Self-Test IVDs: Why Clinical Performance Isn’t Just About Accuracy

Written by HATEM RABEH, MD, MSc Ing

Your Clinical Evaluation Expert And Partner

in
S

I see this pattern repeatedly. A company develops a self-test for glucose, pregnancy, or COVID-19. The core assay performs well in laboratory conditions. The regulatory team assumes that clinical performance for self-testing means proving analytical accuracy.

But IVDR Article 56 and MDCG 2022-2 make something clear: for devices intended for self-testing, clinical performance includes whether non-professional users can correctly perform all steps from sample collection to result interpretation.

This is not about writing better instructions. This is about generating evidence that your intended users, in their real environments, can actually use your device as intended.

What Self-Testing Actually Means Under IVDR

Self-test IVDs are defined as devices intended to be used by lay persons. The critical word is “intended.” If your device is marked for self-testing, you are claiming that non-professionals can operate it safely and effectively without supervision.

IVDR Article 56 requires that manufacturers demonstrate clinical performance throughout the device lifecycle. For self-tests, this explicitly includes usability evidence.

MDCG 2022-2 reinforces this by stating that clinical evidence must address whether lay users can perform the complete testing procedure and correctly interpret results. The guidance also specifies that instructions for use must be validated with representative users.

This dual requirement means your clinical evidence plan must include studies that go beyond laboratory validation. You need data from actual lay users performing the test in unsupervised conditions.

The Evidence Gap I See Most Often

Here is what typically happens. The manufacturer conducts a clinical study. Trained staff collect samples from patients. The samples are tested using the self-test device under controlled conditions. Sensitivity and specificity meet the performance specifications.

The clinical evaluation report presents these results as evidence of clinical performance. The dossier goes to the Notified Body.

The Notified Body asks: where is the evidence that lay users can perform the test correctly?

The manufacturer points to the instructions for use. The Notified Body clarifies: we need evidence, not instructions. Evidence means data showing that representative users, without training, can successfully complete each step of the testing process.

The problem is structural. Analytical performance studies measure what the device can do. Usability studies measure what users can do with the device. Both are required for self-tests.

What Usability Evidence Actually Looks Like

Usability evidence for self-test IVDs must address three critical areas: correct execution of the test procedure, correct interpretation of results, and appropriate action based on results.

First, can users correctly perform sample collection, sample application, and timing? This requires observing lay users as they follow the instructions without assistance. You document where they succeed and where they fail.

I have reviewed studies where users consistently applied too much sample, missed timing windows, or skipped preparation steps. These errors directly impact test performance. If 30 percent of users cannot correctly perform the test, your claimed sensitivity and specificity no longer reflect real-world performance.

Second, can users correctly interpret the result? This includes reading qualitative indicators like lines or color changes and understanding what the result means. Studies must show that users can distinguish positive from negative results and recognize invalid tests.

Third, do users understand what to do with the result? A self-test is only clinically useful if users take appropriate action. This means they must understand when to seek professional care, when to repeat testing, and what the limitations of the test are.

MDCG 2022-2 makes this explicit. The guidance states that usability studies should include participants from the intended user population, including those with relevant limitations like visual impairment or low health literacy.

How Representative Users Are Defined

The phrase “representative users” causes confusion. Manufacturers often interpret this to mean any adult without medical training. That is not sufficient.

Representative users must reflect the demographic and cognitive characteristics of your actual target population. If your device is a pregnancy test, your study population should include women of reproductive age with varying education levels. If your device is a glucose monitor for diabetes management, your population should include elderly users and those with limited dexterity.

I have seen studies where all participants were university students recruited for convenience. That population does not represent the broader lay user base. The Notified Body will question whether your evidence applies to real-world users.

MDCG 2022-2 provides guidance on this. The document states that manufacturers should consider factors like age, health literacy, language proficiency, visual acuity, and manual dexterity when selecting study participants.

This is not about checking boxes. This is about generating evidence that your device works for the people who will actually use it.

The Instructions for Use Problem

Manufacturers often respond to usability gaps by revising the instructions for use. The assumption is that clearer instructions will solve execution errors.

But IVDR requires that instructions be validated with users. This means you must test whether users can successfully perform the test after reading your instructions. If errors persist, the instructions are not adequate.

Here is what I observe in practice. Usability studies reveal that users skip critical steps. The manufacturer adds bold text and warning symbols to the instructions. The next study shows the same errors.

The issue is not always clarity. Sometimes the test procedure is inherently difficult for lay users. Adding more text does not fix a procedure that requires precise timing or fine motor control.

This creates a difficult decision point. If usability studies show persistent failures, you have three options: redesign the device to simplify the procedure, restrict the device to professional use, or generate additional evidence showing that the observed errors do not compromise clinical performance.

None of these options are simple. All of them have regulatory and commercial implications.

Linking Usability Evidence to Clinical Performance

The connection between usability and clinical performance must be explicit in your clinical evaluation. It is not sufficient to have separate usability studies and separate performance studies. You must demonstrate how usability impacts real-world performance.

This means analyzing how user errors affect test results. If users apply insufficient sample volume, what happens to sensitivity? If users read the result too early, what is the rate of false negatives?

I have worked with manufacturers who conducted extensive usability testing but never linked the findings to clinical outcomes. The clinical evaluation report presents sensitivity and specificity from controlled studies. The usability report sits in a separate section describing user errors.

The Notified Body asks: what is the clinical performance when users make these errors? The manufacturer has no data.

This gap can be closed by designing studies that capture both usability and performance. Participants perform the test unsupervised. You collect both the test result and an independent reference standard. You analyze performance based on whether users executed the test correctly.

This is the evidence that demonstrates clinical performance for self-tests. Not just that the device can work, but that it does work when real users perform it.

Post-Market Surveillance for Self-Tests

IVDR Article 56 requires ongoing evaluation of clinical performance. For self-tests, this includes monitoring whether users continue to perform the test correctly as the device reaches broader populations.

Post-market clinical follow-up for self-tests should address usability in real-world conditions. Are users making errors that were not observed in pre-market studies? Are certain user groups experiencing more difficulties?

I see PMCF plans that only track adverse events and complaints. For self-tests, you also need to track usage patterns and user comprehension. This might include post-market surveys, user interviews, or analysis of customer support inquiries.

If post-market data reveals usability issues, you must evaluate whether they impact clinical performance. If they do, you have a corrective action obligation.

This is not theoretical. I have reviewed cases where post-market surveillance revealed that users consistently misinterpreted faint positive lines as negative results. This finding directly impacts claimed sensitivity and requires corrective action.

What This Means for Your Submission

If you are preparing a submission for a self-test IVD, your clinical evidence plan must explicitly address usability. This means designing studies that enroll representative lay users, allow them to perform the complete testing procedure unsupervised, and measure both execution accuracy and result interpretation accuracy.

Your clinical evaluation report must analyze how user performance impacts clinical performance. You cannot present laboratory performance data and assume it reflects real-world performance.

Your PMCF plan must include mechanisms to monitor ongoing usability and detect issues that arise as the device reaches diverse user populations.

This is not about adding more documents to the technical file. This is about structuring your evidence to answer the fundamental question: does this device achieve its intended clinical performance when used by the people it is intended for?

That distinction shapes every part of your clinical evidence strategy. Manufacturers who understand this early build stronger submissions and avoid costly gaps during review.

Final Thought

Self-test IVDs offer significant clinical value by enabling testing outside traditional healthcare settings. But that value only materializes if users can correctly perform the test.

IVDR and the supporting MDCG guidance recognize this. Clinical performance requirements for self-tests are not about laboratory validation alone. They are about demonstrating that your device works in the hands of the people who will use it.

The evidence burden is higher. The studies are more complex. The clinical evaluation must connect usability to performance in explicit terms.

But this is not regulatory burden for its own sake. This is the evidence that shows your device actually delivers on its intended purpose.

Next in this series, we will address how to structure clinical performance studies for companion diagnostics, where performance requirements are linked to therapeutic decisions.

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