IVD Clinical Performance Evaluation: What Most Teams Miss

Written by HATEM RABEH, MD, MSc Ing

Your Clinical Evaluation Expert And Partner

in
S

Clinical performance evaluation for in vitro diagnostic medical devices is not a variation of MDR clinical evaluation. It follows different logic, different evidence hierarchies, and different acceptability criteria. Yet many teams approach IVDR submissions with MDR assumptions, and the result is systematic deficiencies.

I see this pattern repeatedly: analytical performance data presented as if it demonstrated clinical performance. Design verification studies positioned as clinical evidence. Literature searches that never address clinical validity. Performance evaluation reports that read like technical files with a clinical cover page.

This happens because the terminology overlaps, but the substance does not. Understanding what IVDR actually requires for clinical performance evaluation is the difference between a file that progresses and one that stalls at first review.

What Clinical Performance Means Under IVDR

IVDR Annex I, Section 9.1 defines clinical performance as the ability of a device to yield results that are correlated with a particular clinical condition or physiological state in the target population and intended user.

This is not about whether your assay works in the laboratory. It is about whether the results you generate are clinically meaningful for the decisions they inform.

Two questions drive this definition:

Does the test measure what it claims to measure in real clinical conditions? That is clinical validity.

Does the test provide information that supports appropriate clinical decision-making? That is clinical utility when applicable.

Most deficiencies I encounter stem from teams stopping at analytical performance. They validate precision, accuracy, linearity, and sensitivity. Then they assume this proves clinical performance.

It does not. Analytical validation tells you the assay performs reliably under controlled conditions. Clinical performance evaluation tells you the assay performs meaningfully in clinical practice.

The IVDR Evidence Hierarchy That Actually Matters

IVDR Annex XIII lays out the conformity assessment routes. For most IVDs, especially Class C and D devices, clinical performance evaluation relies on clinical performance studies or existing clinical data.

The hierarchy works like this:

Direct clinical performance studies conducted with your device provide the strongest evidence. These studies assess diagnostic accuracy, clinical sensitivity and specificity, predictive values, and correlation with clinical outcomes in your intended population.

Scientific validity evidence establishes that the analyte or biomarker you measure is relevant to the clinical condition. This is foundational. If the association between your target analyte and the clinical state is not scientifically established, no amount of analytical precision saves the file.

Clinical validity evidence demonstrates that your device accurately detects or measures the analyte in the target population and that the results correlate with the clinical condition. This typically requires comparative studies, diagnostic accuracy studies, or correlation with reference methods in clinical samples.

Clinical utility evidence shows that using your test results leads to improved patient outcomes or more appropriate clinical management. This is required for certain high-risk devices and claims.

The problem is not that teams lack data. The problem is they present the wrong data for the wrong purpose.

Scientific Validity: The Foundation That Cannot Be Assumed

Before you evaluate whether your device demonstrates clinical performance, you must establish that the analyte or biomarker you measure is scientifically valid for the intended purpose.

Scientific validity means the association between the measurand and the clinical condition is established in the scientific literature. For example, the relationship between troponin levels and myocardial infarction is scientifically validated through decades of research.

This sounds obvious, but I see files where scientific validity is referenced without documentation. The manufacturer assumes the association is known. The reviewer requires proof that the association is established and that your device measures the relevant form, isoform, or concentration range.

Your literature search must address this explicitly. You need published evidence that:

The analyte correlates with the clinical condition in the target population. The measurement range you provide is clinically relevant. The form or variant you detect is the one that matters clinically.

Without this foundation, clinical validity cannot be assessed. And without clinical validity, you have no clinical performance.

Clinical Validity: Where Most Files Fail

Clinical validity is where the gap between analytical data and clinical evidence becomes impossible to ignore.

MDCG 2022-2 provides guidance on clinical evidence for IVD medical devices. It clarifies that clinical validity requires evidence from clinical samples, collected from the intended use population, under conditions that reflect actual use.

This means your validation studies on contrived samples or spiked specimens do not demonstrate clinical validity. They demonstrate analytical performance in controlled conditions.

Clinical validity requires studies that:

Use clinical samples from patients with and without the condition. Compare your device results to a reference method or clinical diagnosis. Assess diagnostic sensitivity and specificity in the target population. Account for potential confounders and interfering substances in clinical use.

The most common mistake I see is presenting correlation studies with a reference method as if this completes clinical validity. It does not. Correlation with a reference method shows your device measures similarly to another device. It does not show that your results correlate with the clinical condition unless the reference method itself has established clinical validity.

You must trace the evidence back to clinical outcomes or diagnostic accuracy against a clinical reference standard, not just another IVD.

Clinical Utility: When It Applies and What It Requires

Clinical utility is required when your device claims to influence clinical management decisions or improve patient outcomes.

If your device is marketed as a tool to guide therapy, stratify risk, or inform treatment decisions, you need evidence that using your device results in better clinical management compared to not using it or using alternative methods.

This is a higher bar than clinical validity. Clinical validity shows your test accurately identifies the condition. Clinical utility shows that knowing your test result improves care.

I see manufacturers avoid this by limiting their claims. They describe the test results without explicitly stating how they should be used. But if the intended use implies decision-making, the Notified Body will expect utility evidence.

Clinical utility studies are typically comparative and outcome-based. They assess whether patient management differs when your test is used, and whether outcomes improve as a result.

These studies are resource-intensive, which is why many manufacturers try to sidestep the requirement. But for high-risk devices or tests that claim to guide therapy, there is no alternative.

The Performance Evaluation Report Structure That Holds

The Clinical Performance Evaluation Report is the document that synthesizes all your evidence into a coherent argument.

It is not a data dump. It is not a technical file with clinical sections. It is a structured evaluation that appraises your evidence against the performance claims.

The structure should follow this logic:

Start with the scope and objectives. Define the device, intended use, target population, and performance claims clearly.

Establish scientific validity. Present the literature evidence that the analyte is relevant to the clinical condition.

Evaluate clinical validity. Appraise the evidence that your device accurately measures the analyte in clinical conditions and that results correlate with the clinical state.

Assess clinical utility if applicable. Present evidence that using your device improves clinical management or patient outcomes.

Address risks and limitations. Discuss potential interferences, known limitations, and residual uncertainties.

Conclude with a performance determination. State whether the device meets the performance claims and whether the benefit-risk profile is acceptable.

The appraisal is what differates a compliant report from a documentation exercise. You must evaluate the quality of each study, assess its relevance to your claims, and weigh the evidence collectively.

Literature Search Strategy for IVDR: Different from MDR

The literature search for IVDR performance evaluation is not the same as an MDR clinical evaluation literature review.

For MDR, you search for clinical data on equivalent or similar devices. For IVDR, you search for evidence on the clinical validity and utility of the measurand and the diagnostic approach, not just data on competitor devices.

Your search must cover:

Scientific validity of the analyte or biomarker. Clinical validity studies assessing diagnostic accuracy in the target population. Clinical utility studies if your claims imply decision-making impact. Performance data on predicate or similar devices if you use comparative approaches.

The search terms and databases differ accordingly. PubMed and Embase remain primary, but you also need disease-specific registries, guideline databases, and diagnostic accuracy study repositories.

And the inclusion criteria must align with your evaluation objectives. You are not just looking for device data. You are looking for clinical evidence that supports your performance claims.

When Equivalence Arguments Work and When They Fail

Equivalence approaches under IVDR are more restricted than under MDR. IVDR Annex XIII specifies conditions under which you can rely on existing data from another device.

For Class D devices, equivalence is generally not accepted. You need device-specific clinical performance studies.

For Class C and some Class B devices, equivalence may be possible if you can demonstrate:

Technical and biological equivalence to a device with established clinical performance. Access to the clinical performance data of the equivalent device. That the clinical performance of the equivalent device is adequate for your claims.

The equivalence argument must be detailed and justified. You need side-by-side comparison of technical characteristics, analytical performance, and clinical performance claims. And you need documented access to the predicate device data, typically through agreements with the manufacturer.

Most attempts at equivalence fail because manufacturers claim similarity based on technology or target analyte without demonstrating access to the actual clinical performance data.

Post-Market Performance Follow-Up: The Ongoing Obligation

Clinical performance evaluation does not end at conformity assessment. IVDR requires Post-Market Performance Follow-Up (PMPF) as part of your post-market surveillance.

PMPF is the continuous process of collecting and evaluating clinical performance data to confirm that the device continues to perform as intended in real-world use.

This means you need a PMPF plan that defines:

What performance data you will collect and how. The methods for data collection: registries, surveys, literature monitoring, or dedicated studies. The frequency of data review and report updates. The criteria that would trigger corrective actions or updates to the performance evaluation.

The PMPF report updates the clinical performance evaluation report based on new data. It must be submitted periodically as part of your technical documentation updates.

I see PMPF plans that describe data collection without defining how that data will be analyzed or what performance thresholds are being monitored. This fails the requirement. PMPF is not passive data gathering. It is active performance confirmation with predefined evaluation criteria.

What Happens When Clinical Performance Cannot Be Established

Some devices cannot demonstrate clinical performance through conventional studies. For example, novel biomarkers without established scientific validity, or tests for rare conditions where clinical studies are not feasible.

IVDR Annex XIII allows for expert panels or special procedures under certain conditions. But these are exceptions, not alternatives to proper evaluation.

If you cannot establish clinical performance, you have three options:

Limit your claims to what can be supported. Instead of claiming diagnostic performance, claim analytical detection only. Conduct the necessary studies before market access. Delay launch until evidence is sufficient. Seek expert panel review for Class D devices where clinical performance studies are not feasible due to rarity or ethical constraints.

The third option is rarely granted and requires substantial justification.

Most manufacturers facing this situation attempt to reword claims to avoid clinical performance requirements. This usually fails because the intended use still implies clinical interpretation, and reviewers see through the rewording.

The right approach is to generate the evidence. If your device is intended to inform clinical decisions, you need clinical performance data. There is no regulatory shortcut around this.

The Practical Path Forward

Clinical performance evaluation under IVDR is not a documentation project. It is an evidence-based process that determines whether your device can substantiate its claims.

The sequence is always the same:

Establish scientific validity of your measurand. Demonstrate clinical validity with real clinical samples in the target population. Provide clinical utility evidence if your claims imply decision-making impact. Maintain performance confirmation through PMPF.

Skipping steps or substituting analytical data for clinical evidence does not work. The regulatory logic is consistent, and reviewers apply it systematically.

If your file has been flagged for insufficient clinical performance evidence, the reason is usually clear: you presented analytical validation where clinical validity was required, or you relied on equivalence without data access, or you searched for device performance without addressing scientific and clinical validity.

Fixing these deficiencies requires understanding the distinction between analytical and clinical evidence, and structuring your evaluation accordingly.

The manufacturers who navigate IVDR successfully are the ones who start with clinical performance objectives and work backward to design the studies and gather the evidence those objectives require.

They do not assume analytical performance proves clinical value. They demonstrate it with appropriate evidence from the intended use population.

That is the approach that survives review.

Peace,
Hatem
Clinical Evaluation Expert for Medical Devices
Follow me for more insights and practical advice.