Monir Azzouzi: [00:00:00] Welcome to the Medical Device Made Easy podcast. Here is Monir Azzouzi from medicaldevice.com. And today we will talk about clinical evaluation report or CER and more specifically, we will try to help you to perform your literature search on your clinical evaluation report. And for that I invited a guest, a person that is doing that at his company or where the company is doing that. It’s Ed Drower, the director of regulatory affairs from Cite Medical Solutions. So Ed, welcome to the Medical Device Made Easy podcast.
Ed Drower: [00:00:36] Thank you very much, Monir.
Monir Azzouzi: [00:00:37] So Ed, just a quick introduction of yourself and maybe of a medical solution and then we can really help our audience on how they should do the literature search, because I think it’s really an exciting topic for some of the the people in the audience where they are maybe struggling with that. So please just introduce yourself.
Ed Drower: [00:00:58] Sure. Um, again, my name is Ed Drower and I’ve been doing research, filing support filings and reviews for individual medical devices and some combination products both in the US and Europe. With decades of clinical research and drug development experience, I provide a broad perspective to the CiteMed Solution group. We’ve evolved. We’ve involved a lot of preparation that we’re going to talk about a little bit today, about the literature review and how Cite Medical Solutions has approached this. This particular task, which is, I think, very arduous. The Cite Medical Solutions Group was really was really created or founded on a platform of the regulatory needs for this particular literature review as well as technology platform that has provided us some advantages over trying to do this manually. But I’ll go through a little bit more about that as as this talk goes on.
Monir Azzouzi: [00:02:05] Yeah, no, it’s clear. So it’s great. So yeah, as I mentioned, you are mainly helping a lot of your customers about that. And I wanted to ask you. Yeah, to help maybe also the audience to really understand this task of literature search when you are doing a clinical evaluation report and first wanted to ask you mainly what is the best technique to make a literature search? What exactly are the different steps to to consider when we are making a literature search?
Ed Drower: [00:02:35] Yeah. And there’s there’s quite a few steps involved, but I’ll try and break it down simply for for the audience here. The first thing that you really have to understand is a literature search is a systematic approach to trying to find enough information about your product or device so that you can make an assessment of benefit and risk on your device. And this will help flow into the ultimate report that you’re going the reports that you’re going to need, which will be the CR. And then the later on, you’ll have to define your post-market clinical follow up and other types of reports. But just to get back to the actual systematic approach, the first thing that you really want to do is you want to develop a protocol, which is really the guideline of how you’re going to search through all the literature and come up with the ones that are most relevant to support your risk benefit and your product’s safety assessments. And so.
Monir Azzouzi: [00:03:35] Oh yeah, go ahead. So mainly when we are saying to make a protocol, it’s mainly to make a document where we are planning to do this literature search to explain how we do that and to explain also where we will look for those, for those documents or for those articles or all the literature that is existing on on the on any database. Is it correct?
Ed Drower: [00:03:57] Yeah, that’s correct. It is an actual document that it needs to be prepared and is part of the CR. So you’re going to have to have a protocol that outlines exactly what you did and also the output, which is your report of the literature review. So yes, you need both of those.
Monir Azzouzi: [00:04:13] Good. And I if I heard well or if I know well. So it’s mainly something that an auditor can also use to make to prove that to check that you have done really the correct search because you will you will check exactly which database or which keywords or which things you have used on your protocol. And you will you can you can do. I don’t say will do but maybe you can do the same search and check if you have really done it correctly. If I can say.
Ed Drower: [00:04:42] You’re absolutely correct and and they do do check those types of things occasionally just to make sure that it works and that you are really providing all the information. So, yes, it’s very good.
Monir Azzouzi: [00:04:55] And when you have this protocol. So now what are you doing with that?
Ed Drower: [00:05:00] So you’re you’re putting together a protocol, but you really have to understand your product. And that’s usually gotten from a product insert or a label that the product the manufacturer has provided to you. And through those through those documents that you’re provided, you’re going to develop some sort of search terms that are going to be used. And so that’s why it’s important to understand the product that you’re going to do this literature search on.
Monir Azzouzi: [00:05:27] So is it is it something that when we are talking about a certain product, do we need to be qualified really on this product to perform a literature search? Or we should just know the method or some keywords and then we can do it?
Ed Drower: [00:05:43] Well, you really want to do your homework. And the reason why is because there’s going you’re going to find a lot of terms that are probably a little too broad that will give you way too many literature citations back. So you really want to understand your product well enough so that you can narrow the search down to the right amount of of articles that you can review.
Monir Azzouzi: [00:06:08] Yeah, because I think you can get some thousands of articles coming out and then imagine the amount of work to review all that and to and to check what is good and what is bad. But yeah, so we have, we have the keywords that is really important. What about the databases that we have to, to check on?
Ed Drower: [00:06:31] Yeah. So for every search term that you identify, you’re going to probably want to select at least three different databases to search them through. So if you have five search terms, they’ll all have to go through every single database. And there’s a lot of databases out there. Some of the key ones are like PubMed is a key one, which is, I think, similar to what they call Medline. There’s also Embase. There’s there’s other databases that you might have heard of called Cochrane. And then of course, you’ll also find a lot of safety databases from the US. There’s a safety database called Maude, and then there’s additional databases in other countries that are also available.
Monir Azzouzi: [00:07:15] So if your products, for example, if you want to register your products in in Europe, is it okay to search in a database outside of Europe?
Ed Drower: [00:07:26] Oh, absolutely. And actually, some of the larger databases are here or in the US and they’re very well recognized even by Europe.
Monir Azzouzi: [00:07:35] Okay, great. So I think it’s good because we have, as you mentioned, there is a lot of database and we have also to prove to the auditor that the data that we are collecting are really relevant and it can be a good, a good proof that our product is correct. Just one, maybe one comment. So are those databases free of access so you have to pay a subscription for that?
Ed Drower: [00:07:58] Well, most of them are free. You do. There is one that you do have to pay for. I think mbase is one that charges. But there’s you can work out some kind of way of getting through those.
Monir Azzouzi: [00:08:11] And if I tell you that we can also make some research directly on Google, will they accept that or not?
Ed Drower: [00:08:18] Well, I think Google I think there’s a Google scientific database that you can search. The problem that you might have is that it’s it’s very robust. And the other problem that you’re going to have is that just like with all the databases I mentioned, there’s generally a lot of duplication between these databases. They all seem to have access to the same information. So what you’re going to see is that as you do your searches through all these databases, you’re going to end up with hundreds of citations or what we call hits, and there’s going to be a lot of duplications that you have to sort through.
Monir Azzouzi: [00:08:56] So it’s great. So as you mentioned, after we are doing all those searches, we can get a thousand of of articles. So what should we do then? Get this thousand list. So should I cry because I have to review all those? Or what is the next step then?
Ed Drower: [00:09:15] Well, there will be tears shed if you have thousands of articles, and that is unfortunately the manual process in which you have to go through in order to sort this large amounts amount of information out in order to get really the information you’re looking for, the relevant information. So, yes, it’s hard.
Monir Azzouzi: [00:09:38] Yeah, it’s hard. But then is there a way maybe to exclude or some of those articles and to say they are not good enough or they are not relevant? Or is there something that we have to do here?
Ed Drower: [00:09:51] Well, so as you go as you get your list of citations, occasionally there’ll be just titles. When you do your searches, you can set up your parameters on a lot of these databases to not only get the titles of the articles, but also get an abstract. And even some of these you can actually get free articles. But as you’re starting to go through the the list, you’re going to review the title, you’re going to assess whether or not you think it’s relevant or maybe potentially relevant. Then you’re also going to look at the actual abstract and continue your review process to determine whether or not this is an article that you want to obtain and and use in your analysis. So it’s a very long process to go through all those, and especially when you’re having to do it, duplicate times sorting that can be rather problematic. And so I really want to stress that you still have to look at everything all the way through your list. So whatever list that you get, you obviously don’t want a large hundreds of articles to search through for every search term because that’s going to that’s what really takes the most amount of time.
Monir Azzouzi: [00:10:58] Yeah, exactly. So as we’ve said so we have this article, we have these articles, we excluded maybe some duplicates or some that are not really relevant. So what is the next step? Normally we have also when we are looking at the med dev for the clinical evaluation report, we have the method called appraisal appraisal method. So what should we do then with this appraisal method?
Ed Drower: [00:11:24] So so the med dev that you mentioned does define acceptable methods and categories that you have to review each article by. And so I can talk to you a little bit about some of those categories. And again, med dev allows you to adapt what they’ve stated into your protocol. And as long as you’ve defined it well enough, you can select different categories to do your to do your assessments on. So every single article has to be assessed for certain criteria. And each criteria that I’m going to mention is typically given a grade and you can give it grade like a one, two or a three, and you can define that yourself, but they do require that you at least do some assessment of each article. One of the areas is, is whether or not the product is suitable is the article. Talk about the device itself, the particular device, or does it talk about a similar device? Or maybe something that may or may not be related. So you’re going to grade each one of those articles based on just the suitability of the of the device.
Ed Drower: [00:12:35] Another area that they look at is the application is the device being used in the same manner? It could be identical or it might be slightly different. You’re going to have to make a great assessment of that as well. Another area that they look at is the population that it’s going to be used on. So you’re going to say it’s the same population or a different level of population that is actually applied. And then maybe one of the fourth areas that you really want to focus on and that I think the notified bodies are more keen on is the actual quality of the citation. Is it a very high, a high data driven type of article? Is it double blind, randomized, placebo controlled, or is it just something that’s just more of a comment about the product? And each of these things have different levels of importance and that’s what you help to define in your protocol so that the reviewer knows that you’ve defined these and that you’re only picking certain articles that are what they consider relevant based on all of these categories.
Monir Azzouzi: [00:13:39] Now it’s great. So as you mentioned, this method is is created by the company itself. So it’s not something that is coming from a standard or coming from any guidance. It’s something that the company is really creating.
Ed Drower: [00:13:54] It’s a modification of the it’s taking what the med dev has has listed and you’re modifying it to what you feel is appropriate for your product.
Monir Azzouzi: [00:14:03] It’s great. Okay. So I made the search. I have all my 1000 articles. I made the appraisal. So now I have graded each of them and maybe I have really selected the one that are most relevant for my product, for the area, for the population, for the database, etcetera. Now I have maybe okay, 200 articles that are remaining. What should I do with those 200 articles?
Ed Drower: [00:14:31] And this is where the tears really begin, because you’re going to unfortunately have to go through every single one of those articles and categorize every single one as I just described, all those different areas. So you’re not off the hook yet. You still have to define all of those. And what’s even more demanding, I think, in this whole process is that in your reports, the reviewer wants to see all this information tallied up. So you’re going to have to tally up every single category. All the articles you found, all the ones that were relevant, the ones that you thought maybe were potentially relevant and the ones that you just described that weren’t relevant at all. But they want to see this all defined and all all these things tabled out so that in your report they can easily see that you’ve done your diligence in your review process and that you’ve actually found as many things as you can and you’ve actually presented the most relevant work and data for your review.
Monir Azzouzi: [00:15:30] Okay. So you have then I think this those three steps or those three things are really what will take most of the time for anybody that is creating a clinical evaluation report is correct.
Ed Drower: [00:15:46] Well, that’s that is the one that takes most of the time. I should mention that when you’re defining your search terms early on, as you start doing your as you start reviewing them and sending them to the database and getting all your hits, as I said, that takes a lot of time as well. So you may find thousands of hits and you may have to find another search term and then you have to rerun it and then you have to kind of narrow it and narrow it down. And that takes a lot of time as well. So the initial part takes an enormous amount of time to get the right amount of literature searches, the literature, articles to find. And I don’t want to to make that a small part either. And sometimes working with customers and different clients, you have to go back and forth with them to make sure that they agree that these search terms are the ones that you that you really think are most relevant, because that is really setting the stage for all of your subsequent reporting in the MDR that’s coming up. And that’s not only your CR, but that’s what’s going to help you to define what you need for your post-market clinical follow up and as well as your periodic safety update reviews. All of those things are based on these terms. And so getting them right up front is really important.
Monir Azzouzi: [00:17:01] Yeah. And what what now as we talk about timing also and resources. So what how what is the range for the creation of this literature search and to arrive until the the final review of the of the of the literature. So is there a range of time like two weeks or one month or.
Ed Drower: [00:17:21] Well you know it’s funny if you if you if you’re really talking about. Doing this by hand and actually going through the process. I have yet to find someone who will give me an actual number of hours that they’d have to spend to put this together. I think it can range sometimes depending on the product. It could it could be months before you actually can pull together a literature review and a together using a process that the CiteMed Solutions has put together. We have really shortened this whole process. We can probably, depending on working together with clients. And again, a lot of this is process oriented. We tend to want to work with the clients very closely so that we can get this back and forth process finalized very quickly. Once we get this protocol finalized, a review, a literature review can take somewhere between 2 to 3 weeks. So it’s a very short process with our systems. As long as we can work closely with the clients to get this protocol finalized.
Monir Azzouzi: [00:18:30] But when you are talking about your system, so are we talking about people working 24 over seven or what’s what’s what are we talking about here?
Ed Drower: [00:18:39] Well, the system is designed to really harness the our medical writers to really focus on reviews of the critical articles. So our process removes all the duplication and it allows them to categorize as they’re reviewing. So it’s a very concise, systematic approach to the review process. And so once you get this all through, the reviewers get it completely reviewed. The next part of this is really the formatting of the report. And and that is something that we also have done quite well and that we we have found that it consistently is an output that we find is very valuable and it’s easy to review and it’s something that can be generated quite quickly.
Monir Azzouzi: [00:19:25] Okay. So just yeah, let’s let’s continue with Cite Medical solutions. So where are you based again?
Ed Drower: [00:19:31] So we’re a virtual company, but right now we’re based in the Chicagoland area.
Monir Azzouzi: [00:19:38] Okay? And mainly here we are talking about literature search and that you can really help your customers on that. What else also can you help them with?
Ed Drower: [00:19:48] So we also do provide development and reports. We also have our systems in place so that if you want to develop this whole process, when you’re developing your for example, you have to list in it sort of the frequency in which you’re going to do your reports for the post-market clinical follow up as well as the safety updates. If you work within our system, all of the information, all of the search terms, all of the articles, they’re all kept into a our database so that when the time comes for doing a post-market clinical follow up, it’s really more or less of just rerunning our system and rerunning all the search terms and things that you already have developed. And so it becomes a much faster process in finalizing those documents as well. So in that way we’ve been able to provide more of a consistent enterprise solution to all of the reporting needs that are going to be coming up for the MDR.
Monir Azzouzi: [00:20:51] So it’s great. So it means that if you have a customer that comes with maybe two or 3 or 4 cars, the first one will take some time, but the second one third, one fourth one can be quicker because you already know all the profile of the company, of the products and of the terms they are using. Is that.
Ed Drower: [00:21:09] Well, I think you meant. Yeah. For the literature review. Yes that’s true. For every KR is a different typical product, but yes, that would be true. Once you have it in place, it’s almost like it just goes. And, and the nice thing about our solution is you can pick it up at any stage. So if you want the literature review done first, that’s fine. If you want all the subsequent follow ups done, that’s also fine. We’ll work with whatever the needs are. And that’s, I think, what makes us a good hybrid solution to a lot of the clients that we’re working with.
Monir Azzouzi: [00:21:43] No, it’s great. So great add. So really thank you for that. So where people can follow up with you.
Ed Drower: [00:21:51] So if they want a little bit more information, which I hope they would go to citemed.com, that’s C I T E M E D.com, and you’ll find a lot of articles we’ve published up there. We also have some references that might be valuable to some of your listeners in terms of the lit review process. I think we actually even have sort of a protocol or template up there for the literature review as well. So I think all this information is beneficial. We welcome people to use it. Again, this is really the core of what we think the MDR is looking for. It’s to really establish a good, concise, repetitive type of process that that’s very consistent and and very useful for the reviewers, which is really what you’re looking for is you’re notified bodies are the reviewers that we’re trying to address.
Monir Azzouzi: [00:22:47] No, it’s great. So I will put all those details on the show notes and any documents that you you’ll share with the audience, I will put that also in the show notes so that they can just click and and download them. So yeah, that they can. I hope this can really help them to understand this process. And if they have really a request to create a CR or a quick literature search, so go to Cite Medical solution that I think they will be really helping you for that. So really thank you for that. Thank you for your help on educating our audience and I wish you a nice day.
Ed Drower: [00:23:24] Thank you so much and thank you again for allowing me to to do this presentation.
Monir Azzouzi: [00:23:29] Great. Thank you.