PMS Reports – Data Sources
Now let’s look at all that goes into these reports. On our left we see a list of all our internal data sources that is we have all our manufactured health data. This can include your device information, claims, internal testing, risk documents, PMS, FMEAs and device claims matrix. On our right, we have all the data that we collect from external sources that gives us all our peer review literature. We have multiple databases where we can get all this. We also have clinical trial registries and adverse event databases.
Remember all this data goes into documentation for one device back to for one cycle of these reports. These reports have to be maintained. That means you must repeat the process of data collection and analysis all over again when it’s time to update these reports.
Manufacturers must invest time and resources to execute these and if you’re only relying on traditional manual means to collect, you would have to deploy a lot of resources to meet your timelines. This could be a big challenge for all companies that cannot continuously add resources to their teams.
Key Pain Points with Data Sources
I want to touch upon the problems and pain points that could arise when we manually handle all our data sources.
Data collection: The biggest challenge with manual data preparation is that one or more people must take time out from their day to gather data, to use in any form of report. This could also mean that they are taking time away from analyzing this data and making an informed and timely decision. One will have to spend time devising some sort of a tracking system, maybe an Excel sheet or perhaps a pen and paper system. There’s often no central repository for all this information. Manual data is tremendously time consuming, not really a value-added activity for someone like a PhD or a clinician in your team. The time and expertise of these highly skilled resources should be targeted towards actual data analysis versus collection or standardization of data.
Consistency: If your data isn’t widely available, or if your business logic itself isn’t centralized, your teams are reporting on different and conflicting data sets. One department or user may capture and calculate something one way while another calculates it entirely differently. This lack of a single source of truth can cause distrust in your actual data and it usually leads to more manual manipulation at a later point in time. Over time – This can lead to data silos. That means it becomes harder for cross-departmental agreement and collaboration. Problems with consistency could come from, for example, the way people regard their search strategies or mainly be how you capture your article details. Your templates could have inconsistency, or maybe a strategy change in your search protocol itself. We know this happens often and writers must spend additional time collating all this additional data and trying to make it look uniform so that it fits into your original data set.
Reporting: Your manual processes would inevitably mean inaccuracies or errors in calculations and reporting. Simple mistakes like a copy paste error or maybe you put in a wrong V look up formula in your Excel sheet. These compound over time, and these can go unnoticed before it’s too late. Additionally, there can be problems with article traceability, referencing and variations in the report output from one user to another.
Maintenance: Now you remember all the reports that we saw earlier. All of those must be maintained periodically. For us to have consistency with the initial report, we need to ensure that we have everything that we need, right from the original plan, which means you need to have your older search strategy. You need to have all of your previous screening and appraisal data, all your previous reference articles, your source documents from your first report. If you have a lot of individual contributors out of which some of them may not be permanent employees. Their methodology or standardization may be different from what you follow internally and if all their work is not really stored internally in a safe place, then that means all that raw data could be a challenge when it finally comes to these reports at a later point in time.
What if there was a solution to all these challenges while also improving your team productivity, report quality and help you with compliance?
What if technology could help you with bulk import of search data, maybe automatic import of free full-texts, article audit trails give you customizable data collection forms, also give you a centralized repository for all your reusable data, help you with data modularization? This is important because one manufacturer may have different devices, which pretty much follow the same therapeutic area and the disease conditions that treat them are also similar. They can leverage data from one report to another, maybe content like your clinical background, your state of the art, some of the sections from the state of the art, that is maybe a guideline section something about the alternative therapies. All of this can be leveraged for other devices also if they fall under the same therapeutic area.
What if technology could also help you optimize your entire data collection and storage process, and give you automatic triggers for report updates? The answer is YES, it can help you with all of that and it has.
I want to introduce all of you to CAPTIS™. CAPTIS™ is a proprietary technology that we’ve built that allows our team to reduce manual effort and reduce higher quality CERs in a lesser amount of time.
The goal is to have one place where the righter processes, both internal and external data sources in order to create the final report. Now since MDR requires the various reports to be updated on a regular basis. You will always have a high level of effort going into maintenance of these reports.
By working with Celegence and getting access to this solution, you can seamlessly maintain your reports on a regular cadence going forward.
CAPTIS™ – Search Home Page
Let me show you CAPTIS™ in action now. This is CAPTIS™. This is your reports overview page. This is where you’ll get to see a brief summary of what’s happening in your project. You’ll get to see if you have any pending actions against all the articles that you retrieve for your DUE searches or for state of the art (SOTA) searches. We also have a little notice board at the bottom where you can leave comments for yourself. Maybe you need to follow up with someone, or if you have any collaborators on this project, they can also leave comments for you.
This is where all your searches get stored. We see a couple of searches that I’ve run already. To conduct a new search, you click on the search option, and we see a couple of databases listed. Now PubMed, Europe PMC and Google scholar are directly integrated with CAPTIS™ that means any searches that you do on these platforms, you can conduct them on CAPTIS™ directly.
The next three databases Embase, Cochrane and Prospero. You can search the parent websites and import all your article details into CAPTIS™.
Now let’s run a sample search for PubMed. The first data field that we see here is something called SEARCH NAME, where you can assign a name to your search. Whether you’re doing a particular string for your clinical background, maybe you’re doing a search for your alternatives. All of them can be captured here. So, we have our search strings here and what you see in black and in bold like competitor, alternative therapies, your clinical background, you can capture all your search objectives here itself. Meaning because you have multiple strings within the same project. It can get confusing. If you can’t remember why you chose a particular set of keywords or why you ran a particular string with a particular set of filters, you can assign names to yourself so that you can easily identify the objective.
Let’s go back to our search page. Let’s do a very simple search. I have a very basic set of keywords with me here. I’m going to call it a test search. Let this be a search for DUE. Now CAPTIS™ allows you to set your preference. That means do you want to do a search for your device under evaluation? Are you only looking for safety and performance data for your device? or are you doing a search for state of the art? You can choose either of these two options from here, you can assign a search category because we are running a particular full string. We will assign it as a scientific search. You can then set your date range. Let’s just choose one year for now. You can even choose article types, just like you can on PubMed. Let’s say we are only looking for clinical trials for now.
You can even choose your languages and CAPTIS™ also has a species filter like PubMed. You can select and hit search, we got 243 articles on CAPTIS™. We can verify this. If I go here, this will expand the string, like the way PubMed expands the string and then I can hit the globe icon here it’ll take me to my PubMed page. We see the same number of articles.
When you’ve run your search, you can quickly browse through your results to see if these look relevant to your search objective and then if they do, you can simply hit save and CAPTIS™ will save your search. Let’s go back to our screening homepage. So, this is the string that we just ran, and it has captured all our details. But if you look closely, CAPTIS™ has also captured which filters were used when you were running the search.
It has captured the string obviously and we also see which search objective was it that the search was conducted for. We also see which database the search was conducted on and what time, what date range was used. So this is essentially your entire search strategy in one place. This really helps us with traceability of our articles, and it also ensures that we record the correct information on our search protocol in our CERs or any other report pages that you’re writing. This will also eventually help you with your reviews. If your team performs your searches and their reviews on CAPTIS™, the reviewer can just log onto this page and then see what kind of searches were done, what kind of search strings were used. They can go and verify the results.
Let’s say we are doing a SOTA screening and let’s go to level 1 (L1). So systematic literature reviews typically have two levels of screening. Level 1 (L1) is where we are looking at the title and the abstract only. Level 2 (L2) is where we look at the full text to decide. Right now, we are in L1 of my report, and I have a couple of article filters on my right here. Before we jump onto filters, now because we have multiple search strings, and we have multiple databases. It’s quite likely that we will have duplicates. CAPTIS™ automatically identifies these duplicates. Like right now we have 52 articles that CAPTIS™ identified as duplicates and these are automatically excluded for you so that you don’t have to de-duplicate anything on your own.
Now let’s select a couple of filters. Let’s say I want articles from PubMed, and I want all my articles that I’m yet to review. We see 536 articles here. To begin reviewing an article, you can simply click on it, and we see a new window pop up on our right. We see all our article details. We have the article title here; we see article authors. We also have other details like the DOI, PMID and all of these are clickable hyperlinks. You can click on any of these links, and it’ll take you to the original PubMed search.
So, you can see the abstract below. You can read the abstract, take a decision. If this looks something that you want in your report, you can simply hit include here and that’s it.
Let’s jump onto our next article. To exclude an article, you can simply click exclude here and you get a dropdown list of exclusion reasons that you can choose. Now, this list is fully configurable. You can add as many reasons as you want, depending on the process or the template that you follow for your reports.
Let’s say this is a different indication and I’ve excluded that. A cool feature that I really like about CAPTIS™ is article tags. Now let’s say you’re doing your state-of-the-art search and you know that there are multiple objectives that you need to fulfill. That means you have a section for clinical background that you need to write. You also have alternatives that you need to cover. Now, I look at this article in a way, I think this has some great content for my alternative discussion maybe. I can tag it for alternatives so that I remember later why I included this article. Now say this article also has some great content that can be used for my clinical background. I can also assign another tag clinical background and be good.
Let’s also include this one now because we have two very distant objectives in our search. That is the DUE search that is for safety and performance, and this is device specific only. The second objective being state of the art (SOTA). Now, it’s quite possible when you’re doing a review of one objective, and you come across articles which fulfill another objective. Let’s assume this is a SOTA review. Now this article has content for my device. I will not include this for state of the art (SOTA). Rather, I would include this for my device under the evaluation. So, I can exclude this article from this review using an appropriate reason, and I can push this off in my DUE search. CAPTIS™ will automatically push these articles into my DUE workflow, and I can review them as a part of my DUE set of articles.
Other things that you can do is edit article details. Let’s take another article. Let’s assume you did not get your entire information that you were looking for. You need to edit some links. You can always go to the options here and go to edit details, and you can edit any article identified from here itself.