Regulatory Open Forum

Hi all,

I am interested in getting your feedback on the interpretation of Article 88 on Trend reporting.

Being said that EU MDR 2017/745 defines:
Serius incident as:
any incident that directly or indirectly led, might have led or might lead to any of the following:
(a) the death of a patient, user or other person
(b) the temporary or permanent serious deterioration of a patient's, user's or other person's state of health
(c) a serious public health threat
Incident as any malfunction or deterioration in the characteristics or performance of a device made available on the market, including use-error due to ergonomic features, as well as any inadequacy in the information supplied by the manufacturer and any undesirable side-effect

Articles 88 requires that manufacturers have to report any statistically significant increase in the frequency or severity of incidents:
- that are not serious incidents (that in my interpretation are just incidents = complaints according to ISO 13485) or
- that are expected undesirable side-effects
that could have a significant impact on the benefit-risk analysis and which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits

Question 1)
Do you agree that the definition of Incident is very similar to the definition of Complaint according to ISO 13485?

Question 2)
According to article 88, it seems that manufacturers have to perform trend reporting not on all complaints or all expected undesirable side-effects (euse) but just on a subassembly that is defined as complaints/euse that:
- could have a significant impact on the benefit-risk analysis AND which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits

I am asking this since it is a common assumption to consider trend reporting applicable to all complaints and to each associated risk. Instead, it seems that trend reporting is just for risks that are unacceptable when weighed against the intended benefits.
Do you agree with this interpretation?


------------------------------
Andrea Duca
Roma
Italy
------------------------------

Question #1
The definition of complaint and incident are not very similar. A complaint is much broader. Every incident is a complaint, but not the other way around. For example, a package with a label that has the wrong size is a complaint, but not an incident.

Question #2
Article 88 applies to medical devices, not to components, subassemblies, etc. Also, distinguish between trend analysis and trend report. Trend analysis determines if there is a signal, while trend reporting informs the regulators of the signal. One could conduct trend analysis, not generate a signal, and therefore, never report.

Start by classifying each complaint (EN ISO 13485:2016/A11:2021) as an incident (EU-MDR) or not. Classify each incident as serious or non-serious. Report serious incidents following Art. 86. Analyze non-serious incidents following the Article III, PMS Plan.

In Annex III, the PMS plan sets up trend analysis (or most of it) by defining the method to detect a signal and the observation. It doesn't require the frequency at which perform trend analysis, but this is important. As a default, I recommend conducting trend analysis once each quarter with an observation period of four quarters. Determine the non-serious incident rate, plot it over the observation period, and calculate the slope of the linear regression line. If the slope is positive and statistically different from zero, then report under Article 88.

Setting up the test for statistical significance in Excel is easy. Plot the rate against time and tell Excel to show the linear regression line and the equation on the chart. Use the Regression tool in the Analysis Tool Pack to set up the statistical test. It is easy to do and should take less than thirty minutes each time you run the trend analysis.

------------------------------
Dan O'Leary CQA, CQE
Swanzey NH
United States
------------------------------

Original Message:
Sent: 15-Mar-2022 09:59
From: Andrea Duca
Subject: Intepretation of Article 88 Trend Reporting EU MDR

Hi all,

I am interested in getting your feedback on the interpretation of Article 88 on Trend reporting.

Being said that EU MDR 2017/745 defines:
Serius incident as:
any incident that directly or indirectly led, might have led or might lead to any of the following:
(a) the death of a patient, user or other person
(b) the temporary or permanent serious deterioration of a patient's, user's or other person's state of health
(c) a serious public health threat
Incident as any malfunction or deterioration in the characteristics or performance of a device made available on the market, including use-error due to ergonomic features, as well as any inadequacy in the information supplied by the manufacturer and any undesirable side-effect

Articles 88 requires that manufacturers have to report any statistically significant increase in the frequency or severity of incidents:
- that are not serious incidents (that in my interpretation are just incidents = complaints according to ISO 13485) or
- that are expected undesirable side-effects
that could have a significant impact on the benefit-risk analysis and which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits

Question 1)
Do you agree that the definition of Incident is very similar to the definition of Complaint according to ISO 13485?

Question 2)
According to article 88, it seems that manufacturers have to perform trend reporting not on all complaints or all expected undesirable side-effects (euse) but just on a subassembly that is defined as complaints/euse that:
- could have a significant impact on the benefit-risk analysis AND which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits

I am asking this since it is a common assumption to consider trend reporting applicable to all complaints and to each associated risk. Instead, it seems that trend reporting is just for risks that are unacceptable when weighed against the intended benefits.
Do you agree with this interpretation?


------------------------------
Andrea Duca
Roma
Italy
------------------------------

Thank your for the promt reply, Dan.

I wrongly used the term "subassembly" in the place of "subset".

So, do you confirm that medical device manufacturers selling their devices in the EU are only required to evaluate trends for incidents that:
- could have a significant impact on the benefit-risk analysis and
- which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits

I give lot of importance to the "and" that is placed between the two bullet points. So do you?
This would really restrict the scope of the trend analysis to just a subset of incidents (i.e. those that satisfy both bullets points)

------------------------------
Andrea Duca
Roma
Italy
------------------------------

Original Message:
Sent: 15-Mar-2022 10:34
From: Dan O'Leary
Subject: Intepretation of Article 88 Trend Reporting EU MDR

Question #1
The definition of complaint and incident are not very similar. A complaint is much broader. Every incident is a complaint, but not the other way around. For example, a package with a label that has the wrong size is a complaint, but not an incident.

Question #2
Article 88 applies to medical devices, not to components, subassemblies, etc. Also, distinguish between trend analysis and trend report. Trend analysis determines if there is a signal, while trend reporting informs the regulators of the signal. One could conduct trend analysis, not generate a signal, and therefore, never report.

Start by classifying each complaint (EN ISO 13485:2016/A11:2021) as an incident (EU-MDR) or not. Classify each incident as serious or non-serious. Report serious incidents following Art. 86. Analyze non-serious incidents following the Article III, PMS Plan.

In Annex III, the PMS plan sets up trend analysis (or most of it) by defining the method to detect a signal and the observation. It doesn't require the frequency at which perform trend analysis, but this is important. As a default, I recommend conducting trend analysis once each quarter with an observation period of four quarters. Determine the non-serious incident rate, plot it over the observation period, and calculate the slope of the linear regression line. If the slope is positive and statistically different from zero, then report under Article 88.

Setting up the test for statistical significance in Excel is easy. Plot the rate against time and tell Excel to show the linear regression line and the equation on the chart. Use the Regression tool in the Analysis Tool Pack to set up the statistical test. It is easy to do and should take less than thirty minutes each time you run the trend analysis.

------------------------------
Dan O'Leary CQA, CQE
Swanzey NH
United States

Original Message:
Sent: 15-Mar-2022 09:59
From: Andrea Duca
Subject: Intepretation of Article 88 Trend Reporting EU MDR

Hi all,

I am interested in getting your feedback on the interpretation of Article 88 on Trend reporting.

Being said that EU MDR 2017/745 defines:
Serius incident as:
any incident that directly or indirectly led, might have led or might lead to any of the following:
(a) the death of a patient, user or other person
(b) the temporary or permanent serious deterioration of a patient's, user's or other person's state of health
(c) a serious public health threat
Incident as any malfunction or deterioration in the characteristics or performance of a device made available on the market, including use-error due to ergonomic features, as well as any inadequacy in the information supplied by the manufacturer and any undesirable side-effect

Articles 88 requires that manufacturers have to report any statistically significant increase in the frequency or severity of incidents:
- that are not serious incidents (that in my interpretation are just incidents = complaints according to ISO 13485) or
- that are expected undesirable side-effects
that could have a significant impact on the benefit-risk analysis and which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits

Question 1)
Do you agree that the definition of Incident is very similar to the definition of Complaint according to ISO 13485?

Question 2)
According to article 88, it seems that manufacturers have to perform trend reporting not on all complaints or all expected undesirable side-effects (euse) but just on a subassembly that is defined as complaints/euse that:
- could have a significant impact on the benefit-risk analysis AND which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits

I am asking this since it is a common assumption to consider trend reporting applicable to all complaints and to each associated risk. Instead, it seems that trend reporting is just for risks that are unacceptable when weighed against the intended benefits.
Do you agree with this interpretation?


------------------------------
Andrea Duca
Roma
Italy
------------------------------

Annex III(1)(b)(fifth indent) says that trend analysis applies to "the incidents subject to the trend report as provided for in Article 88".

Article 88 says that trend reporting applies to non-serious incidents with an exclusion stated in one of the worst sentences in the English language. When you say "subset" I infer this is the set of all non-serious incidents excluding those in the exception.

In practice, I recommend that you ignore the exception when conducting trend analysis. If you don't get a signal for all non-serious incidents, then you won't get one by excluding the exceptions. This is the most common case, and this approach makes it easy. If you do get a signal, then consider excluding the exceptions and running the analysis again in hopes you will not get a signal on the second pass. If you get a signal the first time, you will likely get one the second time, so the extra classification work may not be worthwhile.

The next problem is to parse the sentence to determine what you may exclude from the analysis.

The sentence says to report "any statistically significant increase in the frequency or severity of incidents that are not serious incidents or that are expected undesirable side-effects that could have a significant impact on the benefit-risk analysis referred to in Sections 1 and 8 of Annex I and which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits".

I'm leaving out the statistically significant part because it applies across the sentence. I've also written this propositional logic format, since it is easier to understand. The logical connectives are in capital letters and the groupings use square brackets.

IF it is a non-serious incident THEN report under Art (88)

IF it is an expected undesirable side-effect AND [it has a significant impact on the Annex I(1) OR Annex I(8) benefit-risk analysis] AND it may lead to an unacceptable risk to health and safety THEN report under Art (88)

From here it is easy to build a decision tree. Applying it may not be so easy.

I recommend that in your EN ISO 14971:2019/A11: 2021 list of residual risks that you include each expected undesirable side-effect related to hazard, sequence of events, hazardous situation, harm sequence.

As you can see, I disagree with your analysis because you say the exceptions apply to incidents whereas they apply to expected undesirable side-effects only.

One more point on reportability. The sentence, "The significant increase shall be established in comparison to the foreseeable frequency or severity of such incidents in respect of the device …" establishes another reporting criterion.

Compare the incidents and expected undesirable side-effects to their location on EN ISO 14971:2019/A11: 2021 risk matrix. I've seen cases where the current trend is flat, but is in a higher frequency of occurrence than determined for the residual risk. So comparison with the foreseeable frequency reveals a statistically significant increase.

------------------------------
Dan O'Leary CQA, CQE
Swanzey NH
United States
------------------------------

Original Message:
Sent: 15-Mar-2022 12:10
From: Andrea Duca
Subject: Intepretation of Article 88 Trend Reporting EU MDR

Thank your for the promt reply, Dan.

I wrongly used the term "subassembly" in the place of "subset".

So, do you confirm that medical device manufacturers selling their devices in the EU are only required to evaluate trends for incidents that:
- could have a significant impact on the benefit-risk analysis and
- which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits

I give lot of importance to the "and" that is placed between the two bullet points. So do you?
This would really restrict the scope of the trend analysis to just a subset of incidents (i.e. those that satisfy both bullets points)

------------------------------
Andrea Duca
Roma
Italy

Original Message:
Sent: 15-Mar-2022 10:34
From: Dan O'Leary
Subject: Intepretation of Article 88 Trend Reporting EU MDR

Question #1
The definition of complaint and incident are not very similar. A complaint is much broader. Every incident is a complaint, but not the other way around. For example, a package with a label that has the wrong size is a complaint, but not an incident.

Question #2
Article 88 applies to medical devices, not to components, subassemblies, etc. Also, distinguish between trend analysis and trend report. Trend analysis determines if there is a signal, while trend reporting informs the regulators of the signal. One could conduct trend analysis, not generate a signal, and therefore, never report.

Start by classifying each complaint (EN ISO 13485:2016/A11:2021) as an incident (EU-MDR) or not. Classify each incident as serious or non-serious. Report serious incidents following Art. 86. Analyze non-serious incidents following the Article III, PMS Plan.

In Annex III, the PMS plan sets up trend analysis (or most of it) by defining the method to detect a signal and the observation. It doesn't require the frequency at which perform trend analysis, but this is important. As a default, I recommend conducting trend analysis once each quarter with an observation period of four quarters. Determine the non-serious incident rate, plot it over the observation period, and calculate the slope of the linear regression line. If the slope is positive and statistically different from zero, then report under Article 88.

Setting up the test for statistical significance in Excel is easy. Plot the rate against time and tell Excel to show the linear regression line and the equation on the chart. Use the Regression tool in the Analysis Tool Pack to set up the statistical test. It is easy to do and should take less than thirty minutes each time you run the trend analysis.

------------------------------
Dan O'Leary CQA, CQE
Swanzey NH
United States

Original Message:
Sent: 15-Mar-2022 09:59
From: Andrea Duca
Subject: Intepretation of Article 88 Trend Reporting EU MDR

Hi all,

I am interested in getting your feedback on the interpretation of Article 88 on Trend reporting.

Being said that EU MDR 2017/745 defines:
Serius incident as:
any incident that directly or indirectly led, might have led or might lead to any of the following:
(a) the death of a patient, user or other person
(b) the temporary or permanent serious deterioration of a patient's, user's or other person's state of health
(c) a serious public health threat
Incident as any malfunction or deterioration in the characteristics or performance of a device made available on the market, including use-error due to ergonomic features, as well as any inadequacy in the information supplied by the manufacturer and any undesirable side-effect

Articles 88 requires that manufacturers have to report any statistically significant increase in the frequency or severity of incidents:
- that are not serious incidents (that in my interpretation are just incidents = complaints according to ISO 13485) or
- that are expected undesirable side-effects
that could have a significant impact on the benefit-risk analysis and which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits

Question 1)
Do you agree that the definition of Incident is very similar to the definition of Complaint according to ISO 13485?

Question 2)
According to article 88, it seems that manufacturers have to perform trend reporting not on all complaints or all expected undesirable side-effects (euse) but just on a subassembly that is defined as complaints/euse that:
- could have a significant impact on the benefit-risk analysis AND which have led or may lead to risks to the health or safety of patients, users or other persons that are unacceptable when weighed against the intended benefits

I am asking this since it is a common assumption to consider trend reporting applicable to all complaints and to each associated risk. Instead, it seems that trend reporting is just for risks that are unacceptable when weighed against the intended benefits.
Do you agree with this interpretation?


------------------------------
Andrea Duca
Roma
Italy
------------------------------