SCCM Resource Library-How to Create Multicenter Research Registries

Video Transcription

Hello and welcome to today's SCCM Discovery webcast on how to create a multi-center research registries. My name is Paul Jung.  I'm a professor of pharmacy practice at the St. Louis College of Pharmacy in St. Louis, Missouri, and I will be moderating today's webcast.  The webcast will be available to registrants within five business days. To access, log into MySCCM.org and navigate to the My Learning tab to access this recording.  Thank you for all of you for joining us. A couple of housekeeping items before we get started.  To submit questions throughout the presentation, type into the question box located in your control panel.  Please note the disclaimer stating that the content that's going to be following is just for education purposes only. So quick introduction.  I want to talk about today's speaker for today. First we have Dr. Alan Watke, who's a professor of medicine in the Boston University School of Medicine.  He's also the section of Pulmonary, Allergy, Sleep, and Critical Care Medicine, as well as the Pulmonary Center at Boston University School of Medicine, as well as the co-director  of the Evans Center for Implementation and Improvement. As well as Dr. Scott Balesta, he's associate professor in the Department of Pharmacy Practice  at the Nesbitt School of Pharmacy at Wilkes University.  Welcome, everybody. This is Scott Balesta. So we are going to start right off with a poll question for everyone.  So we wanted to get a sense from today's attendees if you've ever participated in a clinical data registry. So you can please pull in your questions at this time, your answers.  This will give us a sense of our audience's experience with data registries. Okay, so the majority of no, but it's pretty close.  So again, looks like almost a little bit more than a third, a little less than half of you have participated in clinical data registries at some point.  So our next question is, have you ever started or wanted to start a registry? I'm curious to see how the answer will be to this one.  I have a feeling it's going to be far less in the yes column.  Oh, okay, so yeah, so okay, the majority of you do want to start registry. Well, then you've got the right webinar.  So first, I want to before getting into, you know, our lecture for today, and going over  the objectives, I want to thank SCCM and Discovery and the Clinical Pharmacy and Pharmacology section for inviting Dr. Waukee and myself to present on this topic to all of you.  So we've had some experience in the last couple years with the virus registry, but we'll give you a first general overview.  So the objectives that we'd like to go through today are to discuss the goals and purposes of a registry.  We're then going to outline the process by establishing your team and collecting and storing the data, which of course is probably the crux of a registry.  We're going to review the process for turning that data into knowledge. So how do we disseminate that data? How do we share it with constituents and so forth?  And then finally, Dr. Waukee is going to go over an explanation in a bit more detail of our experience with the virus registry and the accomplishments we've had from that.  So what is a registry? So some of you probably already know what the registry is, 40 some percent of you, 42 percent have been involved in registry at some point in time.  So more of you want to start one. So I gather there's a general familiarity with what registry is, but it's a collection of standardized information in general.  It could be for the three broad categories represented here on this slide. It could be for a group of patients or a particular population.  So you want to study a particular disease state or so forth. So usually patients may share a particular disease or condition, or it may be something  they're exposed to, a particular treatment of a particular risk factor or group of risk factors and so forth.  Generally also it serves a scientific, clinical or policy purpose sometimes. I think we think mostly registries are serving clinical purpose, but there can also be reasons  to enact policies and so forth for public health measures, or even institutional policies across broad health systems and so forth.  But again, it's a collection of that standardized information in general, a bit more beyond what we would see in a normal observational study or a randomized controlled trial.  So that takes us to our next question. So which of the following is a reason or objective for a clinical registry?  Now we're getting into a little bit of what we're going to go into next, and we're going to get a sense of your thoughts on this.  So is it to assess natural history of a disease, determine the effectiveness of an intervention,  measure safety of an intervention or therapy, improve quality of care, or all of the above? All right, 100%, fantastic.  So only one person responded, or the all of the above is a dead giveaway. So not a very good test question to have in an exam. But again, that was kind of the intent.  So really, a registry covers all of these.  So why have a registry? Why do we need registries? What purpose do they serve? So they can be used for multiple different things.  They're not just one purpose to gather large amounts of data and publish papers on them and so forth. So we use them for many, many different things.  They can be used to assess disease history or natural history of a disease course and so forth.  So they can be used to look at the scope of a problem, give a good idea of the incidence or prevalence of the disease or risk factors for a disease and so forth.  They can be used to trend disease over time. So they're not just snapshots in a brief period of time.  They can be used over decades or even longer periods of time to assess some of the important factors that we want to consider with disease states or with therapies and so forth.  They can also be used to describe or estimate survival as well. So we can also use them to determine effectiveness of therapies, not only whether something is  good or bad, positive or negative, but also cost. They can be done to look at comparing one treatment to another in terms of cost or benefit  in some other regards, such as quality of life. We can also look again, as mentioned earlier, at the safety or harm imposed or that patients  would have from a particular therapy or treatment. So it might be specific products or treatments that we were considering, particular interventions.  We can also sometimes compare evaluations of safety and effectiveness to look at safety measures or harm of a particular therapy.  And again, at the end, we can also look at quality measures as well. And these might be applicable to the domains of public health or particular areas of practice in medicine.  So there are some concerns with building and developing and implementing a registry. Patient-centered registries may provide biased data.  So if you have registries that depend on patients to enter data, you may get biased information  if the definitions of the certain fields and data you want to gather aren't implicit, aren't explained in good enough terms that patients would understand to enter that data.  And even researchers as well, if they're entering data and data fields aren't well-defined, you may get biased data as well.  There might also be a lack of standardized data collection. You might get different data samples from the same patients who are applicable or whose  data might be entered into different registries because they have more than one disease state or because that disease state is followed in multiple different registries.  So you may get variation in data collection over time, depending on the patient's progression with their disease or their treatments and so forth.  You might also get competition for patients across registries in the same example. So if you have a patient who qualifies or has data that qualifies for entry into more  than one registry, you might get competition for that patient across registries and so forth.  And also one of the ethical concerns that we'll talk about a little bit later is regarding biobanking samples and so forth, tissue samples, genetic samples, et cetera.  So again, patients need to be aware of what they're consenting to up front if they're providing biobanked samples for use later on in time.  So here's some examples of clinical registries. Of course, we're going to talk more about virus, and Dr. Walke will go over more detail with that.  But there's a few others on here as well that some of you are probably familiar with. The ANZUS trial group, the ICNARC, EpiMed, Reveal.  I can also think of some others, including the STS database, the UCBR registries for cardiovascular interventions and so forth.  So there's a number of registries that are publicly available or available to researchers and clinicians. Again, that I'm sure you can add to this list as well.  So registries and developing one need particular resources to be successful. So we're going to go into more depth of these.  It needs a clear vision, a clear goal to set up front so that the intent and purpose of the registry is defined.  It also needs a good team to be able to not only make the registry, to build it, implement it, but then also to maintain it and so forth.  There also needs to be a willing consortium of sites or patients or a group that are willing to contribute the data.  If you can't get data for your registry, it's not going to be very successful or much use. There are regulatory issues, as we mentioned earlier, regarding patients with some of the  data biobanking and so forth, but also each participating site regarding samples, regarding patient data and PHI that it would need to address.  Also you need a good amount of infrastructure. You need somewhere to store the data. You need someone to be able to clean that data, extract it.  If your registry is going to be very robust or follow a very large and prevalent disease  state, you're probably going to have thousands upon thousands or tens of thousands of patient entries. So you need to have that kind of infrastructure and support.  You also need logistics. Again, in terms of implementing, developing, and even maintaining your registry, you're  going to have to have regular meetings between your constituents, the sites that are collecting  the data, the group that has been put together to develop and plan and implement the registry. So you need to have a good plan for logistics as well.  And then finally, you need to have a good plan of what you want to do in terms of gathering the data, in terms of quality checks for that, and making sure that the integrity of the  data is maintained during the duration and lifetime of the registry. So let's talk a little bit more about each of these. So for the vision, again, you need to have a common goal.  It may start with one person and one thought, as you'll see from the example with virus, but it needs to be a shared goal and a clear path and direction as to what the registry  is going to accomplish, what its aims are, and what the goals are for the data that's going to be collected.  You also need to consider up front who will have access to that data down the line, once the data has been collected and is robust enough to produce some type of results or  disseminate to constituents. How will that data then be used? And again, how will it be disseminated? Will it be publicly available?  Will it be available for researchers to utilize it for different projects and so forth and answer different hypothesis-generating questions?  So those are some of the things to consider up front when developing the initial vision for the registry.  Also, second in our list of things that you'll need is a very robust and well-diverse team, I should say.  So registries that are very successful have a lot of support and multidisciplinary influence within their team.  It can't just be from clinicians, because again, you're going to have to need the bank data and so forth, maybe sometimes terabytes of digital data for data that's been gathered.  So you need to have the background and knowledge of informaticians, data scientists, statisticians, and so forth.  You may also need the influence of patients themselves in terms of what data might be important to gather regarding a particular disease state and so forth.  And then you also need to have those who are well-versed in research methods and so forth  to be able to have a plan of what you're going to do with the data and how that data is going to be disseminated and analyzed as it's gathered and so forth.  You also may consider having an end-user team to provide feedback. So looking back over time, if things are going appropriately and data is being gathered correctly  and entered in a usable manner into the registry. There also needs to sometimes be an external advisory board, again, to provide alignment  with the vision, to make sure that the registry is staying on task and on focus with the original vision of the registry.  And it doesn't deviate from that initial goal of the registry. So again, your consortium is another critical element.  So depending on what you're studying, whether it's something, a treatment safety or efficacy,  or whether you're studying a particular disease over time and so forth, you need to have sites that not only are willing to participate, but that can provide the type of data you're  looking to gather depending on the purpose, again, and vision of the registry. So again, establishing that consortium, gaining interest, developing basically the input of  those participants who are willing to provide data is necessary. So you want to make sure that these centers would have patients or use the treatments  that would qualify them to enter data into the registry as an example. And then you also may need to establish some sort of funding to support the staff at the  site for doing the data collection and so forth. So it may not be feasible for some sites to participate on a voluntary basis.  They may need additional support in terms of funding or staff to be able to gather the data depending on what they're looking for. But also again, the regulatory issues.  This is one of the biggest things to consider up front, not only regarding regulations for site approvals, so whether you need a central IRB or ethics approval for your registry in  addition to approval at each individual center, but you may also need agreements in place between the main registered site and each individual participating center.  So you may need data sharing agreements, institutional authorization agreements. You may also need to produce and keep on file conflicts of interest from researchers at  each site so that we can be assured when data results are published and so forth that there was no undue influential and bias in terms of gathering the data and so forth.  And then finally, when it comes down to time for publication and sharing the results and disseminating the results of the registry or even sharing the data, you also may need  to have in place predetermined contracts for publication and authorship in terms of how the data is going to be shared either in publications or publicly otherwise.  So infrastructure, going not a step back, but infrastructure is something to consider  really up front. So in terms of what kind of amount of data you're going to gather, how it's going to  be stored and so forth, and what kind of software or storage space you may need, especially if it's going to be primarily digital data, which nowadays that's really what we're  talking about, is something to definitely consider up front along with the patient. So if you're going to be downloading 100 or 200 plus fields or thousands of fields of  data from an electronic medical record, you're definitely going to need a good infrastructure to support that and a data platform that can manage either downloading  that data because it may be not feasible to do manually with the electronic medical record platforms and so forth.  So you might also want to consider some of the data that may be public-facing on a website so that you can rapidly share general data that's being gathered in real time, which  is something that we did with Virus with a public-facing website. And also you'll need to have communication platforms in place as well, not only to communicate  with the main core team and participating sites for the registry, but also for researchers in order to be able to access the data and potentially utilize the data for research  purposes.  So logistics is something else that may need to develop at various points in time as the registry grows and so forth.  So initially you'll probably need some type of logistics and communication with not only the core team members of the registry who are developing the registry and implementing  it and then maintaining it, but also the sites and so forth that are going to be gathering, collecting data and so forth that are participating.  This may need to be in terms of operational meetings that would need to be in place to look at data quality, look at potential research projects that may come from the data that's  being gathered and so forth. And then in the end, when you have a large, robust amount of data, whereas the registry  grows or continues to grow, especially if it's a very longitudinal intended registry, we'd have a process in place for conducting research and utilizing that data that's been  gathered. So you may need to have something in place for how researchers can submit and have a review process in place for proposals to utilize the data.  And then also a process for how that data would be shared and published with other  researchers or the public in general.  So data quality was the last item on our list. And the key to any good registry is good data. So in order to make sure that you have good data, again, when you're developing the  infrastructure and concerning the data entry elements for your registry, you've got to have data fields that limit improvisation.  So you want to make sure that it has clear operational definitions, as I mentioned earlier.  Otherwise, you're going to get biased data entered by not only researchers, but potentially patients if they're entering their own data.  Some simple things like making sure that units are clearly defined, using decimals instead of commas, depending on which regions that you're gathering data from across the  globe, if it's a potential multi-center or multi-national research registry. You want to restrict and limit possibilities.  So no free text should, you know, simple things like not entering or being able to enter free text in numeric fields and so forth.  So these are some of the simple things, again, with the infrastructure that need to be considered. And again, that's where sometimes, you know, the help having a web statistician or  implementation on your team would be very necessary up front.  You also need to look at and consider early involvement, again, of a statistician as well to determine, you know, what or in what form data should be gathered to eventually  do analysis and develop results from that data. So if you've gathered free text information on a particular field that really would be  better off as numeric data entry for statistical analysis, again, up front, it would better to have that statistician involved in the early stages of development of the registry.  And then, again, the data needs to be in a readily extractable format. So again, you need to consider how the data is going to be used, what statistical software  packages might be utilized, you know, in terms of best utilization of that data and extraction of it into software packages for eventual analysis.  So data access regulations, again, this is something to consider along with the data you're going to be collecting. So you want to consider, again, who's allowed to access the data.  Will it just be, you know, the developers of the registry and the sites participating, entering that data?  Will you open up the data to other researchers for use in, you know, answering hypotheses that are related to the data in the registry?  Will you make it public-facing, you know, to share with the public to, you know, promote public health in some regard? How long will different constituents have access to data?  Will you give sites access to their own individual data for their own quality improvement?  If it lends itself to that, will you allow the public to access, you know, record recorded history for documentation and following over time?  And then, again, what is the approval process that may need to be in place to allow these various individuals or constituents to have access to the data?  So, again, these are some things that need to be considered up front, but also need to be considered as the registry grows and develops as well.  And then, finally, again, regarding data, you want to consider, because really the intent of the registry may be for multiple different things, but, you know, as researchers, we  tend to want to use the data for research. Again, it may be publicly available data that, you know, promotes public health in some regard  and doesn't get utilized at that high level for research. But, again, some of the data may be used for research.  So, initial vision and goals should really determine the data that's entered and collected.  And so, again, you have to have a clear vision up front to determine the type of data that you want to have in your registry.  And then, again, the data that's entered determines what type of research can be performed. So, if you're entering data that's erroneous or biased or it can't be utilized in your  statistical software package, then you're not going to have usable data to guide and answer research questions.  And so, again, that leads to the quality of the data, leads to the quality of registry-based research.  So, kind of think of it as, you know, if it's garbage data in, you're going to have garbage analysis and garbage data out.  So, you need to consider all these things when, you know, determining the type of data that you need to have entered into the registry.  So, that's been kind of an overview, in general, of some of the elements to consider with developing, implementing, and maintaining registries.  And so, I'll turn it over to Dr. Wachey at this time for some more in-depth exploration of our experience with the virus registry from Discovery. DR. WACHEY  Thank you so much, Scott. Thanks for the wonderful presentation and overview of the general concepts about forming a registry.  And thank you to SCCM as well for inviting me to present today. So, while Scott presented the general concepts about a registry, we're going to take all  of those wonderful things Scott taught us and apply them to what we did with the virus COVID-19 registry.  So, we're going to sort of take it step-by-step through the virus registry and discuss all of those concepts. So, first, what is the virus COVID-19 registry?  I think it's mostly in the name.  The virus is an acronym for Viral Infection Respiratory Illness Universal Study, I believe.  And what it is, it's a platform to study COVID-19. So, this became an international consortium of ICUs collecting data on patients with  COVID-19 fairly rapidly in the early spring or late winter of 2020. But I'll sort of take you through how that happened.  I think going through the step-by-step formation of the virus registry will help give a clear picture to ways that one might approach starting and maintaining a registry.  So, the key thing to note about the virus registry is that this was enabled by the Society of Critical Care Medicine Discovery Network by hosting the registry as well as  by allowing all the logistics and administration and providing a tremendous support for the registry really every step of the way and continuing on to this day.  The Discovery Network, as most of you know, was founded in 2017 with really the goal to harmonize and validate data collection across different hospitals.  And so, the COVID registry was aligned with that goal.  But how did the COVID registry start? Well, work started in March of 2015 as COVID was beginning to rear its head across the world and really starting to come into North America.  I sent a tweet that said, wouldn't it be great to have a COVID registry to answer important non-RCT questions? Anyone thinking about this?  I think I suggested a research question at that time.  And Ogigayek responded, Dr. Kashyap is starting one. Send him an email. So, that was on March 15th and I did send him an email that day.  He said, call me. We spoke. I remember that was a Sunday afternoon and we spoke and came up with a plan that really encompassed much of what you saw today.  And so, we'll go through what that plan entailed.  So, first was to think about what's the purpose of this registry, right? The vision. That was the first thing that Scott was talking about for how you might envision a registry.  So, we came up with some draft aims that we wanted to create as near real-time as possible  COVID data and a dashboard that showed what's the current state of COVID in our ICUs and also what's the current state of the registry.  We had a second aim to study what is the variation in practice, the approach to COVID across hospitals.  And then lastly, to start to evaluate some of the safety and effectiveness of some COVID practices.  Remember, in March 2015, we had pretty much zero data to inform any COVID-specific practice. And so, it was really the Wild West there.  And so, starting to study what people were doing and how they started to change what they were doing over time was one of the main goals of the registry.  Essentially, having a way that the world could see what is happening with COVID in hospitals, but as well as the response to COVID by clinicians.  We did not collect biospecimens. That was a decision made early on with the vision that this wasn't going to be sort of a biospecimen early translational research resource,  but really more of a clinical practice late translational research. So like the evidence to practice end of translation.  So the vision, as I had just stated, was to create an international registry of hospitalized patients with COVID-19 to understand epidemiology, natural history,  practice variation, and quality of care for COVID.  These are the things that Dr. Kashyap and I discussed that Sunday. Who would the team be? We agreed that as Dr. Valeska was saying that we needed a diverse team  with lots of different backgrounds. So we needed clinicians, statisticians, informaticians, epidemiologists, people well-versed in regulatory management  of clinical data, administrators, and data scientists. And so we began to put together those people really over the next few days.  We needed to outreach to sites. So the registry couldn't just be myself and Dr. Kashyap.  We needed more. So as the initial call for establishing a registry  went out over Twitter, the calls to bring in new sites went out over Twitter  as well as through the Society of Critical Care Medicine networks and channels of communication. So we started to outreach to gain different sites.  So this could be a multicenter registry reflective of that U part, the universal practices across the world.  We really the next day, I remember getting to work. So that Monday of March 16th, getting to work on some of the regulatory documents,  starting to write our institutional review board application  for collecting the data, starting to work on data use agreements,  starting to work on the infrastructure. So Rahul, Dr. Kashyap started working on the REDCap data entry and that necessitated  coming up with what we would be collecting in terms of the data. So coming up with our case report forms and then how we might clean and code the data.  And then lastly, thinking about the logistics, how would we meet to make sure that our vision for the registry was being carried out  and also to get feedback from registry participants of what could be improved, what things were unclear and to help and coming up with standard operating procedures  that were clear enough for everyone to follow.  We also began to think about publishing. So what should be the processes for applying  to use the data for research? And then what should be the processes for who should be an author on a study? And we put a lot of thought into that.  So all this work upfront really occurred in the first couple of weeks of thinking about the registry. And I think like most things,  putting in all this upfront work pays off at the end.  So we thought about what types of data should be collective. And this is a screenshot from REDCap, which was chosen as the data entry software,  which is developed by the CTSA in NCATS and NIH.  And so we had divided up the different types of data, as you can see, both by type of data and then timeframe in which it was to be collected.  And by dividing these up into smaller parts,  we think our intent was to have the data collection be more approachable, more feasible for people. Because if you put all of these data collection instruments  together, it was hundreds and hundreds of fields and many hours of work per patient. But what we thought we would do was separate these out  and then stress to people that we are primarily, at this point of the pandemic, concerned with core data one and two. Or, and now at this phase,  please focus on core data five or core data six. So that we had the ability to divide and conquer in many ways the registry by focusing all of the energy  on specific data fields that might be more important at different phases. So this gave some flexibility and divided things up to, we hope and think made things less overwhelming  to collect all of these data that we were interested in.  Here's a timeline of those early weeks that I've talked about. So Rahul and the SCCM Discovery Network started to think about the registry.  And he says March 11th, he saw the tweets on March 15th and we started that day. And by March 20th, I had submitted our IRB at Boston University and Boston Medical Center.  We had designed the first draft of our case report form, all of the information we wanted to collect. We had started to send invitation to sites to participate.  We had created a social media account to begin to network and bring in new sites and had already set up a frequently asked questions page.  So this was a huge amount of work in five days time. But I think at that time we were really just driven by the mission that this was something that was hugely important,  impacting the world that we knew nothing about. And it was really just a great, I think universal drive by everyone we had contacted to try to help in some way.  Everyone wanted to contribute to help, which was wonderful. And really the only way that this could get started. And our IRB also shared in that vision  because it's probably the fastest IRB approval I've ever heard. They approved our IRB application to start collecting data from this registry within three days. So by on March 23rd.  And so we had the first draft of our case report form. And in that three days, now a website set up. So again, incredibly quick this infrastructure going up  such that by March 31st, the registry was launched and we had 16 sites that had IRB approval. We had a standard operating procedure for how and what we should be collecting.  We had this logistics, the orientation call scheduled with weekly investigator calls. And we had 10 patients already enrolled in the registry by really two weeks after the thinking  about forming a registry. So this was really just a rapid infrastructure that was put together that then over time we started to refine.  So that was within two weeks, we had greater than 250 centers begin to become enrolled. And this is a screenshot. I think I took it from last week.  So we had made a dashboard for the registry that showed everyone who wanted to look on the website, sccmcovid19.org, what was happening in the registry.  So you can see in this figure what the registry has become and who's participating. These are yellow dots across the globe of sites that are contributing data to the registry.  I think they represent every continent except Australia and Antarctica. And I think while Antarctica gets a pass, Australia, we would love to have them on board.  And it shows we had 306 ICUs contributing across 28 countries and have now more than 80,000 patients data entered including almost 20,000 patients from the ICU.  So that we thought the dashboard would be helpful to see for people to see what was happening in the registry. And the dashboard was a way both I think  to provide quick feedback to sites who are participating  to show them where their hard work was going to show how quickly data was being entered  really as a way to get feedback, right? So sites are saying, we're putting in all this work where's it's going? The dashboard was there to show here it is.  And I think that was helpful in getting people to continue to participate when the sort of initial drive to respond to COVID started to wear down  as people started to wear down from all of the work. The other reason for the dashboard was really just to give the public a view of what was happening with COVID,  what was happening with treatments for COVID, who was, what were patients getting? It allowed hospitals to give a pretty quick look at what other hospitals were doing.  I know early in COVID, our hospital was not using non-invasive ventilation and high flow nasal oxygen because of a fear of aerosolization  and unclear effectiveness. And I think, just seeing that, oh, look, 25% of patients are getting this and we're not offering it, should we rethink this, right?  So just seeing what was happening in practice was really powerful. Of course, not only just what was happening in practice, but what was happening in outcomes.  So we could, hospitals could start to benchmark themselves against other hospitals and start to audit their own practice and say, you know, in the registry,  patients are staying in the ICU for a median of seven days, but at our hospital, it's way longer. Let's explore why, right? Or the mortality rate in the ICU is 32%. What about us?  So it gives a really, just a benchmark for other hospitals to start to explore their own practice.  Not only was the dashboard helpful, but we had really, the intent of the registry was to provide data for research purposes. And so that was the goal at the beginning,  as you could remember, if you remember my initial tweet, it was proposing a research project about ACE inhibitors and actually that was submitted as a ancillary study  and I think is in the review stage or is somewhere along the line, but that study has been done from our data, which was, I think, really gratifying to see,  from the tweet that someone had actually done that study and it wasn't me, which was good.  So we've had more than a hundred applications to do research using the registry. Those applications went through the process that we had outlined early on  in which people submitted an application. The applications were reviewed by a team of researchers and clinicians and scored and given feedback.  And some people's applications went through on the first try and they were able to get the data, but others that needed a little bit more work  were reviewed by experts in research and given feedback and improved over time and then given the data when they met the sort of threshold of standard  that we had set for research with the registry. I believe at this point, there's about 20 papers that are either in review or already published and you can go to the SCCM website  and scroll through all the papers that have been published so far from the registry.  So this is one example of research that was really just basic epidemiology about the practice. This was a paper that was first authored by J.P. Domek from Mayo Clinic  and this was published in Critical Care Medicine. So this paper was really just showing people what outcomes were for patients with COVID who were hospitalized,  which is really what our registry was.  And so it gave benchmarks. Like I said, they were appearing on our dashboard, but also here was those benchmarks provided in much more detail  across much more finer divisions of types of patients, age and locations and things like that. And so one of the other things we showed with this study  of the outcomes of patients with COVID-19 who required organ support was that for mechanically ventilated patients, depending on the hospital you were admitted to,  there was a wide range of mortality rates. And so that brought up questions as to what types of things might be associated with better or worse mortality across different hospitals.  So this risk-adjusted hospital mortality range was large. And so that motivated additional studies using the registry data.  The registry was not only used for research epidemiology,  but also for quality improvement. So Ogi Gayek and Alex Niven started this STOP virus  quality improvement part of the registry where sites could enroll and have their data fed back to them, again, for audit and feedback and benchmarking,  and also participate in a curriculum for critical care learning. This is just some examples of some of the different lectures that were part of the STOP virus curriculum.  And this STOP virus QI study is now, you know, being turned into other quality improvement curricula across the registry.  And finally, the registry is being used to look at preliminary effectiveness studies, so comparative effectiveness in observational data. So this is just one example on the website  that shows a study of pre-hospital aspirin use and outcomes.  So that was an overview of how we took the concepts that Scott had presented and turned them into the virus registry. But at this point, the virus registry is starting  to wind down for data collection, and we're starting to fill up the empty fields,  look at missing data, finish the cleaning, and think about what are the next steps for the registry. And so in addition to finalizing the registry,  getting it as complete as possible, and starting to think about opening it up to a wider range of participants in terms of data use, we also are thinking about ways  that that infrastructure that the registry formed, you know, this coalition of hospitals across the world that were contributing a harmonized data set,  how can we continue to use that to benefit our patients? So the next steps are really a partnership between the Society of Clinical Medicine,  the Food and Drug Administration of the U.S., the CureID Group, and NIH to start to better automate data collection from hospitals, so not for everyone to manually input data,  but the data to be pulled directly from the electronic health record, to form a coalition of sites with the capacity to quickly and accurately collect data,  to develop and validate methods  that will automate this extraction, to identify common data sets that will automate data collection, really with the goal to start to use the registry  not just for epidemiology and observational studies, but as a platform for rapid, efficient, large-scale pragmatic trials. So if patient characteristics and outcomes  can automatically be uploaded from these registries, then we can do very, very large trials testing different standards of care against each other. So that's really the next goal.  In summary, clinical registries require a clear vision, a coalition of willing sites, low data-entry burdens, and explicit procedures for data use and for research.  Registries are really a wonderful opportunity to form new professional connections, construct new research infrastructure, and conduct, hopefully, in the end, great science.  And really wonderfully, the SCCM is a resource that has great experience and willingness, as shown with the virus registry, to support and assist with making and conducting  and using registries. So thank you very much. We're happy to answer your questions. Thank you, Scott and Alan. We have a question regarding what's the funding  for the virus registry.  So I can answer that, partly, I think, but also happy to have others weigh in. Initially, at the very beginning, the funding was the support  of the Society of Critical Care Medicine's infrastructure, sort of, I think, essentially donated for the purpose of the registry. Early on, we applied for funding  from the Gordon and Betty Moore Foundation, who provided initial support. That was support for the first year. There was also a brief partnership with Janssen,  who provided additional support early on. And that was the main early funding. But at the beginning, it was a volunteer effort for all,  until some funding began to come in to support people. Anyone want to add anything else that I'm missing, funding-wise? No, I don't have anything to add, Alan.  I think that's the funding sources I recall. Was it funding from the CDC, as well, at one point? I can't remember. That's right. I think the CDC funded  the Stop Virus Quality Improvement Initiative.  Thank you. It's a different question. What would you guys say to be the biggest hurdle in terms of setting up a multicenter research registry?  I'm happy to take the first stab at that. I think probably everyone who worked on the virus registry might have a different opinion of what the biggest hurdle was,  because there's lots of hurdles, and it's definitely a difficult process, but it's not impossible, and it's very possible. But I think, for me, some of the more difficult hurdles  are the initial steps of getting sites. So that's one hurdle. So getting people to consider being in the registry is not hard. Getting people to actually be in the registry,  to submit an IRB application, to have their legal team go through a data-use agreement, to iteratively change the data-use agreement  so that everyone agrees of what it should say for every site is a big hurdle, and that requires people with skills that are very different from mine.  So that requires generally a legal team. One area where SCCM and Mayo Clinic, who also helped with all of that, played a big role. So that's a hurdle.  I think, for me, the second big hurdle is making sure that the vision you have for each data field is what is being transmitted on the other end  through like REDCap or through the case report form. So different interpretations of what data should be entered I think was an initial hurdle as well. Those types of things,  like how should length of stay be entered? So you could say length of stay, and people type in 14, and is it hours or days? You know, all of the units, all of those things,  just being incredibly detail-oriented for every field. I think what helps with that is, instead of having free text entry at any point,  is to have like drop-down menus that allow choices because those choices, then you are limiting what can be entered, and by you limiting that centrally,  you're providing an initial level of quality, sort of quality assurance or standardization rather than people being able to enter free text because if people enter free text,  they're going to use periods or commas interchangeably to mean different things and use different units, but if you make them click on something,  that can only mean really the one thing. So those are the two things that I would say. What about Scott? Did you have additional thoughts? Yeah, and I would agree with Alan.  So I was brought into the project with probably those first five days or so to review the case report form for their medication therapy content  in the field that wanted to gather data on medication therapy,  but, you know, was kind of roped in a bit more involvement from that. So I would say, yeah, I agree with Alan, but, you know, the logistics in terms of, you know,  recruiting sites, getting sites to participate, getting sites to complete the data fields  that were necessary and using, entering the data in a useful manner was probably one of the obstacles that I saw. And that probably didn't present itself up front.  Again, there was a lot of, as Alan explained, time crunch to get the CRFs developed early on and so the registry could open in a timely manner.  And some of that was, you know, hindsight in looking at, you know, how that entry fields were developed in the REDCap CRF and so forth. So, you know, from a usability standpoint,  again, those small details in terms of how data is intended to be entered to be useful are very important to consider up front.  And again, bringing in your various members of your team to look at those aspects and then having a statistician or biostatistician looking at the field and saying,  oh, you know, it would really be better if we had that entry for this field, you know, be numeric only or dropdown menus or radio buttons as Alan suggested and so forth.  I would think those would be, you know, those probably aren't the big hurdles, but those are the hurdles that, you know, may rear themselves without really good pre-planning  and so forth.  So I think the biggest hurdle was the time crunch early on, you know, for me in terms of trying to get everything developed, reviewed within those, again,  within about two weeks time, a little more than two weeks time to get the registry open for data entry. So yeah, very, very interesting experience for me personally.  It's also a common scene that for the virus registry, there were over 500 sites who signed up, but not everybody went through the ROB and the DUA.  Right. Some sites, you know, as you can imagine, it's a pandemic and people are incredibly busy and busier than they've ever been in many ways.  And so asking people who are more busy than they've ever been to do more is a tough ask. So I think the fact that we were able to get,  you know, 300 something sites to sign up out of the 500 that initially volunteered, I think that's good. You know, I think the data use agreements and IRB applications is work.  And so not everyone is prepared to do that work for a variety of reasons. And I think that's just expected. And that's why casting such a wide net,  I think was beneficial that really asking anyone to join  who is willing was one of the approaches to try to get as broad a sample as we could, knowing that not everyone would be able to do it.  Another question I was posting is asking, what factors do you think contribute to this huge success for the virus registry?  Yeah, just a good team, you know, that those initial days bringing together the right group of people, and it wouldn't have been able to have been accomplished  without the backing of Society of Critical Care Medicine as well, you know, as a key and critical members of that team. So having a large organization behind you doesn't hurt.  So I think for people looking to start a registry, looking to partner with an organization, you know, whether it's Society of Critical Care Medicine  or some other appropriate organization is helpful, right? Because that's already a huge infrastructure of multi specialties across critical care patients.  You know, it's all sort of part of SCCM already. So they're able to really help provide that logistics and infrastructure needed.  And in addition to that, what Alan said, you know, certainly the support of SCCM, I would think, you know, a lot of the public facing elements  that Alan reviewed, you know, the dashboard, you know, to readily disseminate real time data  to not only the public and SCCM members, but, you know, to participating sites as well, to, again, as he mentioned during his lecture,  to show them what their work and effort is going towards, you know, in real time. I think a lot of those public facing readily accessible elements and the data on the dashboard,  et cetera, really helps with, you know, recruitment and motivation and maintenance of data entry for the registry. Thank you.  Thank you to both of you. This concludes our Q&A sessions for today. Thank you, Dr. Walke and Dr. Balesta.  Thank you for having me. Thank you everybody for attending today's webcast. Again, this webcast is being recorded and the webcast will be available to everybody  within five business days. To access the webcast, log into mysccm.org and go to the My Learning tab to access the recording. And this will conclude our presentation for today.

Video Summary

The webcast discusses the process of creating a multi-center research registry, using the example of the VIRUS COVID-19 registry. The registry was formed in response to the COVID-19 pandemic and aimed to collect data on patients with COVID-19 to understand epidemiology, natural history, practice variation, and quality of care. The registry was made possible through the support of the Society of Critical Care Medicine Discovery Network. The process of creating the registry involved developing a clear vision and goals, assembling a diverse team of experts, reaching out to sites to participate, addressing regulatory issues, creating infrastructure for data collection and storage, and establishing logistics for ongoing communication and maintenance of the registry. The registry used REDCap as the data entry software and had a dashboard to provide real-time information on the current state of COVID-19 in ICUs and the progress of the registry. The data collected in the registry was used for research, quality improvement initiatives, and preliminary effectiveness studies. The registry faced challenges in obtaining site participation and ensuring consistent data entry, but the successful creation of the VIRUS COVID-19 registry demonstrates the value and potential of multi-center research registries in advancing knowledge and improving patient care.

Asset Subtitle

Research, 2022

Asset Caption

The creation of a multicenter registry to facilitate rapid data aggregation in response to COVID-19 has been shown to effectively disseminate real-world experience and generate hypotheses to inform larger trials. The benefits of this mechanism of learning more about diagnostics and identifying potential therapeutics are not limited to COVID-19. This webcast will introduce the organization and creation of multicenter registries and key lessons learned from COVID-19 that could potentially be used not only for future health crises, but also to improve coordination for routine critical illness research. Topics include coordination and logistics, data dictionaries across heterogeneous institutions, and legal implications.

Meta Tag

Content Type Webcast

Knowledge Area Research

Knowledge Level Intermediate

Knowledge Level Advanced

Membership Level Select

Membership Level Professional

Tag Research

Year 2022

Keywords

multi-center research registry

VIRUS COVID-19 registry

COVID-19 pandemic

epidemiology

data collection and storage

REDCap

dashboard

ICUs

improving patient care

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.