Bench testing, biocompatibility evaluation, mechanical and electrical safety testing, software verification, and usability engineering are non-clinical evidence. Under Annex XIV Part A of Regulation (EU) 2017/745, they are not clinical data in the Article 61(1) sense, but they are a mandatory part of the evidence package that supports the CER. Article 61(10) allows, for a narrow set of devices, that a demonstration of conformity based on clinical data is not appropriate. Provided the justification rests on risk management output and on the device-body interaction, clinical performances, and manufacturer's claims. For every other device, a strong pre-clinical package directly shrinks the clinical data gap the CER has to close.

By Tibor Zechmeister and Felix Lenhard. Last updated 10 April 2026.

TL;DR

  • Non-clinical evidence under the MDR covers bench testing, biocompatibility under EN ISO 10993-1:2025, mechanical and electrical safety (EN 60601-1), software verification under EN 62304:2006+A1:2015, usability under EN 62366-1:2015+A1:2020, and the risk management file under EN ISO 14971:2019+A11:2021.
  • Article 61(10) of Regulation (EU) 2017/745 allows, for certain devices, that a demonstration of conformity based on clinical data is not deemed appropriate, with adequate justification based on the risk management output and the specifics of the device-body interaction, the intended clinical performances, and the manufacturer's claims.
  • Article 61(10) is a narrow carve-out, most often relied on for categories such as sutures, staples, dental fillings, crowns, screws, plates, wires, pins, clips, and connectors where state of the art is well established.
  • For every other device, non-clinical evidence supports. Never replaces. The clinical evaluation. It feeds the identification, appraisal, and analysis stages of Annex XIV Part A and reduces the scope of new clinical data required.
  • A strong pre-clinical package lets the CER close the clinical gap with the smallest defensible clinical investigation, non-interventional study, or literature-plus-equivalence package.

Why the non-clinical evidence question matters for your startup

Felix has watched several founders treat bench testing and the CER as two disconnected workstreams. The regulatory team runs the biocompatibility study, the electrical safety lab tests against EN 60601-1, the software team signs off verification reports, and somebody else writes a clinical evaluation report that barely references any of it. Then the Notified Body review comes back asking why the CER does not draw on the evidence the company already has in a drawer.

This is a pure subtraction problem. The pre-clinical work has already been done. The CER is already being written. Connecting the two saves money and strengthens the package. The MDR treats the CER as the analytical layer that sits on top of all the evidence the manufacturer has generated, clinical and non-clinical alike.

The non-clinical evidence question. What the MDR text actually says

The starting point is Article 61(1) of Regulation (EU) 2017/745, which defines clinical data and sets the requirement that conformity with the general safety and performance requirements under the normal conditions of intended use shall be based on clinical data providing sufficient clinical evidence. Annex XIV Part A then describes the clinical evaluation procedure in four stages: identification of available clinical data, appraisal of suitability and quality, analysis of whether conformity is demonstrated, and planning of additional data generation where gaps remain.

Non-clinical data enters this procedure at every stage. Bench testing, biocompatibility, mechanical performance, electrical safety, software verification, and usability engineering all produce evidence the clinical evaluation has to consider when it appraises whether the available clinical data is sufficient. The evidence either closes gaps directly or defines the clinical questions that still need clinical data to answer.

The carve-out sits in Article 61(10):

"Where the demonstration of conformity with general safety and performance requirements based on clinical data is not deemed appropriate, adequate justification for any such exception shall be based on the result of the manufacturer's risk management and on consideration of the specifics of the interaction between the device and the human body, the clinical performances intended and the claims of the manufacturer...". Regulation (EU) 2017/745, Article 61, paragraph 10.

Article 61(10) does not exempt the manufacturer from doing a clinical evaluation. It allows, for specific devices, that the evaluation may conclude that clinical data in the Article 61(1) sense is not appropriate and that the demonstration of conformity rests predominantly on non-clinical evidence plus the risk management output. The manufacturer must substantiate the justification in the Annex II technical documentation.

This is a narrow route, most often relied on for categories where state of the art is well established and the device-body interaction is fully characterised by non-clinical testing. Sutures, staples, dental fillings, crowns, screws, plates, wires, pins, clips, and connectors. For a novel or higher-risk device, Article 61(10) is rarely the right path. For every device not relying on it, non-clinical evidence does not replace clinical data. It supports it.

What counts as non-clinical evidence in the CER

Non-clinical evidence for the purposes of the clinical evaluation includes, at minimum:

  • Biocompatibility evaluation under EN ISO 10993-1:2025, mapped to MDR Annex I Section 10 (chemical, physical and biological properties). The biological evaluation plan, the chemical characterisation, and the endpoint-by-endpoint rationale are direct inputs to the clinical evaluation's appraisal of residual risks.
  • Mechanical and physical performance testing. Dimensional verification, fatigue, wear, burst pressure, tensile and compressive strength, thermal and environmental conditioning, whatever is relevant to the device. Not clinical data, but evidence that the device meets its performance specifications under simulated use.
  • Electrical safety and EMC testing for medical electrical equipment under EN 60601-1 and EN 60601-1-2, including basic safety, essential performance, and electromagnetic disturbance immunity.
  • Software verification and validation under EN 62304:2006+A1:2015, including unit tests, integration tests, system tests, and anomaly lists. The software safety classification (Class A, B, or C) and the verification evidence feed the risk management file and the CER.
  • Usability engineering under EN 62366-1:2015+A1:2020, including the use specification, formative and summative evaluations, and the usability engineering file. Post 146 covers how usability data counts as clinical evidence for safety-in-use claims.
  • Risk management file under EN ISO 14971:2019+A11:2021. The risk analysis, risk control measures, residual risk evaluation, and risk management report. This is the backbone tying the pre-clinical evidence to the clinical claims.
  • Animal studies where ethically and scientifically justified, and simulated-use testing on cadaveric tissue, tissue surrogates, or benchtop models.

The clinical evaluation plan names these sources. The CER analyses them. The linkage to specific general safety and performance requirements in Annex I is what the Notified Body looks for during review.

A worked example. A low-risk device CER built mostly on bench

Consider a Class IIa non-invasive handheld vital-signs monitor for home care. The clinical claims are measurement accuracy within a specified tolerance, reliable alarm behaviour at the threshold, and safe use by a lay user.

The pre-clinical package includes biocompatibility for the patient-contacting surface under EN ISO 10993-1:2025, electrical safety and EMC under EN 60601-1 and EN 60601-1-2, software verification under EN 62304 with the appropriate safety class, a usability engineering file under EN 62366-1:2015+A1:2020 with a summative evaluation on representative lay users, and mechanical durability testing. The risk management file under EN ISO 14971:2019+A11:2021 ties it together.

A well-constructed CER draws heavily on this stack. The accuracy claim is supported by bench testing against a reference method, with literature on comparable devices establishing state of the art. The alarm reliability claim is supported by software verification, the risk analysis of alarm failure modes, and the usability summative. The lay-user safety claim is supported by the usability engineering file and by post-market clinical follow-up under Annex XIV Part B. Equivalence to a legacy device under MDCG 2020-5 provides the remaining clinical backdrop.

The residual clinical gap may come down to a small non-interventional performance study at one clinical partner site, rather than a full interventional trial. This is not Article 61(10). The device still has clinical evidence. The clinical scope is small because the pre-clinical scope is strong.

The evidence-stack playbook for a resource-constrained startup

A disciplined startup builds the evidence stack in this order, with the CER planned against the stack from day one:

  1. Lock the intended purpose and the clinical claims. Every downstream decision derives from the claims the manufacturer intends to make.
  2. Map the general safety and performance requirements in Annex I to the device. Each applicable Annex I section becomes a line item in the evidence matrix. Biocompatibility maps to Section 10, software to Section 17, usability to Sections 5 and 22, electrical safety to Section 17.
  3. Plan the pre-clinical testing against the evidence matrix, not against generic templates. Run the tests the matrix requires. Do not run extras. Do not skip requirements.
  4. Build the risk management file as the connective tissue. Every residual risk is a question the clinical evaluation has to answer. The risk management file and the clinical evaluation plan are developed in parallel.
  5. Write the clinical evaluation plan knowing what the pre-clinical stack will deliver. Identify which claims are answered by bench, biocompatibility, software, usability, literature, or equivalence. And which remain for new clinical data.
  6. Write the CER as the analytical layer. It draws on the pre-clinical package and the clinical data together, appraises each source against the Annex XIV Part A suitability and quality criteria, and concludes whether the general safety and performance requirements are demonstrated. Where Article 61(10) applies, the CER includes the substantiation required by that paragraph.

The subtraction is in the sequencing. A startup that lets the CER trail the pre-clinical work by six months writes the CER twice. A startup that plans the two in parallel writes it once.

Reality Check. Where do you stand?

  1. Have you mapped every applicable Annex I general safety and performance requirement to a specific non-clinical evidence source?
  2. Is your biocompatibility evaluation under EN ISO 10993-1:2025, with a biological evaluation plan explicit about which endpoints the risk analysis requires?
  3. Does your risk management file under EN ISO 14971:2019+A11:2021 trace every residual risk to a control measure or an evidence source that justifies acceptance?
  4. Does your clinical evaluation plan explicitly list which claims are answered by non-clinical evidence and which still require clinical data?
  5. If you are relying on Article 61(10), have you documented why clinical data in the Article 61(1) sense is not appropriate for your specific device?
  6. Is the CER being written in parallel with the pre-clinical stack, or only after the pre-clinical work is "done"?
  7. Have you confirmed the current edition of every harmonised standard you cite against the list giving presumption of conformity to the MDR?

Frequently Asked Questions

Can bench testing replace clinical data in an MDR CER? For most devices, no. Bench testing supports the clinical evaluation but does not replace clinical data in the Article 61(1) sense. The exception is Article 61(10), which allows the clinical evaluation of certain devices to conclude that a demonstration of conformity based on clinical data is not appropriate, provided the justification rests on risk management output and on the specifics of the device-body interaction, the intended clinical performances, and the manufacturer's claims. The Notified Body decides whether the justification holds.

What is the difference between non-clinical evidence and pre-clinical data? The two terms overlap and are often used interchangeably. "Pre-clinical" describes data generated before clinical use. Bench testing, biocompatibility, animal studies where justified, simulated use, and verification testing. "Non-clinical" describes any data that is not clinical data in the Article 61(1) sense. Annex XIV Part A treats them as inputs to the clinical evaluation regardless of the label.

Is biocompatibility testing clinical evidence? No. Biocompatibility under EN ISO 10993-1:2025 is non-clinical evidence. It is mandatory for devices with patient contact because MDR Annex I Section 10 requires it, and it feeds the clinical evaluation's analysis of biological safety. But it does not describe use of the device on human subjects, so it is not clinical data in the Article 61(1) sense.

What does Article 61(10) actually require? Article 61(10) requires that where a demonstration of conformity based on clinical data is not deemed appropriate, adequate justification shall be based on the risk management output and on consideration of the specifics of the interaction between the device and the human body, the clinical performances intended, and the manufacturer's claims. The justification must be substantiated in the Annex II technical documentation. Article 61(10) does not exempt the device from having a clinical evaluation.

How does the risk management file connect the pre-clinical package to the CER? Every residual risk identified in the risk management file under EN ISO 14971:2019+A11:2021 is a question the clinical evaluation has to consider: either the pre-clinical evidence closes it, or clinical data has to. The CER cross-references the risk management output wherever a residual risk is appraised. Without that connection, the CER reads as a separate document, and the Notified Body treats it as one.

Sources

  1. Regulation (EU) 2017/745 on medical devices, Article 61, Article 61(10), Annex I (Sections 5, 10, 14, 17, 22), Annex II, Annex XIV Part A. OJ L 117, 5.5.2017.
  2. MDCG 2020-5. Clinical Evaluation. Equivalence, April 2020.
  3. EN ISO 10993-1:2025. Biological evaluation of medical devices. Part 1.
  4. EN ISO 14971:2019+A11:2021. Application of risk management to medical devices.
  5. EN 60601-1 (IEC 60601-1:2005+A1:2012+A2:2020, with EN amendments including A13:2024). Medical electrical equipment. General requirements for basic safety and essential performance.
  6. EN 62304:2006+A1:2015. Medical device software. Software life-cycle processes.
  7. EN 62366-1:2015+A1:2020. Application of usability engineering to medical devices.

This post is part of the Clinical Evaluation & Clinical Investigations series in the Subtract to Ship: MDR blog. Authored by Felix Lenhard and Tibor Zechmeister. If your CER and your pre-clinical stack are being built on separate tracks and you suspect the clinical evidence scope is bigger than the claims actually require, Zechmeister Strategic Solutions works with founders on integrating the two workstreams so the evidence package answers the claims at the smallest defensible scope.