Literature Search - Notified Body Had Questions About our Rationale

[shiftkoji]

We have submitted our Clinical Evaluation Report to our notified body, and they came back with some cryptic feedback we're not quite sure what to make of it.

We're taking the literature review approach, compliant to MEDDEV. 2.7.1 Rev.3.

The main issue focuses on our Rationale for the Selection. They said the detailed description of different stages of literature search (including identification, appraisal, analysis and conclusion of hits were not available.

We felt we did address this, so I guess I'm reaching out to hear what others have done for this. Perhaps we simply were not thorough enough, or didn't have enough citations? What are others doing to provide rationale for their selection? Does anyone have any examples they're willing to share?

 

[somashekar]

Please read THIS ... Thank you

 

[akp060]

For MDD:
In MEDDEV 2.7.1/Rev 4 (June 2016) Section 8.2 says "Literature searching is used to identify data not held by the manufacturer that are needed for the clinical evaluation". What I understand from this is if I am the manufacturer and data (results from investigation/ other clinical experience) that is not held by me, I can do a literature search for either my device w.r.t literature that cites result by using my device (assuming I do not have an equivalent device) OR search for literature/clinical study results for a similar device to support the safety/performance of my device (assuming I have an equivalent device). Is that correct?

So above there are two options that I can think of. However, if in the latter case the literature does not support the performance can I skip literature search in case there is no literature for my device (former option)? and use only data held by me to cite compliance to ER 1, 3 and 6

 

[Raisin picker]

The point is that in the end, you need sufficient data. For ER 3, your own clinical trials might suffice. For safety, you should collect all data available:

• literature (clinical trials, reviews, case reports) on own device, equivalent device, similar device
• competent authority databases on equivalent device, similar device (you should know about incidents of your own device anyway)
• ongoing or finished clinical trials (clinicaltrials.gov or clinicaltrialsregister.eu) can lead to additional data sources
• Registers
• ...

So, it's not "OR", but "AND". When you take the MEDDEV literally (noone does, including NBs), you need to demonstrate that you considered all potential data sources. And this will be enforced more with MDR.

If, of course, your device is really novel, and there is no equivalent or similar device, you have to rely on your own data.

Currently, for example, there is a discussion about Breast Implant-Associated Anaplastic Large Cell Lymphoma. It might be related to certain textured implants, or not. There is very few data, so you have to collect everything you find if you work with silicone implants (not necessarily only breast implants). Competent authorities are looking into this, not only NBs.

 

[monoj mon]

Below is an excerpt from what I understood by "Data retrieved from literature" as mentioned in this MEDEV guidance.

Besides the clinical data that you have mentioned, other literature regarding the current clinical practice/alternative options to treat the intended medical condition shall also be considered under this "data not held by the manufacturer" concept. You will need to explain the current knowledge/state of the art for the intended condition and if there is any alternative option without using your device then you will need to explain why using your device is beneficial and under what circumstances.

 

[akp060]

If, of course, your device is really novel, and there is no equivalent or similar device, you have to rely on your own data.

Here do we even need to go through Literature search strategy, keyword identification and the rest?

 

[Raisin picker]

I'd recommend to do literature search anyway. If only to prove that there really is nothing comparable. Chances are, the NB wants to see this.
You will do this anyway, for development, and you might find some helpful preclinical data.

Last time I did this I was surprised how much preclinical or purely scientific studies had been performed in that field. Most of them had no connection to a product in development, so I haven't heard of them before.

 

[Weeder]

The point is that in the end, you need sufficient data. For ER 3, your own clinical trials might suffice. For safety, you should collect all data available:

• literature (clinical trials, reviews, case reports) on own device, equivalent device, similar device
• competent authority databases on equivalent device, similar device (you should know about incidents of your own device anyway)
• ongoing or finished clinical trials (clinicaltrials.gov or clinicaltrialsregister.eu) can lead to additional data sources
• Registers
• ...

So, it's not "OR", but "AND". When you take the MEDDEV literally (noone does, including NBs), you need to demonstrate that you considered all potential data sources. And this will be enforced more with MDR.

If, of course, your device is really novel, and there is no equivalent or similar device, you have to rely on your own data.

Currently, for example, there is a discussion about Breast Implant-Associated Anaplastic Large Cell Lymphoma. It might be related to certain textured implants, or not. There is very few data, so you have to collect everything you find if you work with silicone implants (not necessarily only breast implants). Competent authorities are looking into this, not only NBs.

We have somewhat of a similar case where there is no device similar to ours in the market. However, there is a lot of literature available on our own device. This data comes from clinical studies, research studies and articles appearing in medical journals where our device was used. My question is this - Is there a limit to how much data should be appraised? Say if you have already appraised over a hundred pieces of literature and have shown that the device is safe and effective, is that enough or do you continue to appraise each and every piece of literature that is being published and where your device was used? Where and how do you draw the line?

 

[Raisin picker]

Hi Weeder, in that case, you have to tighten the screw at the literature selection process. Define your selection criteria/ appraisal criteria in a way that only a reasonable number of publications are selected (e.g. by a minimum number of study subjects, or by publication date).
Do also take care to remove duplicate data. I once had a product with tons of publications as well, but these publications were all based on a few long term clinical trials where yearly interim reports were published, and reports about specific aspects (from that trials), and "reviews" with subjectively selected publications (from that trials). In that case, I tried to weed out all but the most current "official" study publications, and so could drop most of that publications.
And be aware that you have a luxury problem, many companies look at a list of publications in the range of zero ;-)

 

[Weeder]

Hi Weeder, in that case, you have to tighten the screw at the literature selection process. Define your selection criteria/ appraisal criteria in a way that only a reasonable number of publications are selected (e.g. by a minimum number of study subjects, or by publication date).
Do also take care to remove duplicate data. I once had a product with tons of publications as well, but these publications were all based on a few long term clinical trials where yearly interim reports were published, and reports about specific aspects (from that trials), and "reviews" with subjectively selected publications (from that trials). In that case, I tried to weed out all but the most current "official" study publications, and so could drop most of that publications.
And be aware that you have a luxury problem, many companies look at a list of publications in the range of zero ;-)

Excellent advice, thanks.