www.science.org
Open in
urlscan Pro
104.18.34.21
Public Scan
Submitted URL: https://news.google.com/rss/articles/CBMiZkFVX3lxTE8ybnJDeXo0U09ZOWlWMGZreUJLN3J4a1MtclM5cDM0LVdLaVQ0QXpoR0RLZVBLUUVOZmR...
Effective URL: https://www.science.org/content/podcast/saving-wildlife-ai-and-randomized-trials-go-remote
Submission: On August 27 via api from DE — Scanned from AU
Effective URL: https://www.science.org/content/podcast/saving-wildlife-ai-and-randomized-trials-go-remote
Submission: On August 27 via api from DE — Scanned from AU
Form analysis 8 forms found in the DOM
Name: defaultQuickSearch — GET /action/doSearch
<form action="/action/doSearch" name="defaultQuickSearch" method="get">
<fieldset>
<legend class="sr-only">Quick Search anywhere</legend>
<div class="input-group option-0 animation-underline"><label for="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac010" class="sr-only">Enter Search Term</label><input type="search" id="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac010" name="AllField"
autocomplete="off" placeholder="Enter Search Term" data-auto-complete-max-words="7" data-auto-complete-max-chars="32" data-contributors-conf="3" data-topics-conf="3" data-publication-titles-conf="3" data-history-items-conf="3" value=""
class="quick-search__input form-control autocomplete"></div>
</fieldset>
</form>Name: publicationQuickSearch — GET /action/doSearch
<form action="/action/doSearch" name="publicationQuickSearch" method="get">
<fieldset>
<legend class="sr-only">Quick Search in Journals</legend>
<div class="input-group option-1 option-journal animation-underline"><label for="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac011" class="sr-only">Enter Search Term</label><input type="search" id="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac011"
name="AllField" autocomplete="off" placeholder="Enter Search Term" data-auto-complete-max-words="7" data-auto-complete-max-chars="32" data-contributors-conf="3" data-topics-conf="3" data-publication-titles-conf="3" data-history-items-conf="3"
value="" class="quick-search__input form-control autocomplete"><input type="hidden" name="SeriesKey" value="science"></div>
</fieldset>
</form>Name: publicationQuickSearch — GET /action/doSearch
<form action="/action/doSearch" name="publicationQuickSearch" method="get">
<fieldset>
<legend class="sr-only">Quick Search in Journals</legend>
<div class="input-group option-2 option-journal animation-underline"><label for="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac012" class="sr-only">Enter Search Term</label><input type="search" id="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac012"
name="AllField" autocomplete="off" placeholder="Enter Search Term" data-auto-complete-max-words="7" data-auto-complete-max-chars="32" data-contributors-conf="3" data-topics-conf="3" data-publication-titles-conf="3" data-history-items-conf="3"
value="" class="quick-search__input form-control autocomplete"><input type="hidden" name="SeriesKey" value="sciadv"></div>
</fieldset>
</form>Name: publicationQuickSearch — GET /action/doSearch
<form action="/action/doSearch" name="publicationQuickSearch" method="get">
<fieldset>
<legend class="sr-only">Quick Search in Journals</legend>
<div class="input-group option-3 option-journal animation-underline"><label for="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac013" class="sr-only">Enter Search Term</label><input type="search" id="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac013"
name="AllField" autocomplete="off" placeholder="Enter Search Term" data-auto-complete-max-words="7" data-auto-complete-max-chars="32" data-contributors-conf="3" data-topics-conf="3" data-publication-titles-conf="3" data-history-items-conf="3"
value="" class="quick-search__input form-control autocomplete"><input type="hidden" name="SeriesKey" value="sciimmunol"></div>
</fieldset>
</form>Name: publicationQuickSearch — GET /action/doSearch
<form action="/action/doSearch" name="publicationQuickSearch" method="get">
<fieldset>
<legend class="sr-only">Quick Search in Journals</legend>
<div class="input-group option-4 option-journal animation-underline"><label for="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac014" class="sr-only">Enter Search Term</label><input type="search" id="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac014"
name="AllField" autocomplete="off" placeholder="Enter Search Term" data-auto-complete-max-words="7" data-auto-complete-max-chars="32" data-contributors-conf="3" data-topics-conf="3" data-publication-titles-conf="3" data-history-items-conf="3"
value="" class="quick-search__input form-control autocomplete"><input type="hidden" name="SeriesKey" value="scirobotics"></div>
</fieldset>
</form>Name: publicationQuickSearch — GET /action/doSearch
<form action="/action/doSearch" name="publicationQuickSearch" method="get">
<fieldset>
<legend class="sr-only">Quick Search in Journals</legend>
<div class="input-group option-5 option-journal animation-underline"><label for="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac015" class="sr-only">Enter Search Term</label><input type="search" id="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac015"
name="AllField" autocomplete="off" placeholder="Enter Search Term" data-auto-complete-max-words="7" data-auto-complete-max-chars="32" data-contributors-conf="3" data-topics-conf="3" data-publication-titles-conf="3" data-history-items-conf="3"
value="" class="quick-search__input form-control autocomplete"><input type="hidden" name="SeriesKey" value="signaling"></div>
</fieldset>
</form>Name: publicationQuickSearch — GET /action/doSearch
<form action="/action/doSearch" name="publicationQuickSearch" method="get">
<fieldset>
<legend class="sr-only">Quick Search in Journals</legend>
<div class="input-group option-6 option-journal animation-underline"><label for="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac016" class="sr-only">Enter Search Term</label><input type="search" id="AllField9ea9b667-c3ce-40d1-870b-f7e001bbac016"
name="AllField" autocomplete="off" placeholder="Enter Search Term" data-auto-complete-max-words="7" data-auto-complete-max-chars="32" data-contributors-conf="3" data-topics-conf="3" data-publication-titles-conf="3" data-history-items-conf="3"
value="" class="quick-search__input form-control autocomplete"><input type="hidden" name="SeriesKey" value="stm"></div>
</fieldset>
</form>Name: defaultQuickSearch — GET /action/doSearch
<form action="/action/doSearch" name="defaultQuickSearch" method="get" id="quick-search-from-e7029926-2304-4dc3-a2d3-f6c677f7bbc70" role="search">
<fieldset>
<legend class="sr-only">Quick Search anywhere</legend>
<div class="input-group option-0 align-items-center animation-underline animation-underline--primary"><label for="AllFielde7029926-2304-4dc3-a2d3-f6c677f7bbc70" class="sr-only">Enter Search Term</label><input type="search"
id="AllFielde7029926-2304-4dc3-a2d3-f6c677f7bbc70" name="AllField" autocomplete="off" placeholder="Enter Search Term" data-auto-complete-max-words="7" data-auto-complete-max-chars="32" data-contributors-conf="3" data-topics-conf="3"
data-publication-titles-conf="3" data-history-items-conf="3" value="" class="quick-search__input form-control autocomplete">
<div class="input-group-append"><button type="submit" title="Search" class="btn quick-search__btn p-0"><i class="icon-search align-middle"></i></button></div>
</div>
</fieldset>
</form>Text Content
Advertisement * * news * careers * commentary * Journals * * * Log in * Become A Member science science advances science immunology science robotics science signaling science translational medicine science partner journals Quick Search anywhere Enter Search Term Quick Search in Journals Enter Search Term Quick Search in Journals Enter Search Term Quick Search in Journals Enter Search Term Quick Search in Journals Enter Search Term Quick Search in Journals Enter Search Term Quick Search in Journals Enter Search Term Searching: Anywhere AnywhereScienceScience AdvancesScience ImmunologyScience RoboticsScience SignalingScience Translational Medicine Advanced Search Search TRENDING TERMS: * cancer * climate * artificial intelligence * postdoc * aging Log In Become A Member Donate Quick Search anywhere Enter Search Term science.org * news * careers * commentary * Journals * science * science advances * science immunology * science robotics * science signaling * science translational medicine * science partner journals * Custom publishing * newsletters * collections * videos * podcasts * blogs * visualizations * prizes and awards * authors & reviewers * librarians * advertisers * about * help * * * * * * * Terms of Service * Privacy Policy * Accessibility HomePodcastsSaving wildlife with AI, and randomized trials go remote Back To Podcasts * Podcast SAVING WILDLIFE WITH AI, AND RANDOMIZED TRIALS GO REMOTE * 22 Aug 2024 * 2:00 PM ET * By Sarah Crespi, Christine Peterson Share: * Facebook * Share on X * Linked In * Reddit * Wechat * Whatsapp * Email Gerrit Vyn/NPL/Minden -------------------------------------------------------------------------------- Play Pause Go ten seconds backward Go ten seconds forward Mute Unmute 00:00 28:33 Full Transcript Download Subscribe Apple Spotify Overcast Pocket Casts -------------------------------------------------------------------------------- Related article To reduce wildlife deaths caused by fences, scientists are turning to AI BY Christine Peterson News 16 Aug 2024 First up this week on the show, uncounted kilometers of fences are strung across the globe. Researchers know they interfere with wildlife migrations and sometimes make finding food and safety difficult for animals. But they don’t know where all these fences are. Freelancer science journalist Christine Peterson joins host Sarah Crespi to discuss how artificial intelligence and aerial photos could help create fence inventories and eventually reopen spaces for native species. Next, Azizi Seixas, interim chair of the University of Miami Miller School of Medicine’s department of informatics and health data science and a professor in the department of psychiatry and behavioral sciences, discusses his review on decentralized randomized trials. Randomized, controlled trials based in a research center or centers have long been the gold standard for determining the effectiveness of a medical intervention. This week on the podcast, Seixas argues that distributed research designs with home-based measurements and reporting have the potential to speed up research, allow greater participation, and make the results of studies more equitable. This week’s episode was produced with help from Podigy. About the Science Podcast -------------------------------------------------------------------------------- TRANSCRIPT 0:00:05.8 Sarah Crespi: This is the Science Podcast for August 23rd, 2024. I'm Sarah Crespi. First up this week, countless kilometers of fences cross the globe, locking animal migration, sometimes penning wildlife away from food and safety. Wyoming-based freelance writer Christine Peterson joins me to discuss how conservationists can find the fences by using AI and aerial photography and then make a plan to open the land once again for native species. Next up on the show, Azizi Seixas talks about decentralizing clinical trials. Instead of having participants commute into a research center, they are equipped with the supplies and digital access they need to take part in research. The intent is to make participation easier, trials speedier and more true to life. This week in science, Wyoming based freelance journalist Christine Peterson wrote about a novel use of artificial intelligence for conservation purposes, harnessing AI to find fences in the western United States. Hi Christine, welcome to the Science podcast. 0:01:15.1 Christine Peterson: Hi Sarah, thanks so much for having me. 0:01:17.5 SC: Oh sure. So can you set up the fence problem for us? Like why would we need to find them? How are they lost, and why is it bad that we have fences? Or why would scientists wanna find them? 0:01:28.8 CP: This is one of those funny hiding-in-plain-sight things. As someone who's been in Wyoming for decades, fences are everywhere around the West, I'm staring at them right now. And they're really useful, they keep cows where they belong, they keep sheep where they belong, they separate property lines, they've been here for centuries, but unlike roads, where agencies like the Forest Service or the Bureau of Land Management or even private landowners, everyone has kept track of where roads are. Nobody has ever kept track of fences, and so they just are built when they're needed and sometimes they start to fall down, and more are built in the same place, or they're built out of all kinds of different materials that were available at the time. So now we are in a place where no one has any idea how many fences there are, but when there have been a few number of surveys done, one notable one in a county called Sublette County in Wyoming, where the researcher mapped 7000 kilometers of fence just in that one county, which as a reference is the length of the US-Mexico border out and back. So it's double the length of the US-Mexico border just in one Wyoming county. The problem becomes that every fence that an animal like a pronghorn or a mule deer or an elk has to cross is a hurdle. So that's just one more barrier to them getting to where they need to be. 0:02:55.7 SC: Are there any good examples of the impact of these stray fences or abandoned fences on the animals that we're talking about? 0:03:02.1 CP: Yeah, there was, and this is in some ways the genesis of this story. But the last winter, winter of 2023, was particularly harsh in stretches of Wyoming, it was really snowy, it was very cold and very windy, and so there were periods of... A lot of times in a winter, there are periods of reprieve, where some of the snow melts and animals can get to ridge lines and can find a little bit of food, that kinda thing. And there was a stretch of the red desert, which is a big chunk of land in Central Western Wyoming, where collared pronghorn... There were 45 pronghorn that had been collared for a project, started to die at the end of the winter. And ultimately about half of the pronghorn died, thousands of pronghorn in the overall herd height died. And when the researcher went out and collected the collars that had fallen off, or still attached to these pronghorn that had died, he uploaded the data and realized that they had just been bouncing around. 0:04:01.1 CP: The maps almost looked like a ping pong ball, where they're just bouncing back and forth and back and forth and back and forth up against fences. And some of the fencing is along interstate, some of it's just pie traffic on the interstate, and then a lot of it's just these fences in the middle of nowhere. And some of them don't really need to be used anymore at all probably and some of it still could be used, so there's efforts to replace it. But what he found was one individual pronghorn which really stuck out had wandered well over 400 kilometers over the course of a winter looking for somewhere to go to escape the snow but never managed to get more than about 50 kilometers away from where it started. So it had just walked and walked and walked and walked and never really got anywhere 'cause it was blocked all over the place. 0:04:48.1 SC: Sounds like it would be really difficult. Talking about 7000 kilometers in one county alone, there's a vast expanse on the western end of the United States. I can just imagine the fences adding up and adding up. How did this research that you heard about from the Senckenberg Biodiversity and Climate Research Center and the Microsoft AI for Good Lab... How do researchers from there try to use AI to solve this problem? 0:05:14.4 CP: The project is in its very early stages, but the proof of concept is there, which is what they presented at The Ecological Society of America conference last week. For these initial stages, they used pictures taken from airplanes that were at a resolution of 60 centimeters per pixel, which was enough to allow AI to begin to identify fences from the air. So basically, then AI is able to scan all of these images at rates much higher than a human could ever do to then look for where fences are and where fences are to try to start building an inventory. 0:05:51.7 SC: This is very early stages and the idea is to make a catalog, what would be the next step? Would you then decide which ones to take down or who would have to do that even? 0:06:05.0 CP: There are already efforts underway. There's a fair bit of money right now looking at either taking down fences or also replacing them with wildlife-friendly fencing. There's a lot of fence that's called woven wire. It essentially looks like a whole bunch of squares in the bottom. And for a species like pronghorn that prefers crawling under fences or mule deer, which about half the time prefer crawling under fences, that's just impenetrable, they can't get through. And so then in areas where you still need fence, 'cause you still have cows that need to be contained or sheep need to be contained, you can replace that woven wire with single strands that are a certain amount above the ground and that allows pronghorn and mule deer to go underneath. So what happens with this data then is it goes to land managers and conservationists, landowners to work together to try to figure out and identify, Okay, here clearly these fences are problematic and, do we need them anymore? Can we take them out? Do we just need to replace them with fence that are easier for wildlife to cross? 0:07:11.8 SC: Right. So you can scope the project, you can figure out what the priorities are. 0:07:15.9 CP: Yep. 0:07:16.4 SC: All that stuff. That's great. So I could see that they could make these like large interconnected areas so the animals are able to move into spaces where food might be more available in different parts of the winter. 0:07:29.2 CP: For sure. Yep. There is already a really large effort underway that has been going on for some years now to map migration routes for wildlife and that especially in big landscapes like the American West but also these big rangeland, grassland systems all over the world in Tibet, Australia, chunks of Africa, and those starting to get an idea of where animals are moving would like to move. And then you can go in presumably with something like a fence inventory and say, Okay, this is clearly an important migration route and we've got 27 fences that cross it, how can we work to make some of that easier for wildlife? 0:08:07.2 SC: We haven't actually talked about satellites. This initial study was using photos from airplanes, but we do also have access to satellites, imagery of the land. Would that be something that could be adapted to conducting a fence inventory? 0:08:20.9 CP: Absolutely. Yep. That's where they hope to go next, is to start using satellite images. Then they can expand it out more broadly around the West and even around the world. 0:08:30.1 SC: Is this abandoned fence problem a universal thing? Is it not just the American West? 0:08:35.8 CP: Yeah. And I guess just for the note, the abandoned fences is a... It's a little bit more because a lot of these fences are still being used, where it's one of those like Elise has. Elise has some property and hasn't maybe run cows there in 10 years but still might want to. Some of them are still certainly in use or could be in use, so it's not just abandoned fence, it's just fence in general. But yes, there are fence... And there are fences going up... Still a lot of fencing going up in a lot of places in the world, and so this is definitely a global problem. And it's something that I think a lot of people really wanna work on because it's a conservation issue that has some solutions that especially if you're replacing it with wildlife-friendly fencing that in some ways is almost a win-win because it allows somebody who needs the fence to still have the fence but because it's been redone in a way that's good for wildlife, it can still allow for some movement. 0:09:35.5 SC: Thinking about AI, it's just this hot topic, it's come up in so many different fields over the past few years. What are some of the other ways that artificial intelligence could be used to help with conservation? What are people looking into there? 0:09:49.7 CP: Yeah, that's a great question. There's a lot going on, it looks like, and a lot of it's still pretty early, but it's everything from one researcher I was talking to, did a huge sound study with birds, he's beginning to use artificial intelligence to then sort that out for him. And one point that a researcher made that I felt was... He feels is very important, is that artificial intelligence isn't... AI isn't going to be replacing human observation, it isn't going to be replacing all the local knowledge and what humans know about the land and wildlife and the environment. It's just a tool to help us be more efficient and in a lot of ways, gather, categorize, inventory this huge amount of data that we have in a way that then becomes useful or more useful for people to start making decisions. 0:10:40.2 SC: I heard the same thing at MBARI, the marine research station now on the West Coast. They have experts that can identify fish and coral and all kinds of stuff underwater from video. But those are incredibly valuable people who can do that and they could train AI to do it and then they just solve the tough cases, the edge cases, and they don't spend all their time being like, Eel, eel, eel. 0:11:05.4 CP: Right. Exactly. 0:11:05.5 SC: It's definitely a possibility to think of it as a partnership that saves that human experience for the tough stuff. 0:11:12.5 CP: Right. Yep. Absolutely. 0:11:14.3 SC: Very cool. All right, Christine, thanks so much for talking with me. This is super interesting. 0:11:19.8 CP: Yeah. Thank you. I appreciate it. 0:11:22.5 SC: Christine Peterson is a freelance journalist based in Wyoming. You can find a link to the story we discussed at science.org/podcast. 0:11:31.0 SC: Stay tuned for a discussion of how decentralized randomized controlled trials could increase participation and improve outcomes. 0:11:44.3 SC: Randomized control trials based in a research center or centers have long been the gold standard for determining how effective a medical intervention is, but in a review this week in science, Azizi Seixas and colleagues argued that decentralized randomized trials are the next step for medicine, that a decentralized approach will speed up research, allow greater participation, and make the results of studies more equitable. 0:12:08.6 SC: Hi Azizi, welcome to the science podcast. 0:12:10.3 Azizi Seixas: Thank you for having me, Sarah. 0:12:12.5 SC: Sure, so can you just start us off with the basics of what is the difference between what we think of as a randomized control trial and a decentralized randomized control trial, the DCT? 0:12:23.1 AS: Yeah. No, thank you so much. Really appreciate the opportunity to speak with you and your audience. So typically, a randomized trial pretty much is where you have different groups and you manipulate different exposures and you track over time how people fare, how do they perform? Typically, those trials are very site-specific and site-based, meaning a participant has to come into a health system or a clinic or they have to go to some research environment that may not reflect the person in the real-world environments. Contrastingly, a decentralized trial allows us to still do the same types of procedures for the most part, but instead of having it being site specific, that participant or patient can do this from the comfort of their homes or their work or wherever. And I think that, as you said, allows us to invite a whole lot more individuals to participate. 0:13:27.3 SC: Yeah. You really think about... There's a lot of barriers to participating in research, and part of it is where you live and how close you are to say a research hospital. If you're rural, if you don't have a car, it's not gonna be easy for you to get it. And that is actually changing the profile of the populations that get involved in clinical trials. 0:13:49.0 AS: It's estimated by the FDA that less than 5% of eligible participants actually engage or participate in a clinical trial, so then the question is, what happens to that 95% and what are the barriers that might be preventing people from engaging? Pharmaceutical companies, researchers, scientists, they want to be able to capture a sample that is representative because they wanna ensure that their findings and results are translatable and applicable to all, for the most part. And decentralized trials, what it allows us to do is to increase the opportunity and the level of access to the folks who say they are unable to participate. So you mentioned transportation. It's noted that the median travel time to a certified clinical site is about two hours. Like who has that time, right? 0:14:50.4 SC: That's a sacrifice, that's dedication, that's commitment. 0:14:54.7 AS: Exactly. 0:14:56.3 SC: And we do want people to do that for science, but it might be asking a little bit much. 0:15:01.6 AS: It is. I always jokingly say to our team, No one wakes up saying, I wanna participate in a research study today. 0:15:09.8 SC: No. They do not. There's also this expectation that by increasing participation, you might have more robust results. So does that address some of the problems we see translating from like clinical trials that have such a control population that are excluding people who can't get to a hospital, for example, so they might have a comorbidity? Is this going to get us better results from these incredibly expensive time-consuming endeavors? 0:15:33.8 AS: We believe it will provide better results, but let me qualify better results. So what we say is that we believe in real-world data, because when you look at one of the many challenges that fraught clinical trials and science in general is the lack of replicability. And so what we argue is that the lack of capturing real-world data may actually explain many of the discrepancies that we see between studies. So real-world data encompasses a wide range of health-related information such as electronic health records, claims data, billing activities. And you can get it from different products, you could get it from registries. Patient-generated data like wearables. And what it allows researchers to do is that typically in very deterministic types of trials, they're oftentimes very sterile, they don't reflect through someone's real world, they're too controlled. And what we're saying is that yes, we can get a little bit more innovative in capturing more noisy data so that we can capture the individual in a 360-degree perspective. And I think that's the advantage of real-world data. So for some of the science folks out there, this will allow you to be able to test confounding effects, artifactual effects. 0:17:06.1 SC: One thing that comes up in the review is the digitization of a lot of this process, so having people in their homes answering questionnaires or vlogging progress or even as you say, wearing something that's monitoring certain vitals. Even though it sounds really important to decentralizing things, it sounds like it might have some challenges for patients or for researchers looking to standardize this. 0:17:32.0 AS: Absolutely. We tried our best to not necessarily present decentralized trials as a panacea that can cure all the ills that fraught clinical trials. By no means. Decentralized trials also has significant limitations as well. And some of these have to do, as you say, the digitization... And if we unpack digitization, you're talking access to these types of digital devices. And what we highlight is a model that we have created at the University of Miami, it's called YOU DECIDE. 0:18:08.5 SC: It's an acronym, just so people will know. Yeah. Okay. 0:18:10.4 AS: It's that acronym. It's YOU DECIDE. Exactly. YOU DECIDE framework addresses issues around digital literacy, as well as digital divide, as well as providing social care solutions, because we know that when a participant or a patient is noncompliant or not adherent to the procedures of a study, it probably has nothing to do with their cognitive ability or level of understanding or any of those types of things. It may have to do with lack of access to a cell phone or a mobile device or lack of broadband and internet. Or they may just be challenged because they're working two, three jobs and they don't have the transportation or they can't pay for their utilities. And so we provide additional supports around those to address those as best as we can. 0:19:01.7 SC: Right. So if you're recruiting people who are 80-plus into your study and you're like, Here's the cell phone app that you need to install and here's the wearable that you need to update your drivers, you're just really setting them up to... 0:19:13.0 AS: To fail. 0:19:14.1 SC: To fail. And so if you have support for that... And similarly, as someone doesn't have stable internet, which we saw a lot during COVID, when students like were, I'm from Indiana. They didn't have a dedicated device that they could use at home for schooling when that was suddenly required of them. That kind of lack of access is also something that needs to be considered when you're digitizing your trial. 0:19:34.2 AS: Absolutely. We recognize as well that when you have ubiquitous assessment and engagement with participants, it can lead to participant burden and engagement, that people will tap out. Are you gonna be sending another text message? 0:19:49.4 SC: Oh no. Yeah. 0:19:51.4 AS: People will tune out. That's another thing. And we also describe the fact that there are decentralized trials, but then you also have a subset of digitally facilitated trials as well. 0:20:04.1 SC: Like a hybrid. 0:20:05.1 AS: Exactly. And they're not the same. And so one of the things that we speak about is the lack of standardization in how the procedure of decentralized trials should go. And we highlight this in seven core stages of research, from recruitment to screening, to informed consent to randomizing and assigning people, and then exposure to intervention and monitoring and providing incentive and study debrief. 0:20:34.9 SC: Yeah. I do wanna ask you about this Tufts study that you mentioned, Tufts, the university in Boston that mentioned about like evaluating how decentralized trials fit in the big picture of pushing things from the lab to the clinic to the public. Can you talk a little bit about that? 0:20:48.4 AS: So this was a study where they wanted to look at the return on investment. How do you make the business case for decentralized trials? One of things that you want to be able to establish is whether or not if the decentralized trials can help to reduce costs. And that's huge, right? 0:21:09.1 SC: Because we're talking about maybe giving people hotspots and cell phones and training them on things. Is that saving you money rather than having a clinical center with staff that just come in and type it in the computer for you? 0:21:19.0 AS: Exactly. Typically, when people spoke about providing a value, what's the value add of decentralized trials? Typically, we will always keep it either very practical, from an administrative standpoint, like, It will help with this, it would increase efficiency. And then also people philosophically would say, Decentralized trials, it makes a lot of intuitive sense, who wouldn't agree to the fact that you want to subscribe to some kind of modality that would allow people to participate and reduce barriers? I think what the Turf study did, just to sum it up, is that they were able to provide what the economic investment is and what the ROI was for them in terms of looking at what the net present value is. So they found that... I believe that they found a $10 million return on a $2 million investment in phase two trials, and about a $39 million return on a $3 million investment on phase three trials. 0:22:21.3 SC: And this came out of speeding things up in a lot of cases. 0:22:25.2 AS: We saw this really during the COVID vaccine trials, right? 0:22:29.9 SC: Yeah. 0:22:30.7 AS: Because we were able to say, We can have this incremental amount of investment to set up the infrastructures, and not just the infrastructure but to ensure that the dissemination and implementation aspects of a study is well oiled. And they were able to show that a return on investment is... And I think that's where beans research, research scientists and executives love these kind of pharmaceutical companies. Is why I think most of them really consider decentralized trials as a potential success story for them to embrace. And I think that's one of the values of that Tufts study, and we highlight that in our people. 0:23:12.4 SC: If people take this up, how does it fit with how we regulate clinical studies? That's something that multiple governments all over the world... They wanna have a look in and make sure that things are safe, they're private, that they're meeting standards. Is that something that you have to worry about differently with decentralized clinical trials? 0:23:29.1 AS: I think you do for a variety of reasons, but I don't think you abandon your foundational principles of doing good ethical research. What we highlight in the paper are the challenges that we see with big-data generation, data transfer, HIPAA compliance, or whether if you're in Europe or elsewhere, whatever the governance of that... Do you set up a governance structure as to who owns those data? Is the patients who owns those data or is it the research scientists or the funders? And we tried to, we didn't have a lot of Caper real estate to go into detail. 0:24:09.2 SC: No. You never do. 0:24:11.9 AS: So we tried to highlight those as areas that still need critical assessment, and we need some policies around that to then translate those into practical applications. 0:24:25.9 SC: We're gonna talk about studying studies, so what do you think... What kind of research, what kind of new data do you need to better design and implement decentralized clinical trials? What do you wanna learn more about, or what kind of research is being done into this that you think is important? 0:24:41.1 AS: Wow, this is a good question. I think where we are headed, it's not just the new types of data, so we talk about... In my capacity as chair of informatics on health data science, in our research vertical, we focus on what we call "digital biomarkers." When you look at disease and health, it's never black and white, the progression of disease is on a continuum. And what we want to be able to do as researchers is to focus on prevention and not just management, and what will allow us to be able to prevent diseases, to be able to predict who will get those disease. And so if we try and figure out the periods of the onset of disease, and if a decentralized trial, whether it be through digital biomarkers, will allow us to capture things far earlier than we would have previously, then that would significantly lead to reductions in prevalence and instance rates of chronic conditions and other types of health conditions. We are focusing on traditional types of biometric signals that can be captured over a period of time called "digital biomarkers," but we are seriously investing our resources on time in what we call more virtual immersive technologies because it's very likely at some point, we'll be able to start capturing auras, right? 0:26:13.2 SC: Okay. 0:26:13.3 AS: No, here's why. 0:26:14.3 SC: Technical auras, right? 0:26:16.8 AS: Technical auras. 0:26:17.4 SC: Technical. Scientific auras. 0:26:18.2 AS: Yeah. Well, so here's the deal. Someone who has a migraine, what is typically the triggering symptom that migraine patients typically talk about when they know they're gonna get the migraine? Feel these auras. But how do we measure those? Is it just through subjective report? So wouldn't it be cool if we have some form of device, some solution that will be able to structure the onset of those types of physiological auras where people no longer have to suffer? 0:26:49.7 SC: There is some kind of... We don't know how to call it, but a state that's measured through many, many different facets that has appeared as a signal when you had enough data from tracking enough people. 0:27:01.5 AS: Exactly. And I think not just what we wanna capture, but what we're looking at, which we try to focus on in the paper, is not just the discovery piece but the translational aspects of it. So what we're working on, which we didn't really say this much in the paper, is that we're using our decentralized studies to capture as much data where we can build digital twins of people so that we can start focusing more on precision medicine and the construct that we created called Precision and personalized population health. 0:27:34.4 SC: Okay. Thank you so much, Azizi. 0:27:36.3 AS: Thank you, Sarah. Much appreciated. Thanks for having me. 0:27:39.0 SC: Azizi Seixas is the interim chair of the Department of Informatics and Health Data Science and a professor in the Department of Psychiatry and Behavioral Sciences at the University of Miami Miller School of Medicine. You can find a link to the review that we discussed at science.org/podcast. 0:27:56.5 SC: And that concludes this edition of the Science Podcast, if you have any comments or suggestions, write to us at sciencepodcast@aaas.org. We also see comments on Spotify. You can listen there or Overcast or Apple Podcast. To find us on those apps, search for "Science Magazine" or you can listen on our website, science.org/podcast. This show was edited by me, Sarah Crespi, and Kevin McLean. We had production help from Megan Tuck at Podigy. Our music is by Jeffrey Cook and Wen Koi Wen. On behalf of Science and its publisher, AAAS, thanks for joining us. doi:10.1126/science.z6rlb01 ABOUT THE AUTHOR SARAH CRESPI mailShare on X AUTHOR Sarah Crespi is the Senior Multimedia Producer at Science and host of the Science Magazine Podcast. View more CHRISTINE PETERSON Share on X Christine Peterson is a freelance journalist in Laramie, Wyoming, covering wildlife and the environment. View more -------------------------------------------------------------------------------- SCIENCE PODCASTS 22 Aug 2024 Saving wildlife with AI, and randomized trials go remote By * Sarah Crespi, Christine Peterson 15 Aug 2024 The origins of the dino-killing asteroid, and remapping the scientific enterprise By * Sarah Crespi, Elizabeth Culotta, Shraddha Chakradhar, Meagan Cantwell 8 Aug 2024 Why humidity is worse than heat, and studying the lifetime effects of famine By * Sarah Crespi, Meredith Wadman View more ADVERTISEMENT RECOMMENDED 1 May 2020 Why Are Clinical Trials So Complicated? 8 Nov. 2017 When Small Trials Convince 24 Mar. 2020 The Latest Coronavirus Clinical Trials January 2023 Biological therapies need definitive randomized controlled clinical trials ADVERTISEMENT Skip slideshow FOLLOW US * * * * * * * Get our newsletter * NEWS * All News * ScienceInsider * News Features * Subscribe to News from Science * News from Science FAQ * About News from Science * Donate to News * CAREERS * Careers Articles * Find Jobs * Employer Hubs * COMMENTARY * Opinion * Analysis * Blogs * JOURNALS * Science * Science Advances * Science Immunology * Science Robotics * Science Signaling * Science Translational Medicine * Science Partner Journals * AUTHORS & REVIEWERS * Information for Authors * Information for Reviewers * LIBRARIANS * Manage Your Institutional Subscription * Library Admin Portal * Request a Quote * Librarian FAQs * ADVERTISERS * Advertising Kits * Custom Publishing Info * Post a Job * RELATED SITES * AAAS.org * AAAS Communities * EurekAlert! * Science in the Classroom * ABOUT US * Leadership * Work at AAAS * Prizes and Awards * HELP * FAQs * Access and Subscriptions * Order a Single Issue * Reprints and Permissions * TOC Alerts and RSS Feeds * Contact Us FOLLOW US * * * * * * Get our newsletter © 2024 American Association for the Advancement of Science. All rights reserved. AAAS is a partner of HINARI, AGORA, OARE, CHORUS, CLOCKSS, CrossRef and COUNTER. back to top * Terms of Service * Privacy Policy * Accessibility Get the latest news, commentary, and research, free to your inbox daily. SubscribeNot Now ✓ Thanks for sharing! AddToAny More…