Clinical Evaluation Report#

The Clinical Evaluation Report states the clinical benefits and safety characteristics of the device, based on clinical data. It is the output of the Clinical Evaluation Plan.

While the content of the Clinical Evaluation is simple, writing it, coming up with the right structure and forming a sensible line of reasoning (equivalence) can be a bit tricky.

These are the guidance documents on Clinical Evaluation. If you’re the person writing it, you should read them:

MDCG 2020-1, 2020-5, 2020-6

MDCG 2020-13: Quite helpful as it gives you an idea of the structure.

MEDDEV 2.7.1 rev. 4. (mostly for MDD, but still a good starting point; especially the list of proposed headings for a report at the end of the document).

Finally, this Clinical Evaluation Report is for Medical Devices. If you are an IVD manufacturer you will write a Performance Evaluation Report which has many similarities but some essential parts are different, e.g., Post-Market Performance Follow-Up and procedural descriptions of evaluation and validation).

Product#

Name: <product name>
Version: <product version>
Basic UDI-DI: <insert UDI-DI, if/when available>

Table of Contents#

A list of the sections is presented below. You have to update this manually when you edit headings as this is a markdown file (sorry - guilty dog face).

List of Abbreviations
Product
Relevant Documents
Scope of the Clinical Evaluation
Device
Clinical Background, Current Knowledge, State of the Art
Type of Evaluation
Equivalence
Literature Search
Clinical Data
Post-Market Activities
Conclusions
Date of the Next Clinical Evaluation
Dates and Signatures
Qualification of the Responsible Evaluators
References

1. List of Abbreviations#

Abbreviation	Explanation
AE	Adverse Event
BfArM	German Federal Institute for Drugs and Medical Devices
CE	Communauté Européenne
CER	Clinical Evaluation Report
CI	Confidence interval
DHF	Design History File
DMR	Device Master Record
EUDAMED	European Database for MEdical Devices
FDA	Food and Drug Administration (US)
IFU	Instruction for Use
LoE	Level of Evidence
MAUDE	Manufacturer and User Facility Device Experience
MEDDEV	Guideline for Clinical Evaluation of Medical Devices
MDD	Medical Device Directive, European Directive 93/42/EEC
MDR	Medical Device Directive, 2017/745
PMS	Post Market Surveillance
PMCF	Post MArket clinical Follow Up
Rev	Revision
SOP	Standard Operation Procedure
T	Tendency
C	Comparability
WHO	World Health Organization

2. Product#

Name: <product name>
Version: <product version>
Basic UDI-DI: <insert UDI-DI, if/when available>
UMDNS-Code:
GMDN-Code:

The classification is based on the following criteria:

Select one of these two, based on whether you’re going for MDD or MDR compliance.

Annex IX of the European Directive 93/42/EEC (MDD): <if applicable>
According to the EU Regulation 2017/745 (MDR) Annex VIII Rules: <if applicable>.

3. Relevant Regulatory References#

4. Scope of the Clinical Evaluation#

Note: This section is copy-pasted from the Clinical Evaluation Plan.

The following section can be part of the Clinical Evaluation Plan or pasted here with reference to the respective chapters. Remember that you can update the Clinical Evaluation Plan during your evaluation (e.g., your search criteria are not sufficient).

The approach according to the MEDDEV 2.7/1 rev. 4 includes five logical procedural stages in order to evaluate the performance and safety data of the medical device:

Step 0 (“Scope”)

Step 1 (“Identification”)

Step 2 (“Appraisal”)

Step 3 (“Analysis”)

Step 4 (“Report”)

During the working process, these five steps are iterative and influence each other. The report shows the steps in a sequential way. In the present report (step 4), the steps 0 through 3 are corresponding to the following chapters:

List sections of the Clinical Evaluation Report here which cover the steps above.

5. Device#

5.1 Device Description#

Copy-paste your Device Description (which includes the Intended Use) here. If it’s not done yet, remember to do it later :)

Also, reference your Essential Requirements Document here.

5.2 Clinical Benefits, Outcome Parameters#

Medical device claims are statements from accompanying documents, marketing material, your website, etc. that include information about the performance and safety of the medical device (information by the manufacturer). The clinical evaluation is done in order to determine whether those claims are confirmed by sufficient clinical evidence.

So, if your website claims that your device cures back pain in 50% of patients after 14 days, here’s the place to list that claim and show explain how you’ll prove it.

Claims that have not been stated so far (e.g., in the IFU) are described below and sorted according to their meaning for performance and safety. In this clinical evaluation, it was determined that the claims are sufficiently supported by clinical data.

Performance-related product claims:

Claim 1
Claim 2

5.3 Clinical Safety, Methods for Analysis#

Describe your safety parameters, i.e., which things should your product fulfil so that you consider it safe? And your methods, i.e., how will you prove that your product fulfils those safety parameters? A method could be a literature search for past studies, but you could additionally do a Post-Market Clinical Follow-Up to double-check whether that’s actually true for your device.

Safety-related product claims:

Claim 1
Claim 2

5.4 Acceptability of Benefit-Risk-Ratio#

After you’ve defined your benefits and safety parameters, which combination of those is acceptable to you? In the case of most software devices (and apps), you’ll probably have subtle benefits (e.g., better disease management, early detection of relapses) while low safety concerns (e.g., disease progression unlikely, not killing anyone).

6. Clinical Background, Current Knowledge, State of the Art#

This chapter focus on literature and guidelines that describe the current state of the art and other topics. It is similar to the literature / introduction chapter of papers.

It makes sense to differentiate between “context” and “pivotal” data:

Context data describes the state of the art (commonly the introduction / literature part of papers)

Pivotal data is used for the appraisal, i.e., that’s the data describing the actual study and outcome(s). In the best case, the pivotal data is about the actual device you’re claiming equivalence to.

6.1 Clinical Background & Current Knowledge#

Describe the clinical context of the disease you’re treating: How are patients currently treated? Which symptoms do they have, which diagnostic modalities are being used to establish a diagnosis, which treatment options exist? What are the benefits and drawbacks of current treatment options?

6.2 State of the Art incl. Alternative Treatments#

Given the current treatment options, what is the preferred, “state of the art” treatment? What are its benefits and drawbacks? Are there recent scientific achievements (studies, new technologies, software) which may be promising to improve this state-of-the-art treatment? Also, what are alternative treatments?

7. Type of Evaluation#

Note: This section is copy-pasted from the Clinical Evaluation Plan.

8. Literature Search#

8.1 Literature Search Methods#

Copy-paste from Clinical Evaluation Plan.

8.2 Literature Appraisal Criteria#

Copy-paste from Clinical Evaluation Plan.

You can describe it as a flow chart, e.g.:

Initial search with all publications from the relevant databases. 2.1 Potentially relevant papers (from the first skimming of article titles / abstracts) 2.2 Irrelevant papers

Potentially relevant papers found in irrelevant papers 4.1 Relevant paper used for context 4.2 Relevant paper used for a single appraisal (pivotal data)

8.3 Literature Search Protocol#

A table which lists your actual literature search results. For each entry, you should decide whether it’s acceptable (based on your appraisal criteria) or not. I pre-wrote some tables to give you an idea of the structure below. You could separate the tables based on the database where you did the search (PubMed, Google Scholar).

It could make sense to put this in a separate document (rather: a spreadsheet).

Here are some bullet points from the guidance:

Literature search protocol provided

Literature search reports provided

Full list of retrieved articles provided

Full list of excluded articles provided, with reasons for exclusion Full text copies of relevant documents available

8.4 Database Search Overview#

Database	Search term	# Hits	# Evaluated Abstracts	# Potential Relevant Publications

8.5 Database: PubMed#

Title	Author	Year	Summary	Relevant? Why?
				e.g., similar design, similar features

8.6 Database: Google Scholar#

Title	Author	Year	Summary	Relevant? Why?

8.7 Database: Cochrane#

Title	Author	Year	Summary	Relevant? Why?

9.8 Literature Search Report#

Briefly summarize how many studies you reviewed, how many you deemed acceptable and why you didn’t include the unacceptable ones (probably because they didn’t conform to the appraisal criteria).

Describe a gap analysis which information could not be found during your search e.g., specific functionalities of your device, limited number of clinical data in publications.

8.10 Evaluation of study Quality#

8.10.1 Level of Evidence (LoE)#

Take a sufficient classification to evaluate the quality of your study. You can use the LoE from MDCG 2020-6 or from the American Heart Association.

8.10.2 Tendency (T)#

The MEDDEV 2.7/1 rev. 4 requires that literature is used that confirms as well as questions the suitability of the evaluated medical device.

Positive: + (Confirms clinical suitability)

Negative: - (Does not confirm clinical suitability)

Indifferent: i (No statement to the clinical suitability possible)

8.10.3 Comparability (C)#

For the consideration of the relevance of literature data for the clinical evaluation it is necessary to make a statement about the comparability to the medical device to be evaluated. This statement about the comparability is made according to the described criteria from chapter “Equivalence” above.

8.10.4 Single Appraisal of Searched Clinical Studies#

This chapter is highly important for medical device evaluation. Here you take all the publications from the literature search and evaluate them to prove equivalence with your device and prove the respective product claims. The latter can also be proven by other activities (e.g., PMS, equivalent device, usability testing, result of adverse events etc.).

8.10.5 Literature Data on the Performance of the Medical Device#

Focus here on equivalent devices that support your performance claims.

Do the paper description in the same way e.g.:

First, describe the study with its main characteristics regarding the quality and relevance are made. (Study design, patient number, mean age inclusion and exclusion criteria, used product, endpoint, results of performance, conclusion of the author).

Complications, side-effects, adverse events.

A consideration of the study according to the criteria. (LoE, T, C)

In the following, the selected literature references, which were categorized as “relevant” (appraisal), are evaluated. This was done according to the following scheme:

8.10.6 Literature Data about the Safety of the Medical Device#

Focus here on equivalent devices that support your Safety claims. So you do not need to describe the LoE since potential risks can also arise from case studies (which might have a low LoE).

9. Clinical Data#

9.1 Clinical Data From Literature#

List all the clinical data you got from studies which matched your appraisal criteria.

9.2 Clinical Data from Clinical Study Databases#

List all the clinical data you got from studies (clinical trials.gov, ANCTR, DRKS, WHO etc.).

9.3 Clinical Data From Adverse Event Databases#

List all the clinical data you got from studies which matched your appraisal criteria (BfArM, MAUDE, FDA Medical Device Recall, EUDAMED (when applicable)).

9.4 Summary and Appraisal of Clinical Data#

Summarize all the clinical data from above :)

9.5 Analysis of the Clinical Data#

Analyze the clinical data with a focus on whether your targets of clinical benefits and safety were fulfilled.

10. Equivalence#

NOTE: first of all, under the European MDR, equivalence data is not as significant anymore as it used to be under the MDD. That’s because the MDR now actually requires you to have access to the clinical data that you are referencing to support your own device’s safety and performance. Access to the data of competitor devices is usually rather tricky though. That’s why we’re using equivalence in our template only as a another indicator on the side.

It’s important though to understand why you’re conducting this equivalent device comparison: you want to show that another device is very safe and has a reliable performance. You then refer to those datapoints and, after demonstrating that your device is equivalent to the other, argue that those datapoints also support your device profile. That means:

Make sure to include search terms for the equivalent device(s), such as the brand name, in your literature search (para. 8)

Make sure to include search terms for the equivalent device(s), such as the brand name, in your literature search (para. 9)

You don’t really analyze the resulting information in the above paragraphs, but draw your main conclusion from this below in para. 10.2.4.

10.1 Equivalent Device#

Describe the equivalent device you’re comparing yourself to, mainly its Intended Use.

10.2 Demonstration of Equivalence (Technical, Biological, Clinical)#

Here you have to demonstrate that your device is equivalent to the Equivalent Device. You accomplish that by creating a table in which you list certain characteristics, and describe those characteristics both for the Equivalent Device and for your device. The idea is that your device is mostly the same in most characteristics.

The tables are split into general stuff (first table), then Clinical, Technical and Biological equivalence as per the guidance documents. I pre-filled some of the table rows for you as they should be universal, e.g., Intended Use, Medical Indication(s) and Programming Language. But definitely feel free to add additional rows which are useful for comparing your device to the equivalent one. Maybe you’re using recurrent neural networks and the equivalent device is, too? Then add that.

	Equivalent Device	This Device
Intended Use
Medical Indication(s)
Device Classification
Principle of Operation	Stand-alone Software	Stand-alone Software

According to standard methodology, the level of similarity or equivalence of a comparable medical device to the evaluated medical device is divided into the three categories “clinical”, “technical” and “biological”.

Within each category there are four steps for the evaluation of the equivalence level:

Equivalence Level	Score
Very Equivalent	3
Nearly Equivalent	2
In some aspects similar	1
In some aspects similar	0

After a score has been assigned to all three categories (clinical, technical, biological), those scores are summed up. The total equivalence is determined based on the sum of those values:

Sum	Equivalence
8-9	Very Equivalent
6-7	Nearly Equivalent
3-5	In some aspects similar
0-2	Very different

For software devices, adjust the score correspondingly. For example, you could say: “As the biological category typically does not apply for software devices, the total equivalence is determined based on the sum of those values as follows” and use an adjusted score: 6 // 5 // 3-4 // 0-2.

Clinical Equivalence#

The device is considered clinically equivalent if it is:

Used for the same clinical condition
Used for the same intended use
Used at the same site in the body
Used in a similar population
Not foreseen to deliver significantly different performances
Similar performances such as the expected clinical effect, the specific intended purpose, the duration of use, etc.

Technical Equivalence#

The device is considered technically equivalent if it:

Is of a similar design
Is used under the same conditions of use
Has a similar specifications and properties
Has a similar intensity of energy, tensile strength, viscosity, surface characteristics, wavelength
Has a similar surface texture, porosity, particle size, nanotechnology, specific mass, atomic inclusions
Uses a similar deployment methods
Has similar principles of operation and critical performance requirements

Biological Equivalence#

For stand-alone software:

Not applicable. The device doesn’t come in contact with human tissue or body fluids.

For medical devices which are a part of a hardware medical device:

Note: Medical device monitors are also medical devices according to 60601-1.

The device is considered biologically equivalent if it:

Is similar or has similar effecting materials which are in contact with body tissue and fluids

Optional: Effectiveness Equivalence#

The device is considered to have equivalent effectiveness, if it:

Has similar Sensitivity, Specificity, Accuracy, NNP, PPV, etc.

10.2.1 Clinical Equivalence#

Note: The MDR doesn’t explicitly state that the device needs to be used for the same medical indication, gender and duration of use. But it should be used for the same clinical condition or purpose including similar severity and stage of the disease.

	Equivalent Device	This Device
Clinical Condition
Disease Stage
Site in Body
Population: Age, Anatomy, Physiology
Clinical Effect
Duration of Use
Significant Performance Difference

Clinical Equivalence Score: <enter number between 0 and 3>

10.2.2 Technical Equivalence#

	Equivalent Device	This Device
Software Algorithm
Programming Language
Graphical User Interface (GUI)
Web-based Application
Inputs
Outputs

The following table can be added if it is a hardware-related Medical Device:

	Equivalent Device	This Device
Design
Dimension
Operating Condition
Physical Characteristics (e.g., weight)
Data transmission

Technical Equivalence Score: <enter number between 0 and 3>

10.2.3 Biological Equivalence#