely analytical?"}
  B -- Yes --> C{"Built on real negative clinical matrix?"}
  C -- Yes --> D["Contrived data can support analytical claim"]
  C -- No --> E["Artificial matrix or transport medium only"]
  E --> F["Hold claim: repeat LoD or matrix study in real clinical matrix"]
  B -- No --> G{"Are native clinical specimens available?"}
  G -- Yes --> H["Run clinical study in intended population and each claimed matrix"]
  H --> I["Present clinical and contrived datasets separately"]
  G -- No --> J{"Qualifying rare analyte or Section 564 emergency?"}
  J -- No --> K["Hold clinical claim: do not extrapolate from spiked panel"]
  J -- Yes --> L["Document gap, state labeling limitation, or accept post-authorization condition"]
Decision flow for separating analytical claims from clinical claims in a contrived validation package

Why the matrix, pooling, and analyte form change the answer

Two validation reports can both describe their samples as contrived and still carry vastly different evidentiary weight. When a reviewer reads a contrived study summary, four technical variables determine whether the data can even support the analytical claim—let alone bridge toward a clinical limitation.

First is individual negative clinical matrix versus pooled negative matrix. Spiking analyte into individual known-negative clinical specimens from distinct donors exposes the assay to realistic donor-to-donor variation in endogenous proteins, mucus, particulate matter, cellular debris, and mild inhibitors. Pooling negative specimens from several donors into one homogeneous batch dilutes outlier inhibitors toward an average. FDA's November 2026 emerging-pathogen guidance reflects that distinction: it defines contrived clinical samples as known concentrations of a microorganism or analyte placed into individual (not pooled) human specimens known to be negative, and generally recommends individual negative matrix for confirmatory LoD and for constructing contrived specimens, while accepting individual or pooled natural negative matrix for LoD once more specimens become available (FDA final guidance PDF). A pooled-matrix LoD can therefore be a legitimate analytical result and still say little about donor-level failures in a clinical panel.

Second is real clinical matrix versus transport medium or artificial matrix. FDA's final emerging-pathogen guidance states that liquid transport media without clinical matrix are not considered real clinical matrix, and that simulated or artificial specimen matrix, or recombinant antigen, should not be used in an LoD study because it does not accurately mimic actual patient samples (FDA final guidance PDF; summarized by MedTech Current). Diluting an organism into clean viral transport medium tests reagent reactivity, not whether the assay extracts and amplifies the target from mucus, epithelial cells, or whole blood.

Third is analyte form. Spiking live or whole inactivated pathogen tests whether lysis, nucleic acid extraction, and purification release the target efficiently. Spiking purified genomic RNA, plasmid DNA, or synthetic transcripts can bypass those extraction barriers. FDA's emergency guidance accordingly asks that synthetic material closely mimic the natural target; for an RNA virus, that means synthetic RNA rather than synthetic DNA in most cases. For host biomarkers, Arete Biosciences notes that a spiked standard may not replicate the isoforms, glycosylation variants, or bound versus free fractions present in native specimens, and that recovery, stability, and binding behavior of spiked analyte often differ from analyte that has lived in the patient matrix. A spiked immunoassay calibrator that reacts well in buffer therefore does not by itself show how the assay behaves with the endogenous protein in patient serum or plasma.

Fourth is concentration distribution. Contrived panels can be weighted toward comfortable multiples of LoD, where detection is close to certain, while patient specimens span a continuous biological range that includes low-titer early or late infections near the clinical cutoff. FDA's emergency guidance addresses this directly: when a clinical performance evaluation is not a prospective study, it recommends that the evaluation include samples that appropriately represent the range of pathogen levels expected in clinical specimens from the intended-use population (FDA final guidance PDF). A high agreement rate derived from mid-to-high spikes therefore says little about clinical sensitivity. When contrived specimens are paired with native specimens, for instance to cover a measuring range under 21 CFR 866.5830 for brain trauma assessment tests, concentration levels should be documented explicitly alongside the native clinical anchor rather than summarized as an aggregate clinical detection rate.

Experimental VariableControlled in a Contrived StudyUncontrolled or Different in Native SpecimensRisk if Extrapolated to Clinical Claims
Analyte formPurified nucleic acid, recombinant protein, cultured virus, or cell line.Intracellular or encapsulated pathogens, degraded fragments, isoforms, and bound biomarker complexes.Overestimates extraction efficiency and epitope binding in patient samples.
Matrix compositionSingle buffer, transport medium, or screened negative matrix.Variable mucin, hemoglobin, lipids, patient medications, and endogenous nucleases.Misses false negatives from inhibition and false positives from background cross-reactivity.
Concentration and distributionOperator-chosen levels, which may sit at fixed multiples of LoD.Skewed low-titer tail in asymptomatic, early-stage, or treated patients.Artificially inflates positive agreement relative to intended-use sensitivity.
PoolingPooled negative background is sometimes used to create uniform volume.Each patient sample presents an unaveraged matrix profile.Masks specimen-level invalid rates, clogging, or donor-specific interference.
Replicates and sitesRepeated testing of prepared aliquots in a controlled lab.Collection technique variation, transport delays, and point-of-care operator handling.Underestimates total pre-analytical and clinical workflow imprecision.
Comparator alignmentTruth is known by gravimetric or volumetric spike target.Clinical truth relies on an imperfect comparator or composite reference standard.Hides discordant-resolution bias and clinical misclassification (see discordant IVD analysis).

What FDA device-specific rules and special controls require

In the United States, there is no single horizontal regulation that defines contrived specimen rules identically for every IVD. Instead, binding requirements appear in device-specific regulations and special controls in 21 CFR Part 866, while nonbinding recommendations appear in device-specific or emergency guidance. Separating binding CFR text from nonbinding guidance is essential when defending a validation plan.

Two codified special controls in 21 CFR Part 866 illustrate exactly how FDA treats contrived samples as narrow analytical substitutes rather than silent replacements for clinical specimens:

  • 21 CFR 866.5950 (Genetic health risk assessment systems): Under 21 CFR 866.5950, accuracy must be evaluated using fresh clinical specimens collected and processed consistently with the instructions for use. Archived clinical samples may substitute or supplement fresh samples when fresh collection is impractical. Contrived samples or human cell line samples are permitted only in some limited circumstances, must mimic clinical specimens as much as feasible, and must provide an unbiased evaluation. Crucially, the regulation requires that accuracy results from clinical specimens be presented separately from results generated with contrived samples or cell lines. In that rule, accuracy means agreement with bidirectional Sanger sequencing or another method FDA identifies as appropriate, so it is a genotype-call accuracy requirement rather than a clinical-sensitivity rule; its value here is the specimen hierarchy and the separate-reporting principle.

  • 21 CFR 866.5830 (Brain trauma assessment tests): Under 21 CFR 866.5830, precision testing must include at minimum one unmodified clinical specimen from the intended-use population near the medical decision point. Contrived specimens generated by pooling multiple samples or spiking purified analyte may be used to cover the measuring range, provided they are prepared to mimic clinical specimens as closely as possible. Clinical performance itself must be demonstrated in the intended-use population with clinically valid endpoints and explicit handling of banked samples.

FDA's October 2009 Class II Special Controls Guidance Document for Respiratory Viral Panel Multiplex Nucleic Acid Assays remains a frequently cited example of how FDA separates analytical spiking from clinical performance. That document is device-specific (scoped to assays under 21 CFR 866.3980), dated, and explicitly nonbinding, so it should not be treated as a universal protocol for unrelated IVDs. Even within that scope, its structural logic is clear:

  • LoD should be determined for each specimen type and each analyte, which can be accomplished by limiting dilutions of infected cell samples into non-infected cell samples—an explicitly analytical use of contrived material.

  • The guidance warns that spiking experiments may not necessarily model the in vivo scenario accurately when evaluating potentially interfering substances.

  • Clinical performance is expected from prospective studies in the intended-use population; in rare cases where a virus has low prevalence, banked preselected specimens may supplement prospective collection, but their results should be presented separately and calculated as positive and negative percent agreement, with a labeling limitation because that performance does not reflect performance or prevalence in the intended-use population.

  • Clinical samples are expected from each claimed specimen type and matrix, rather than extrapolating from the easiest matrix to collect.

Teams sourcing negative matrices and clinical specimens should also verify donor consent, characterization, and stability documentation early, as outlined in our biological specimen sourcing guide, so analytical matrices are traceable before pivotal runs begin. In the submission dossier itself, keeping separate data tables for prospective clinical specimens, banked preselected clinical specimens, and contrived or cell-line challenges—with separate confidence intervals and explicit specimen-source labels—ensures reviewers can audit each claim without reconstructing mixed denominators.

What the EU IVDR requires instead

Manufacturers preparing dual-market submissions often assume that an analytical package accepted by a US reviewer—or an emergency contrived dataset—can be dropped directly into an EU technical documentation file. Under Regulation (EU) 2017/746 (IVDR), FDA and EU requirements are not interchangeable, and Annex XIII draws a structural line between analytical performance and clinical performance.

Annex XIII requires the Performance Evaluation Report (PER) to integrate three distinct pillars: scientific validity, analytical performance, and clinical performance (see our IVDR performance evaluation guide). Through Annex I Section 9.1, which Annex XIII Sections 1.2.2 and 1.2.3 reference, analytical performance covers parameters such as analytical sensitivity, analytical specificity, trueness, precision, accuracy, limits of detection and quantitation, measuring range, linearity, cut-off, and control of known interference and cross-reactions. Clinical performance covers diagnostic sensitivity, diagnostic specificity, positive predictive value, negative predictive value, likelihood ratios, and expected values in normal and affected populations.

Under IVDR Annex XIII, clinical performance may be demonstrated through one or a combination of three sources: clinical performance studies, scientific peer-reviewed literature, or published experience gained by routine diagnostic testing. However, the regulation establishes a clear default: clinical performance studies must be performed unless due justification is provided for relying on other sources of clinical performance data. Annex XIII Section 2.1 adds that the purpose of clinical performance studies is to establish or confirm aspects of device performance that cannot be determined by analytical performance studies, literature, or previous experience gained by routine diagnostic testing.

In practice, that means a contrived analytical dataset cannot silently fill the clinical performance section of an IVDR Performance Evaluation Report. Even when an assay targets a rare analyte where prospective clinical studies are genuinely unfeasible, the manufacturer must document a formal justification under Annex XIII explaining why clinical performance studies are not required or feasible, check whether specific EU common specifications apply, and manage any resulting clinical performance study obligations under IVDR performance study requirements. Where peer-reviewed literature or published routine diagnostic experience supports part of the clinical performance claim, the Clinical Performance Report must still trace how that evidence maps to the device's specific intended purpose, target population, and specimen types rather than treating analytical spiking equivalence as self-evident clinical proof.

The emergency-pathway exception, and what it does not erase

On 10 November 2026, FDA announced the availability of its final guidance, Validation of Certain In Vitro Diagnostic Devices for Emerging Pathogens During a Section 564 Declared Emergency (Federal Register notice 2026-22883, docket FDA-2024-D-2707; replacing the January 2025 draft; downloadable as the final guidance PDF). This guidance applies strictly to pre-EUAs, EUA requests, or tests offered under an applicable enforcement discretion policy during an active Section 564 public health emergency declaration. It is nonbinding and does not create a general marketing pathway for 510(k), De Novo, or PMA clearance or approval.

The November 2026 guidance illustrates why regulatory flexibility during sample scarcity is a sequencing tool, not a redefinition of clinical evidence. Natural clinical specimens remain FDA's preference for clinical performance evaluation. Early in an outbreak—when prospective and archived patient samples are unavailable—contrived samples may be acceptable to support initial authorization.

However, FDA explicitly links that flexibility to the limitations of contrived testing: an EUA granted on contrived validation data would likely carry a condition of authorization requiring the manufacturer to evaluate clinical performance using natural patient specimens once feasible. Published regulatory analysis by MedTech Current highlights three additional boundaries in the final guidance that reviewers should track:

  • Real clinical matrix vs. artificial matrix: The guidance frames contrived samples as known amounts of organism or analyte added to individual known-negative clinical specimens, noting that transport medium alone is not a real clinical matrix and artificial matrix should not be treated as a patient specimen in an LoD study.

  • Analytical specificity remains expected: Cross-reactivity and interference studies are generally expected when clinical evaluation does not rely on natural patient specimens.

  • The 30-positive / 30-negative example is not a universal rule: Once clinical samples and a suitable comparator are generally available, FDA cites at least 30 positive and 30 negative clinical specimens from the intended-use population and specimen type as a typically adequate example for an established highly sensitive viral nucleic acid test. That figure is an example for that specific context—not a universal sample-size standard across all analytes, technologies, or non-emergency submissions.

The final guidance itself adds context in a footnote: before COVID-19, FDA issued emergency authorizations for tests based on 50 contrived positive and 50 contrived negative clinical specimens, while for similar products outside a declared emergency it has generally expected an all-comers study of natural clinical specimens until at least 50 positives were obtained (FDA final guidance PDF). Those figures describe FDA's historical practice in that setting. They are not a sample-size rule for other analytes or for routine premarket submissions, but they show that FDA's own practice has long distinguished contrived panels from natural-specimen studies.

A ten-question review for a contrived-only package

Before finalizing IFU wording, an internal design review, or a regulatory submission summary based on contrived data, run the validation package through these ten questions:

  1. What exact claim is the dataset supporting? Separate analytical claims (LoD, precision, interference) from clinical claims (clinical sensitivity, specificity, PPV, NPV). Never use analytical detection rates as clinical sensitivity.

  2. What is the negative matrix source? Verify whether spikes were prepared in individual known-negative clinical specimens from the target anatomy, rather than pooled matrix, transport medium alone, or artificial buffer.

  3. Was negative matrix pooled before spiking? If donors were pooled, confirm whether individual-donor matrix effects and inhibitor tolerance were evaluated separately.

  4. Does the spiked analyte match the clinical form? Check whether live/whole inactivated organism, cultured cells, synthetic RNA/DNA, or recombinant protein exercises the full extraction, lysis, and detection workflow—or bypasses critical pre-analytical steps.

  5. How are spike concentrations distributed relative to LoD? Check whether contrived positives are clustered at high multiples of LoD or include low-level challenges near the cutoff and medical decision point.

  6. What comparator benchmark was used? Remember that contrived truth reflects gravimetric or volumetric input; it does not test agreement against an independent clinical comparator or reference method.

  7. Is there a native clinical anchor, and is it reported separately? Confirm whether at least some unmodified clinical specimens from the intended-use population were tested (as 21 CFR 866.5830 requires for precision testing of brain trauma assessment tests), and ensure clinical and contrived results are never pooled into a single accuracy percentage (21 CFR 866.5950 requires separate presentation for genetic health risk tests).

  8. Does evidence cover every claimed specimen type? Confirm that each claimed specimen type (such as nasopharyngeal swab, nasal swab, saliva, serum, or plasma) has dedicated matrix and clinical backing rather than assuming cross-matrix equivalence.

  9. Are population and prevalence boundaries respected? Ensure the package does not calculate PPV or NPV from contrived or banked preselected panels, and does not extrapolate to untested symptomatic, asymptomatic, or pediatric cohorts.

  10. Are limitations stated transparently in labeling and the PER? Verify that the IFU and performance evaluation report disclose where contrived or banked preselected specimens were used and state any resulting clinical uncertainty.

When to escalate before the claim is written

Rare genotypes, low-prevalence pathogens, and invasive sample types such as cerebrospinal fluid or tissue biopsies can make prospective native positives genuinely scarce. In those programs, starting with a well-designed contrived analytical study in real negative clinical matrix is often the only practical first step. Escalation is about deciding how to bound the claim honestly rather than forcing contrived data to do clinical work.

  • Hold or narrow the IFU claim: If only contrived data support a newly added specimen type, variant, or analyte level, keep the claim in analytical terms or defer adding that clinical claim until native specimens are collected.

  • Document the exact gap and justification: Under both FDA device-specific expectations and EU IVDR Annex XIII, record why native samples were unavailable, what matrix and analyte form bridged the gap, and what post-market or supplemental study will close it.

  • Align with regulators before locking pivotal studies: When scarcity forces reliance on banked preselected specimens, cell lines, or contrived panels for a primary clinical claim, use an FDA Pre-Submission (Q-Submission) or structured notified-body dialogue—as appropriate to the jurisdiction—to confirm acceptable sample composition and separate reporting rules before running the pivotal protocol.

  • Never treat future studies as present evidence: A planned post-authorization commitment, EUA condition, or pending clinical collection is a regulatory obligation, not completed performance data.

Scope and methodological limits: This article is a documentary analysis of public FDA regulations, FDA guidance, Federal Register notices, EU IVDR text, and published regulatory commentary as of November 2026; it does not predict agency or notified-body review outcomes. FDA guidance documents are nonbinding, device-specific CFR rules govern only their respective classifications, and EU IVDR requirements apply solely within the EU framework. No public dataset quantifies overall acceptance rates or review timelines for contrived-only clinical submissions.