— Also in TTHealthWatch: balloons or stents for coronary revascularization
October 24, 2026 • 9 min read
TTHealthWatch is a weekly podcast from Texas Tech. In it, Elizabeth Tracey, director of electronic media for Johns Hopkins Medicine in Baltimore, and Rick Lange, MD, president of Texas Tech Health El Paso, look at the top medical stories of the week.
This week's topics include balloons or stents for coronary revascularization, clinician use of artificial intelligence (AI), a model for predicting adverse infant outcomes in labor, and a more equitable way to distribute donor hearts.
Program notes:
0:40 Clinician use of AI clinically
1:40 703,000 active monthly users
2:40 How helpful is this?
3:40 More up-to-date
3:50 Prognostic model to estimate fetal compromise
4:50 Data from singleton births
5:50 Fit4Labour model
6:50 Could be used 8 hours before a problem
7:11 Drug-coated balloons or stent
8:11 Indications are symptoms or high risk
8:40 U.S. donor heart allocation
9:40 Best interests of their patient
10:40 Similar or significantly higher area under the curve
11:40 Can identify a higher-risk group
12:54 End
Transcript:
Elizabeth: Is there a way to use data gathered during labor to predict adverse outcomes?
Rick: Are drug-coated balloons or stents best for people that have coronary artery disease?
Elizabeth: Is there a better way to determine who gets scarce donor heart transplants?
Rick: And how do clinicians use artificial intelligence?
Elizabeth: That's what we're talking about this week on TTHealthWatch, your weekly look at the medical headlines from Texas Tech University Health Sciences Center in El Paso. I'm Elizabeth Tracey, a Baltimore-based medical journalist.
Rick: And I'm Rick Lange, president of Texas Tech Health El Paso.
Elizabeth: Rick, we're sliding into election season. There's a lot going on about artificial intelligence, so I'd like to turn first to JAMA and take a look at that particular, what shall we call it, resource in medicine.
Rick: There's a huge amount of interest in using artificial intelligence in medicine, but there's really not a lot of evidence on how clinicians currently use these tools. And that's what this study attempted to address. They wanted to characterize the clinician use of AI-enabled clinical information platforms.
So they used the one that is really most popular among clinicians. It's called OpenEvidence. And they queried it from January of 2024 through June of 2026. They looked at just under a million clinicians, over half of those being physicians, the other being PAs [physician assistants] and nurse practitioners. They submitted 271 million queries.
When they first looked in January of 2024, there were only 355 active users on a monthly basis. Fast forward to the end of the study, 703,000 monthly active users. The average monthly queries per active user rose from about 18 to 40. Physicians within a decade of graduation were more likely to use it. Residents and fellows were more likely to use it. There were certain specialties like oncology and hematology, internal medicine, and family medicine that were likely to use it.
What were the queries addressing? Primarily addressing clinical decision making, about 52% of the time, particularly looking at the treatment and management of patients. And they rarely used it to make a diagnosis. And oftentimes it was things outside of their specialties. It's really taking hold.
Elizabeth: Well, and unsurprisingly, I would suggest, and I certainly am not surprised by the piece of data relative to how many years ago you went to medical school and how often you are using it. That makes a lot of sense to me because these are tools that they're accustomed to in many aspects of life, not just in practice. So I think one of the things that I would like to know is, OK, so how helpful was this? Was there any qualitative data where they surveyed the users and asked about that?
Rick: No. In fact, that's the really interesting thing. It doesn't address the accuracy of the information that was provided. It doesn't address whether it altered patient care and whether it improved care at all. So these are the follow-up studies.
Elizabeth: And so I'm familiar, and you are too, with UpToDate, which I've used many, many times. And what this is striking me as is just a modification of that, the utility of being able to query what's going on with the latest in whatever it is, therapy for triple-negative breast cancer, for example, and all of the constellation of symptoms that might arise in somebody who's undergoing treatment and how to manage those. That might be a really great thing to be able to access.
Rick: Different from UpToDate in which a single author writes on a topic and updates it every 6 to 12 months, what OpenEvidence does, it holds content partnerships with several major medical publishers. It is more up-to-date than UpToDate. So you're right. It is an iteration. Now we need to know, gosh, does it actually improve patient care?
Elizabeth: Always the most salient question.
Let's turn to The Lancet. This is a prognostic model for use at labor onset to estimate the risk of severe fetal compromise at birth. I didn't realize that nobody has ever really done this before. We're all familiar with fetal monitoring or cardiotocography, or they abbreviate CTG, which is used to monitor the fetal heart rate during labor.
What they tried to do was develop a novel prognostic model for prediction of severe fetal compromise at birth by combining known clinical risk factors with the first 20 minutes of this monitoring. What they did in this study, which is retrospective, is they first developed their model based on a big database. And then they said, let's train it and test it on another one. And then they externally validated it on the third.
They included data from singleton births at 36 weeks and 0 days of gestation or more who were undergoing CTG monitoring, whether labor was spontaneous or induced. And they included in their model development factors like maternal age, gestational age, nulliparity, preeclampsia, diabetes in pregnancy, thick meconium when they were able to obtain that piece of information, and maximal recorded maternal temperature in labor. They also included fetal heart rate at the baseline, short-term variability, decelerative capacity, number of accelerations, and number of prolonged decelerations.
There were three mutually exclusive groups in birth outcomes: severe compromise, intermediate, and healthy. Ultimately, they created four risk groups. They were looking at 147,000-plus births with this piece of data, 20 minutes of CTG at the inception of labor. What they were able to show is that their model fit for labor had a sensitivity of 33.4% and compared with the clinical benchmark sensitivity of just shy of 40%. So they say it compares favorably. They opined that this particular model could help to produce interpretable data in real time so that clinicians could act on that and employ methods to hopefully improve fetal outcomes.
Rick: These are measures that are routinely available. When the severe compromise occurred during delivery, it usually occurred about 7 or 8 hours later. So what this is supposed to be is an alarm. The risk overall of having a severe compromise is less than 1%. This improved assessing those individuals, but a sensitivity of 33%, you can't really hang your hat on it. The next question is, can we actually do things that decrease the risk to the newborn child?
Elizabeth: That's exactly what I think. I would point out, however, that they say that this particular measure could be used 8 hours before there's a problem that arises. And it seems like that's a pretty robust window to be able to intervene if intervention is possible.
Rick: And you're right. What you do is you want to have an intervention that is successful at preventing the severe compromise. That's what needs to be tested. Elizabeth...
Elizabeth: Let's move on...
Rick: ... to The Lancet. If you have coronary artery disease, which is better, a drug-coated balloon or a stent? We know that drug-coated balloons are better in small blood vessels because when you put a stent in, it occupies some of the lumen. But what about in larger vessels?
That's what this study did. It was a systematic review and a meta-analysis of all the randomized controlled trials comparing drug-coated balloons to stents. Over 7,800 patients were identified. They realized that there was really no difference at all with regard to cardiovascular death or with regard to myocardial infarction, or was there restenosis.
Here's the caveat, though. When they looked at the large vessels, they did much better with the stent, about half as likely to have restenosis in the large vessel with a stent than with a drug-coated balloon.
Elizabeth: What about the medium vessels?
Rick: I'm going to say that you just divide it into small and non-small.
Elizabeth: Let's talk about how often it's really indicated to put something into a small vessel anyway.
Rick: The indications for doing percutaneous coronary intervention are to relieve symptoms or for certain vessels that are so high risk, the person is more likely to be alive if we open the vessel than if we just treat with medications. And then once you take the picture, at that point, you determine whether it's a small or a non-small vessel.
Elizabeth: Fit this into the paradigm of practice right now.
Rick: Yeah. To use drug-coated balloons with small vessels. In non-small blood vessels, a stent is probably better.
Elizabeth: Turning back to JAMA, then, let's take a look at this development and validation of an adult medical urgency score for U.S. donor heart allocation.
I see a lot of patients in the chaplain role who are awaiting transplant and have paid a lot of attention to this idea of organ allocation and the ethics, clearly, that are related to that. With hearts, there's no exception, yet one more of those organs, it's a scarce resource where trying to develop equity is a really important idea.
The authors cite the fact that one in five heart transplant candidates in the U.S. die or are delisted without receiving a transplant. And currently, hearts are allocated by a six-status therapy-based system that's been criticized because of its limited risk stratification and susceptibility to gaming.
I find the use of gaming, I just have to say editorially, a little bit pejorative. I think I might use something else because I don't think clinicians do that. What they're looking for is the best interest of their patient. The fact that they might try to modify things so that their patient might be able to receive a transplant, I get it. It's a deeply human thing.
Having said that, what they did was look retrospectively at all adult transplant candidates from January 2019 to August 2025. They split this into training, validation, and test data sets. They had a marginal structural model of 6-week mortality without transplant that they estimated, and they called this the U.S. candidate risk score or the US-CRS. Finally, they converted that into an allocation policy and a match-run simulation of waitlist reordering compared with the current system.
Almost 22,000 heart transplant candidates divided into these different groups, and their model included five laboratory values and four therapeutic devices that the patient would be currently using. Their system, they say, had similar or significantly higher time-dependent area under the curve than the current six-status system. The area under the curve predicted by their model was 0.807 versus 0.693 for the current model. And they conclude that their model prioritized sicker candidates with higher mortality without transplant compared with the current medical urgency statuses.
Rick: I'm going to take issue with the gaming because that's what happens. You're right. The physicians are doing what's in their patients' best interest, and that's true. There's no question about it. But what they don't have is the global picture because there are people around the country that need heart transplants and we don't have enough donors. We want to make them available to the sickest patients.
How do physicians game the system? If you put a mechanical device, like a balloon pump, in somebody, then suddenly now they've got to move to the top of the list. So what this score tries to do is it tries to level the playing field, not only about devices, but about these other blood tests. If we use those, we can really identify a higher-risk group.
Elizabeth: I thought the editorial that ostensibly accompanies this virtually does not mention heart transplant in its entire discussion does talk a lot about liver transplantation. It notes that in 2025 donation after circulatory death accounted for 43% of deceased donor organs, and that's a change. There are a lot of other changes, including xenotransplantation and utilization of AI models, what I'm going to call lab-grown organs, and they're going to radically change this picture.
Rick: There will be more availability of different types of transplants, but it doesn't take away from the fact is that they need to be equitably distributed. Getting this right from the beginning will help us as we add more options to the transplant candidate.
Elizabeth: On that note then, that's a look at this week's medical headlines from Texas Tech. I'm Elizabeth Tracey.
Rick: And I'm Rick Lange. Y'all listen up and make healthy choices.