Compiled and interpreted based on Article 61.10: Clinical Evaluation based on non-clinical data (BSI, 2025)

Important Note (Statement of Rigor)

This article only provides an objective summary, logical structuring, and Chinese-language restatement of the BSI white paper Article 61.10: Clinical Evaluation based on non-clinical data. All views strictly follow the original intent of the white paper, without introducing additional conclusions or extrapolative interpretation. Any “acceptable/unacceptable” situations discussed herein correspond only to the example scenarios provided in the white paper and do not constitute a prediction regarding any specific product.

I. Main topic addressed by this white paper

The white paper focuses on a frequent point of debate in clinical evaluation under the EU MDR: whether certain lower-risk devices (Class IIa/IIb non-implantable) may, in “exceptional circumstances,” demonstrate conformity without relying on clinical data, and instead rely primarily on non-clinical data (non-clinical / pre-clinical data).

BSI’s core objective is to clarify its interpretation of MDR Article 61(10), its scope of applicability, and review expectations: Article 61(10) is a route for “exceptional circumstances,” not an alternative route that can be used simply because clinical data are unavailable or difficult to collect.

II. BSI’s key definitions and position on “non-clinical data vs. clinical data”

1) What are non-clinical data (non-clinical / pre-clinical)?

The white paper defines non-clinical data as data relevant to safety and performance that do not arise from actual use in humans. Common sources include laboratory/bench testing, animal/cadaver/phantom testing, biocompatibility, usability testing, simulated-use studies, and modeling. Their value lies in providing strong indications or predictions regarding clinical performance and risk.

2) A common point of confusion for software (SaMD):

Even if software training/testing uses retrospective human datasets or images, the white paper still classifies the resulting outputs as non-clinical data. Only when the software is used by intended users within the intended clinical environment and workflow, and is evaluated against the workflow without the software, does it constitute clinical data.

3) Why do clinical data remain important?

The white paper emphasizes that clinical data come from actual use of the device or an equivalent device in humans, whether from pre-market clinical investigations or post-market real-world use. Their value is that they provide objective, verifiable evidence that the device performs as intended in the real clinical environment and does not cause additional or unexpected harm.

III. BSI’s “bottom line” on Article 61(10): it may be used only in exceptional circumstances

The white paper clearly states that relying solely on non-clinical data to demonstrate conformity is based on an assumption: that if a device is well designed and manufactured, it will perform as intended in a disease-state setting without causing unexpected harm. However, this assumption does not always hold true in clinical use, so Article 61(10) should be used only in exceptional circumstances.

BSI provides two very clear review principles:

The applicant must be able to fully justify why clinical data are “not appropriate” for demonstrating safety and performance.
“Lack of clinical evidence,” “difficulty collecting clinical data,” or “insufficient clinical data” are not acceptable reasons, and are even less likely to be accepted when similar or equivalent devices already have clinical data.

IV. Which devices are more likely to qualify, and where is the greatest controversy?

The white paper considers that, for some devices, conformity can be more convincingly demonstrated based on non-clinical data, for example disinfectants, sterilizers, medical refrigerators, and basic tools (such as forceps, surgical gloves, and drapes). The reason is that their performance and safety parameters can be sufficiently covered by non-clinical testing and standards, including product standards and technical standards.

By contrast, for devices used in more specific or limited clinical scenarios (such as biopsy needles, devices used with specific implants or specific procedures, SaMD, and system components), whether clinical data are “appropriate” and whether Article 61(10) can apply are far more controversial. The white paper explains BSI’s position through factor-by-factor analysis and scenario examples.

V. The “five elements of Article 61(10)” that BSI expects you to address one by one

The white paper breaks down whether Article 61(10) can be substantiated into several elements that must be demonstrated and documented. Their common feature is that the justification must follow the logic chain of the Clinical Evaluation rather than merely stacking test reports. Below are the core points emphasized by the white paper item by item:

1) Applicable device scope (classification)

Article 61(10) cannot be used for Class III or implantable devices; the potentially applicable scope is limited to Class IIa and Class IIb non-implantable devices. The paper also mentions discussion around certain Class IIb implantable WET devices, but it more strongly suggests considering the “totality of evidence” route in MDCG 2020-6 Appendix III.

2) Clinical performance

The clinical role/function of the device must be clearly defined: whether it is part of a system or a standalone device, and what impact device failure would have on patient health outcomes.

3) Manufacturer claims — one of the easiest grounds for outright rejection

The white paper distinguishes between clinical claims and non-clinical (technical) claims:

Non-clinical claims: such as ease of use, design, durability, sterility, etc., which are not directly linked to health outcomes.
Clinical claims: positive effects on patient health, patient management, or public health that can be measured (such as procedure time, healing rate, or reduction in complications).

If a device makes clinical claims, clinical data are required, and Article 61(10) is not applicable.

At the same time, claims may be explicit or implicit. For example, a device is implicitly expected to perform at least comparably to the current State of the Art (SOTA) alternative. If peer or benchmark products commonly make clinical claims under the prevailing SOTA, but you choose to make only technical claims, this may raise questions about whether the device conforms to SOTA.

4) Traceability of risk management and SOTA

The white paper requires key clinical risks to be identified from the SOTA discussion and traced into the risk management documentation. If you claim that the effectiveness of risk controls can be verified solely by non-clinical methods, you must provide sufficient justification and consider both SOTA and benefit-risk acceptability.

5) Clinical benefit: “direct vs. indirect” must be distinguished

The white paper identifies a common misconception: some manufacturers claim that a device has no clinical benefit and therefore does not require clinical data. However, under the GSPR, the device must demonstrate a favorable benefit-risk profile; without clinical benefit, conformity cannot be demonstrated. Only devices that provide indirect clinical benefit may have room to rely primarily on non-clinical data.

6) Device-human body interaction

The type and duration of contact, the tissue/anatomical site involved, and whether the interaction is novel must be assessed; if the interaction is novel, clinical data are usually required.

7) Acceptable reasons why “clinical data are not appropriate” (not appropriate ≠ insufficient/inadequate)

The white paper emphasizes that “not appropriate” is different from adequacy/sufficiency; Article 61(10) is an exception. Examples of more acceptable reasons include:

A clinical investigation would expose patients to unnecessary burden or risk, raising ethical concerns (for example, in emergency settings).
Meaningful, measurable, and significant clinical endpoints cannot be identified (for example, for some simple, long-used general-purpose devices with a well-established safety profile).
The device has no direct medical effect and cannot be evaluated in a patient clinical setting (for example, sample refrigeration equipment or sterilization equipment that depends on technical endpoints such as temperature control or microbial log reduction).

VI. Documentation requirements: using Article 61(10) does not mean “no CER required”

The white paper makes clear that even when Article 61(10) is used, a CEP, CER, and PMS/PMCF plan are still required. In addition, a clear and robust Article 61(10) justification document must be provided, covering each requirement of Article 61(10) point by point (risk management, clinical performance, claims, device-human body interaction, etc.), and it will be subject to Notified Body review; insufficient justification may lead to rejection of this route.

In the CER, the white paper requires:

A SOTA discussion to understand the clinical context, identify clinically relevant risks, and confirm whether there are any unknown or unaddressed clinical data sources (for the device itself or similar/equivalent devices).
Based on the SOTA, identify the safety and performance outcome measures that need to be supported by non-clinical testing and define acceptance criteria.
Summarize the non-clinical evidence: applicable common specifications, harmonized standards, product standards, as well as animal/phantom/cadaver studies, usability testing, etc.; compare the results against the predefined acceptance criteria and demonstrate that the benefit-risk profile is acceptable.

The white paper also highlights an "extremely rare" scenario: at the initial conformity assessment stage for a newly marketed device, clinical data may be unavailable. In very rare cases, Article 61(10) may be applied during the initial conformity assessment, with clinical data collected in the post-market phase and presented at recertification. However, this still requires a robust justification covering all elements of Article 61(10), and other clinical pathways, including equivalence or data from similar devices, must already have been considered.

7. The role of PMS/PMCF: Article 61(10) does not mean "no PMCF"

The conclusion of the white paper makes this explicit: Article 61(10) does not exempt a device from PMS/PMCF. For Article 61(10) devices, PMCF is usually limited to "general PMCF," such as literature screening and collection of user feedback. Its purposes include confirming continued alignment with the State of the Art, identifying new risks or verifying risk estimates, and confirming that no new clinical data have emerged for the device itself or for similar/equivalent devices.

At the same time, the white paper states that "specific PMCF" is generally not expected for Article 61(10) devices (for example, registry studies, clinical investigations, or patient-level surveys), because these activities are intended to answer specific clinical questions and collect clinical data, which would directly conflict with the position that clinical data are not appropriate.

8. Seven scenarios provided by the white paper: how BSI determines "acceptable" vs. "not acceptable"

The following only restates the example scenario conclusions in the white paper and the rationale it provides, to help readers understand BSI's review framework:

Scenario 1: Patient monitoring support software, an "electronic diary" app (Class IIa) (Conclusion: acceptable)

• The device function is to record, store, and display patient data, without providing treatment recommendations; the benefit is to support patient management (indirect benefit).

• There are no clinically relevant claims; the main risks are technical or use-related (data recording, account deletion, cybersecurity, etc.), which can be demonstrated through usability, software verification, and conformity with standards.

• The State of the Art assessment concludes that the performance outcomes are technical in nature, and no similar/equivalent devices were identified as requiring clinical data support.

Scenario 2: Dental imaging management / enhanced visualization software (Class IIa) (Conclusion: acceptable)

• It does not perform diagnosis or treatment decision-making, and only provides informational support; the benefit is to facilitate patient management (indirect benefit).

• The claims are technical (enhanced visualization, distance measurement) and can be supported by non-clinical testing; there are no clinical claims.

• Key clinical risks (such as unnecessary radiation exposure from repeat imaging, incorrect diagnostic information, and safety issues) are controlled through validated usability and technical testing measures; the State of the Art did not identify any need for clinical outcome measures.

Scenario 3: Peracetic acid high-level disinfectant (Class IIb) (Conclusion: acceptable)

• Used for device decontamination, with the indirect benefit of infection prevention; no "clinical endpoint" requiring clinical data support can be defined.

• The claim is broad-spectrum decontamination efficacy, which can be demonstrated through non-clinical data and standards covering microbiological effectiveness, toxicological safety, simulated use, material compatibility, and related evidence; there are no clinical claims.

• Safety concerns and systematic misuse / use beyond the intended scope are monitored through PMS/general PMCF.

Scenario 4: Vertebral augmentation cement delivery device (linked to a specific implant/procedure) (Class IIa) (Conclusion: not acceptable)

• Although the device claims only an indirect benefit, the State of the Art assessment concludes that it may affect procedural outcomes (such as cement leakage and postoperative fracture rate).

• Similar devices are typically supported by a combination of procedure-related indirect clinical evidence and non-clinical testing to demonstrate safety and performance.

• In this example, the manufacturer planned specific PMCF (surveys/studies), which contradicts the position that clinical data are not appropriate; BSI recommends using the cumulative evidence pathway in MDCG 2020-6 Appendix III instead.

Scenario 5: Biopsy needle (Class IIa) (Conclusion: not acceptable)

• The device is a surgically invasive device and, although serious clinical risks are uncommon, they do exist (bleeding, infection, pneumothorax, etc.), and the device forms part of a high-risk procedure.

• Although some endpoints can be assessed using non-clinical methods, the purpose of the device is to obtain samples from different anatomical structures, and the transferability of preclinical environments is limited (differences between living and non-living tissue affect performance).

• In this example, clinical data from regions outside the EU were already available for initial conformity assessment, and the manufacturer also intended to collect post-market user feedback/registry data, indicating that clinical data are appropriate and can be collected; Article 61(10) is therefore excluded.

Scenario 6: Microprocessor-controlled injection system (Class IIb) (Conclusion: not acceptable)

• The State of the Art shows that comparable devices typically make clinical claims that are supported by clinical data (for example, performance related to imaging diagnostics).

• If the assessed device does not make these clinical claims, it would be unable to demonstrate that it achieves or exceeds the State of the Art; this pathway is therefore not accepted.

Scenario 7: 0.9% saline for wound/skin care (Class IIb) (Conclusion: not acceptable)

• The State of the Art discussion identifies clinically relevant endpoints (such as infection rate, wound healing, mucociliary clearance time, symptom improvement, quality of life, etc.), showing that clinical data are, in principle, appropriate.

• Although conducting clinical investigations for this mature technology may be impractical or not resource-justified, the example manufacturer did not adequately consider other sources of clinical data; the cumulative evidence pathway in MDCG 2020-6 Appendix III is more appropriate.

9. A "directly usable decision checklist" distilled from BSI's position

If you want to self-assess, using the logic of the white paper, whether Article 61(10) may be claimed, you at least need to answer each of the following:

Is the device Class IIa or Class IIb and non-implantable? (Class III/implantable devices are excluded outright.)
Are there any explicit or implicit clinical claims? (If there are clinical claims, it is excluded.)
Does the device provide an indirect clinical benefit or a direct clinical benefit? (A direct benefit usually requires clinical data.)
Could the device affect patient outcomes within the system/procedure? The greater the impact, the more relevant/appropriate clinical or indirect clinical data become.
Does the State of the Art show that comparable devices generally use clinical data to support key performance? If so, making only technical claims may be challenged as not aligned with the State of the Art.
Is the device-human interaction novel? Does the type/duration/site of contact involve high-risk areas? (Novel interactions often require clinical data.)
Can you provide a legitimate justification for why clinical data are not appropriate (ethical reasons, inability to define meaningful endpoints, no direct medical effect, etc.), rather than simply stating that data are unavailable or difficult to collect?
Are you prepared with a CEP, CER, PMS/PMCF, and an Article 61(10) justification document that addresses each required element point by point?
Have you avoided planning "specific PMCF" clinical data collection activities? (Otherwise the position becomes contradictory.)

10. Conclusion: the "one core message" BSI conveys through this white paper

Throughout the paper, BSI repeatedly emphasizes that the default starting point for clinical evaluation under the MDR is Article 61(1)—clinical data are required to generate sufficient clinical evidence; Article 61(10) is only an exception pathway. To use Article 61(10), you must be able to demonstrate, in a logically consistent clinical evaluation chain, that clinical data are not appropriate, and build a reviewable, traceable, documented justification based on factors such as the State of the Art, risk management, claims, type of benefit, and device-human interaction. This logic and philosophy are highly aligned with Clinsota's practical approach, and they also help some companies save substantial time and cost through a rigorous closed-loop evidence-based medicine strategy.

Deep Dive into the BSI White Paper: MDR Article 61(10)—How to Complete a Clinical Evaluation Using Non-Clinical Data in “Exceptional Circumstances”