Podcast: AI Makes Catheters 100 Times Safer

author avatar
Podcast: AI Makes Catheters 100 Times Safer

In this episode, we talk about a new AI designed catheter that is 100x safer than and how it came to be after a passive conversation about fun facts between two researchers.

In this episode, we talk about a new AI designed catheter that is 100x safer than and how it came to be after a passive conversation about fun facts between two researchers.

This podcast is sponsored by Mouser Electronics


(3:13) - Aided by AI, New Catheter Design Prevents Bacterial Infections

This episode was brought to you by Mouser, our favorite place to get electronics parts for any project, whether it be a hobby at home or a prototype for work. Click HERE to learn more about the type of medical devices that can benefit the most from additive manufacturing!


What's going on, folks? Welcome back to the Next Byte podcast. And in this one, we're talking about catheters. That's right. Are you excited to learn about how catheters give people infections? Probably not, but that's okay, because we're gonna talk about how the folks over at Caltech are solving the catheter bacteria problem. So, if you're excited, buckle up and let's get into it.

I'm Daniel, and I'm Farbod. And this is the NextByte Podcast. Every week, we explore interesting and impactful tech and engineering content from Wevolver.com and deliver it to you in bite sized episodes that are easy to understand, regardless of your background. 

Farbod: All right, folks, we're back. We're talking about all good things, medical devices. But before we jump into today's episode, let's talk about today's sponsor, Mouser Electronics. Now you're probably thinking medical devices, Mouser, how does that line up? Well, friends, let me tell you. Well, by now you should actually know, but Mouser is the world's biggest electronic supplier. And by the nature of being the world's biggest electronic supplier, they got some really cool connections to folks in academia and in industry. What that means is that as cool things are coming down the pipeline, they hear about it and sometimes they write about it. For example, in today's Mouser resource that is very relevant to today's topic, they're talking about the types of medical devices that are suitable for additive manufacturing. They actually, first of all, start off by saying low primer on additive manufacturing, what it is, the types of medical devices that we commonly see in the field, they break it down into two buckets. You got the, let's say, sophisticated complex ones that have electronics embedded then you have the not so complex ones because they are just, let's say, hip implants. They're complex by design, but they don't.

Daniel: Yeah, still might be complex.

Farbod: For sure, but they don't have integrated circuitry. So, they're like, if the complex geometry is what you're going for and you wanna make sure you're getting it exactly right for that person or for that use case, then additive manufacturing could be a great way to go. They also talk about surgical guides. For example, if you're doing dental surgery and you wanna line up a drill hole to be in an exact spot you can do a scan of that person's mouth, come up with a perfect surgical guide. They talk about prosthetics, and where they kind of end it off is there's this whole new generation of bioprinting happening, and the future looks very bright for additive manufacturing's role within the medical field because of that. I thought it was pretty interesting. Obviously, we've done a lot of topics about medical devices, 3D printing, bioprinting and whatnot, but again, as a primer resource, this was already spot on and it only took me like what two three minutes to read and digest all the information in there.

Daniel: Yeah, I agree man. Right. Mouser being at the tip of the spear as they are with new technology does a great job of communicating, kind of like medical 3D printing for dummies is almost what this resource is. And that's why we always turn to them as a technical resource when researching for the podcast and for other technical projects. And that's why we're happy to have them as sponsor and have their article linked in our show notes.

Farbod: Absolutely, if you guys want, as always, you can find the article in the show notes. And with that said, let's jump into today's article. This one's coming out of Caltech. I feel like we haven't been to Caltech in a hot minute. We haven't talked about anything out of Caltech in a while. But I'm going to say it, I know it's early. I know it's early. But this might be one of my favorite episodes that we're going to do this year.

Daniel: Oh, let's go, man. I thought you might've had PTSD from studying fluid mechanics in college, but.

Farbod: No, I'm coming back for round two after this. I'm into it. So, let's talk about what's going on. This group of researchers with varied backgrounds are trying to come up with a way to create a catheter that prevents bacteria from swimming up. Now, why are they doing this? Well, the statistics are actually pretty interesting between 15 to 25% of hospital patients receive a catheter during their visit, which is, it's a pipe that's going up your urethra to help you urinate. Now I knew that catheters were used, I just didn't know how common they were, like up to a quarter of everyone going to a hospital is getting one. And interestingly enough, one of the most common bacterial infections in a healthcare setting comes from catheters.

Daniel: Well, that's what I was gonna say. I was aware, I think, I would say, I was largely aware of the vast percentage of hospital patients that ended up needing a catheter at some point during their stay. I had no idea how common it was to get a bacterial infection as a result of a catheter. That it makes me cringe. Just think it gives me such a visceral reaction thinking about there's something already which feels intrusive into your body. Also providing a porthole for bacteria to infect up your urethra that does not sound pleasant given the vast percentage of folks that receive a catheter during their stay, I can only imagine the immense cost, the immense complication, the immense uncomfortability that comes alongside with bacterial infections, if not, real lethal health risk as well.

Farbod: That's what I was going to say. Imagine you have patients who are immune compromised and they require a catheter and now there's an added risk for something that is so essential for them to have during their stay. And I don't know, I kind of thought they would have figured out that you're not getting a bacteria from needing to pee, figured out by now. But it's a difficult problem. And you brought up cost. This is costing America 300 million US dollars every year. That is insane.

Daniel: And it's just a portion of the rest of the world's population that receives catheters as well, right?

Farbod: And the reason that this is happening is pretty interesting too. This is the throwback to fluid mechanics you were talking about. Essentially the way the flow, the fluid flow in the catheter works is that the fluid moves slower on the wall, closer to the wall and faster in the middle. A phenomena that is referred to as the Poiseuille flow.

Daniel: And I think it's pretty intuitive to understand that toward the boundaries around the outside of this catheter tube the fluid isn't gonna flow as fast because it's got friction with the interface of the inside of the tube. So, the middle of the flow, you can think about a river, right? The middle of the river is going to have the fastest flow because the flow on the outside is the one that's interfering the, the water's colliding with the, the banks of the river. It's a very similar phenomenon here with the Poiseuille flow. But the problem is bacteria can kind of take advantage of this Poiseuille flow and they call it like a two steps forward, one step back pattern. They kind of like flow around the most flow around the tubular structure in a unique motion that allows them to basically advance their way up the catheter. And if you can imagine, advance your way up the catheter into the body by taking advantage of the slow flow toward the outside. And this movement that bacteria exhibit inside tubular structures, because of the poiseuille flow phenomenon, basically increases the risk of infection in medical settings. It allows bacteria to take advantage of this fluid mechanics phenomenon and swim their way up the catheter flow against the direction of flow and into the body.

Farbod: Yeah. Well, what is that saying? While you were busy sleeping in fluids class, I was mastering the flow. Um, no one says that. I just came up with it. Please laugh.

Daniel: I've never heard that, but I'll laugh to make you feel good.

Farbod: Thank you. Thank you. That's what friends do. But this is where probably my favorite portion of any of our episode starts is a good story. This is where the story really starts. Now we're at Caltech. There's a postdoctoral student, Dr. Zhou, and Dr. Zhou is a chemical engineer. Dr. Zhou is sharing this phenomena with a professor of mechanical engineering, Professor Daraio, and he's just saying, hey, I just learned about how the bacteria progress through these catheters. It's a two step forward, one step back. I think it's pretty interesting. Now Professor Daraio's response wasn't like, oh, that's cool. Let's move the conversation along and let's talk about our day, whatever she was like, this is very interesting. What would a practical response to this be? Like a very analytical approach. And the reason, apparently Professor Daraio brought that on is because her entire research focuses around the different geometries and how they interact with the environments that they're put in.

Daniel: And when you're saying the geometries here, right? I think her specific focus was on the geometry of the inside of the catheter tube, right?

Farbod: Correct. Correct. So, she was wondering like, is there any sort of manipulation that we could do to deter this progression of the bacteria through the inside of the catheter? Now, this is how we're going to start to get into the core of the sauce. What did they come up with, Daniel?

Daniel: Well, I think it's really interesting. They take advantage and borrow some geometry from nature. I think it's shark’s fin is what inspired the triangular protrusions along the interior wall of the tube if you think about it a triangle shaped spikes that that mimic the shape of a shark's fin along the Interior walls of the tube these structures create turbulence, right? And they kind of bounce the bacteria back into the center of the flow where the flow is fast and downstream. So, as opposed to the outside walls where there's some turbulence and maybe some vortices, right? Where the bacteria is able to take advantage of the slow flow near the walls, these triangles push bacteria back into the center where there's a jet stream flow downstream out of the catheter tube. And it basically makes sure that anything that could potentially be floating upstream and taking advantage of the poiseuille flow, whether it be bacteria or other particulate matter, etc, is forced back into the middle of the stream and then sent down and out of the catheter as opposed to up into the body.

Farbod: Yeah. And I mean, you already said it, but just to summarize, the geometry gives you advantages on two sides, right? One, the bacteria can't just like slide up the wall and not worry about the middle flow anymore. The geometry forces them to encounter that middle flow every time the spike goes up, right? So, they got that jet stream coming at them. And the geometry of those spikes are very reminiscent to that of a airplane wing where you have these massive vortices that are coming off this turbulence that is further causing them to get yanked back and go downstream instead of being able to work their way back upstream. So that's the genius behind it. And that's already cool as is. But obviously, as you, we're mechanical engineers, simulation is our best friend before we wanna make anything. As you're trying to iterate on your design and you wanna do some simulation, that's great. You wanna start doing some experimentation as well, right? Well, mechanical engineers, being as great as we are. We actually need help from other people as well. In this case, they reached out, Zhou actually reached out to another friend who was working at the biology lab, Dr. Wan, to say, hey, would you be able to help us out with this? Because I think it's aligned with what you're interested in.

Daniel: And Dr. Wan was studying small worms named after our friend, Nema, nematodes.

Farbod: I'm sure he's gonna appreciate that shout out.

Daniel: Hopefully you don't hit me after that one, Nema, but studying small, like microscopic worms.

Farbod: 5 to 10 micrometers, I think.

Daniel: These really, really small worms studying their movement pattern. So, what a perfect team to come together and say, hey, let's use, and I think what they ended up borrowing from this nematode studying team was kind of a two-phased approach where they use 3D printing to prototype something and then high-speed cameras to monitor fluid flow around those geometries. They were able to borrow a lot of this techniques that the nematode team was using and were able to quickly rapid prototype using 3d printing and then use the high-speed cameras to monitor actual bacterial progress and see which geometry might be best for limiting bacterial spread again up through the catheter tube into the body.

Farbod: Absolutely. And what was cool is, Dr. Juan, when they reached out, they were like, Dr. Juan was like, yeah, like, this is pretty up my alley. I love studying the trajectory of how these small worms works. I was excited to get on this. And again, it just felt like this perfect storm of this story coming together and this idea coming together, but it doesn't stop there. It doesn't stop there. Now that they had this data from the 3D printing and the high-speed cameras and all that, they were like, all right, cool. Now we wanna even iterate more on our design, right? We want to run more simulations and see how the different curvature of these fins could impact our ability to eject these bacteria even more.

Daniel: And they at this point have probably had a basic simulation with which they used to create their initial prototype of the design. They've got a couple iterations. And they've even got physical testing results with real high-speed cameras tracking the flow, monitoring bacterial progress. Speaking of someone who works with engineers that use optimization studies, you've got all the recipe you need to run an optimization study. You've got an initial simulation with some projected results. You've got then physical testing results and a couple iterations of those to correlate back to the initial simulation. At this point, they were ripe and ready to try and feed this back into the simulation loop and get an optimized design, a theoretical geometry that they should use that takes advantage of the physical testing results and the simulation and trying to understand what's the best possible in this case, shark’s fin shape and size to help limit bacterial progress up the catheter tube and into the body.

Farbod: Yep, absolutely. And that's where they collaborated with the AI laboratory within Caltech to optimize their design. Now, you might be wondering, well, what are they doing with them? We've used generative design tools before. Like, you give it a certain parameter, you run a certain simulation, and then it creates potential options for you that you can test against. Running these simulations, especially if you're adding fluid components to it as well, is very computationally heavy. At times it can take days and it looks like in their case It was also taking days to iterate on every design and then test and then get the feedback feed it back in repeat the process and that's just very time-consuming.

Daniel: Well, and one of the problems right, is if you're trying to study optimization using kind of a brute force method like you mentioned that takes days. You're not necessarily intuitively changing the geometry and looking for expected results you're looking at all different types and shapes and sizes geometry and figuring out which one's best. And then that's the one that you proceed with. So, you know, instead of looking at a few key combinations that you can expect to be successful, you're looking at thousands of different combinations and then picking the one that's best based on the results.

Farbod: Absolutely.

Daniel: And that's what takes days, not minutes, right?

Farbod: But they hacked it. They hacked it with the help of the AI lab. Which lab is this? This is the Anandkumar, I think, lab for artificial intelligence.

Daniel: Yeah. Computing and mathematical sciences.

Farbod: There you go. And they are using these neural operators, which is this artificial intelligence method of processing the data for this optimization problem. And they were able to get the feedback they needed that would usually take them days within minutes, which allows you to iterate quicker and then essentially move forward with your project much, much faster than you could before.

Daniel: My interpretation, my understanding of the way that that AI optimization process worked is by narrowing down the number of different simulations and the number of different parameters that were changed in a simulation to arrive on what you believe to be the optimized optimal design much quicker, right? As opposed to, like we said, trying thousands of different combinations, right? Different size, different shape, different number of fins, how far they protrude into the, into the tube, whatever they were able to use AI to help guide the simulation to focus on a few key parameters that generated the optimal result, again, in a matter of minutes, not a matter of days, which is really interesting.

Farbod: Absolutely. But all that, what has it led to? What is the so what of this entire multidisciplinary effort? Well, they have been able to reduce the movement of bacteria by two orders of magnitude. That's 100 times. Which is crazy, and the implementation of it is kind of what blows my mind because when I was thinking about the original problem, bacterial infections, if you want to take the chemical approach to this problem by like prescribing antibiotics, you always run the risk of having the bacteria adapt to it. And then like, at least in America, we have a problem of overprescribing antibiotics. So, you have you can create super resistant bacteria.

Daniel: Well, I would say that that's one problem, right? Creating antibiotic resistant bacteria. Those are known to fester in the medical space, right? Because of the vast use of antibiotics in the medical realm. But in addition to all that, I mean, just the pure side effects of antibiotics on the patient's body, that alone is enough of a reason to try and look at this, I wouldn't say it's completely passive approach, but look at this passive approach that doesn't involve side effects to the patient's body. This is something purely related to the geometry of the catheter tube design. If I'm a patient with a catheter, I'm not gonna know the difference whether mine has shark fins on the inside or not. It's not gonna cause a decrease to my quality of life, a decrease to my health. It's actually gonna be no difference except for the fact that it's 100 times less likely that bacteria makes its way up the catheter and causes an infection in my body.

Farbod: Exactly.

Daniel: And I thought about it, hundredfold decrease in upstream bacterial movement. I'm not sure how much these tubes might cost to implement one day in the future, but I looked, I remember the figure that you said, bacterial infections cost $300 million US dollars annually in the US alone, right? That's only a portion of the world. I thought for sure if we're able to reduce bacterial infections associated with catheter installation. If we're reducing the cost of those infections down to $3 million per year, as opposed to $300 million per year, that looked to me as an opportunity to invest up to $297 million in upgrading our catheter tubes to ones with these designs. And if you invest any less than $297 million, you've saved money.

Farbod: You've won, yeah.

Daniel: Which is crazy. That's a lot of money that we could invest to upgrade our catheter infrastructure. I'm not sure how much it costs to replace every single catheter of two in the US, but I imagine $297 million would be a great start.

Farbod: And again, you might have a lot of that upfront cost that seems like a lot of a burden, but the reality is down the line. Like I feel like the population's growing, people needing hospital visits is just going up. This is just gonna save you in the long term. So, to our fans that are listening who might be interested in the medical devices field, this is an opportunity. This is a great opportunity to collaborate with these amazing folks at Caltech to make all of us a little bit safer if we have to go to the hospital.

Daniel: Well, and I was going to say, this is really inspiring to me. Noting how it started with a pretty innocent conversation between a professor and a post-doc researcher saying, hey, I learned about this phenomenon that happens in catheter tubes. And through a series of collaborations between that first professor with the post-doc scholar in chemical engineering post-doc scholar in biology, working all the way through to chemical engineering and then mathematical sciences, right? It's a really interdisciplinary team. It shows just how complex bioengineering problems can be, but it all started with a series of conversations. It reinstills my faith kind of in the fact that you can truly stumble upon a life-changing innovation, just truly have conversations with other people who are bright and are willing to talk about it.

Farbod: Or by listening to a podcast.

Daniel: That's what I was going to say is maybe you'll get struck by that lightning bolt with an idea of something that you need to change or something that you can fix or apply a new piece of technology in your life to solve a problem. Maybe you'll gain inspiration from this podcast. So, I would love…

Farbod: That’s the dream, isn't it?

Daniel: I would love for one of these stories one day to be like the first step, the first domino that fell before the giant domino toppled over and then saved. $300 million a year in catheter infections is like someone listening to the podcast and hearing this conversation.

Farbod: We'd have made it then. That's the dream. But let's try to do a quick wrap up of the episode. Yeah. The long and short of it is if you go to a hospital, the people around you, 15 to 25% of them, they're gonna need a catheter. And the crazy thing is that catheter infections are very common. Like it cost the US taxpayer or just the US patient $300 million every single year to deal with this problem. And the reason that happens is because the flow inside of a catheter, it's very fast in the middle, but very slow on the sides. That's called the Poiseuille flow phenomenon. And the bacteria take advantage of this by crawling up the side where it's slow, two steps forward, and they take one step back in the middle. These folks at Caltech, they were just having a passive conversation about how this phenomena is happening with bacteria. One of them proposed a solution by using shark’s fin designs to slow down that progression and then using the jet in the middle to eject all the bacteria. One thing led to another. They kept optimizing. They worked with biologists and AI researchers and they've came up with a new catheter that reduces the rate of bacteria going up by a hundred times.

Daniel: Crazy.

Farbod: Wild. And that's why this is probably one of my favorite episodes of the year.

Daniel: Yeah, for sure. And it all comes down to fluid mechanics, which if I recall was something that was not your favorite when you were studying mechanical engineering.

Farbod: Lowest grade I got. All four years, I had a B-.

Daniel: Yeah, me too.

Farbod: B-, I'll never let it go, but you know what? It's okay. We've come a long way.

Daniel: That was quite a long pause before you decided you were okay.

Farbod: There's a lot of pain in there.

Daniel: Yeah, I could tell.

Farbod: Folks, before we wrap up, quickly wanna give a shout out to our fans in Hungary. You guys made us trend. I think we were number 114 in Hungary, which is awesome. I think it's the first time we're trending in Hungary as well. We owe you guys a thank you. I listened to the Google translation at least 10 times, so I really hope I don't butcher it if I do. Please forgive me.

Daniel: Actually, let me give a correction quickly before you give your Hungarian thank you. We were the number 27 podcast in Hungary. I think that warrants an extra thank you from us in English and then hopefully an awesome thank you in Hungarian.

Farbod: Oh my God, the pressure's on now.

Daniel: I wanted to crank the pressure up a little bit before you did it.

Farbod: Thank you, yes. What are our friends for if not pressure?

Daniel: Yeah.

Farbod: To our friends in Hungary, Köszönöm, hógy meghallgattak.

Daniel: What does that mean?

Farbod: Thank you for listening.

Daniel: I needed the translation because I imagine your Hungarian is bad enough that they would need the English translation.

Farbod: OK, you know what? You've done enough. Everyone, thank you so much for listening. And as always, we'll catch you in the next one.

Daniel: Peace.

As always, you can find these and other interesting & impactful engineering articles on Wevolver.com.

To learn more about this show, please visit our shows page. By following the page, you will get automatic updates by email when a new show is published. Be sure to give us a follow and review on Apple podcasts, Spotify, and most of your favorite podcast platforms!


The Next Byte: We're two engineers on a mission to simplify complex science & technology, making it easy to understand. In each episode of our show, we dive into world-changing tech (such as AI, robotics, 3D printing, IoT, & much more), all while keeping it entertaining & engaging along the way.


The Next Byte Newsletter

Fuel your tech-savvy curiosity with “byte” sized digests of tech breakthroughs.