This month’s episode of “News You Can Use” on HealthcareNOWRadio features news from the month of April 2024
News You Can Use with your Hosts Dr Craig Joseph and Dr Nick van Terheyden
The show that gives you a quick insight into the latest news, twists, turns and debacles going on in healthcare withmy friend and co-host Craig Joseph, MD (@CraigJoseph) Chief Medical Officer at Nordic Consulting Partners and myself, where every diagnosis comes with a side of humor. We hope you stay curious, stay engaged, and keep seeking the truth in healthcare in a world that thrives on information.
Buckle up as we dive into the ER of excitement, the ICU of irrationality, and the waiting room of wacky wisdom in this month’s show that features a review of:
- AI and the Various Models in Healthcare
- Paywalls
- User Design and Divorce
- Clinical Research into Sleep Apnea
- The National Physician Advisor Conference
AI, Apnea, and Awkward Divorces
We open analyzing the latest developments in AI and healthcare technology, discussing the proliferation of large language models like Claude and Llama, speculating that users may come to rely on different models for different use cases, just as one might choose different vehicles for different trips. There is the intriguing study on using AI to draft responses to patient portal messages, which didn’t save physicians much time but did boost morale by preserving empathetic language.
User interface design emerged as another hot topic, with the doctors relating the anecdote about a lawyer accidentally divorcing clients due to confusing software UI.
The conversation turned to recent clinical research, including the connection between obstructive sleep apnea, atrial fibrillation, and stroke risk. While not a formal peer-reviewed study, the analysis demonstrated the power of mining large datasets to uncover important medical insights.
And finally, an inside look at the National Physician Advisor Conference in Coronado, California featured high-quality content from the clinicians who manage denials, appeals, and other administrative aspects of hospital billing and reimbursement.
We hope you enjoy our take on the latest news and developments in healthcare and want to help you keep untangling the web of information, dodging the sensational pitfalls, and emerging victorious, albeit a little dizzy, on the other side. In the end, the stories we uncover, and the discussions we ignite, all shape the narrative of our shared future. We want to hear from you especially if you have topics covered or questions you’d like answered. You can reach out directly via the contact form on my website, or send a message on LinkedIn to Craig or me.
Until next week keep solving healthcare’s mysteries before they become your emergencies
xx
Listen live at 4:00 AM, 12:00 Noon or 8:00 PM ET, Monday through Friday for the next week at HealthcareNOW Radio. After that, you can listen on demand (See podcast information below.) Join the conversation on Twitter at #TheIncrementalist.
Listen along on HealthcareNowRadio or on SoundCloud
Raw Transcript
Nick van Terheyden
First off this week, we have the latest news in AI and large language models. And what’s going on with paywalls and cyber attacks. So AI, it seems to be our sort of trending topic these days. And where are we with that? Well, I think we’re in this sort of continued improvement, but not the sort of revolution that we all felt when chat GPT emerged on the scene. And it’s still incredible. But one of the things that I’ve been reading a number of times now is folks are saying some of these other models are pretty good. And I’ve had some explicit experience with the other models. They come with different pros and cons, but certainly Claude, which I’ve tried used, I was super impressed with when I tested it out with a test. And it did really, really well. And then I hear Larmer, which I think is the Facebook group has leapfrogged or is potentially leapfrog over some of the other models. So I think, overall, I’d say interesting times keep testing all this stuff out. But no, no quantum leaps ahead. But there’s some other things going on as well. Right?
Craig Joseph
Well, absolutely. And to your point with the different models, I think we’re we’re, it’s kind of like going to be what kind of cardio like, you know, they’re, they’re all going to be good and want to come out with something good. And to some people, they make a choice of a car based on how big the cup holder is.
And this is apparently true in the United States, cup holders in the UK? No,
Craig Joseph
well, we have them they’re required by law, I actually don’t know if that’s true, but they shouldn’t be. And some people do make those decisions. And so I think what you’re gonna see is, oh, well, I use chat GPT for this, and I use llama for that. And, you know, you’ll get to, you’ll get to your favorite.
Nick van Terheyden
So it’s like your garage, where you’ve got, you know, if you’ve got to get somewhere fast, you take the Porsche and you know, if it’s the family trip, you’re taking the Range Rover, SUV, and you know, you pull out the different cars at different times. And Craig,
Craig Joseph
I feel like you’re you’re contributing to stereotypes that all physicians are incredibly wealthy. And so I don’t have either of those kinds of cars. Those are not decisions that I get to make, Nick, apparently, I’ve I’ve made some bad decisions in my life and not becoming a brain surgeon.
Nick van Terheyden
Well, I’m sorry to hear that. But, you know, hopefully, at some point, you’ll get to choose your cars.
Craig Joseph
Well, you know, I we did want to I did want to kind of mention the article that came out in JAMA last week about physicians using an artificial intelligence to prepare a response to patients who query them via the patient portal. So many people have something like my chart, or might be branded a different thing. But you know, you send a message to your doctor and say, Hey, I’ve have the symptoms, or this things come back or you mentioned, I should do something if this thing happened and inherits now happening, what should I do? And physicians have gotten over the last couple of years, especially since the pandemic started, really overwhelmed with just so many messages, many of which would have been phone calls or in person visits. And so we’re trying to do as much as we can to help physicians with this with this deluge. And one thing is to see if an AI can kind of craft that message for you, and then let you just tweak it a little bit. And what they found. What they were hoping to find was that this was going to save physicians time, because they could just look at it and go yep, that’s a really great answer. Click Send. What they found was it didn’t really save a lot of time, in fact, because physicians were fairly heavily editing some of the suggestions. Now some of this is because AI is not the godsend, we all think it is in terms of being able to diagnose and treat some of it however, I’m glad you said that I, I’m there for that. Some of its though, in the AI is defense, it doesn’t get access to the entire medical chart. At this point, it’s only getting a small subsection of fairly small subsection of the information that the physician and all the humans know about you. And so it’s a little unfair to be asking it to do these amazing things. That said, it didn’t save a lot of time. It did, though, however, make physicians happier, and it kind of decreased burnout and increase morale. And what they, what they found when they studied, hey, what changes did these physicians make to the AI generated response? It often left in the physicians left in the sentence, or two or three that was typically at the beginning of the message with kind of that the empathy part, you know, what some of us might call the fluffy part, like, wow, really sorry to hear that this happened, you know, if it’s bad news, you know, this is unfortunate, but we’re going to work through it, that that kind of thing, physicians left that in and actually appreciated that. And I think that’s a popular part of the message that the patients are seeing. And that was probably missing a lot. Because physicians, we’re going to try to get through these this massive of notes and responses as fast as possible. Thank
Nick van Terheyden
you. I gotta say, I think it’s a physician thing. Maybe it’s not. And you know, there’s an n of two here because, well, no, maybe some more from from the study, but my own personal experience with notes, and let’s just talk about regular email, I get an email, I’m replying, I hammer out literally 123 You know, here are the answers. And then actually, I go back up to the top, and I put the nice niceties in and the stuff at the end, I do it afterwards. Because you know, it’s important, but it’s not sort of built into my natural. So I’m not surprised to hear that. I’m, you know, and I think it’s good, because that bleeds through and I think it’s important for the patients.
Craig Joseph
I you know, I get it. The closest I can I can kind of relate to this is when patients used to call me your parents, I’m a pediatrician, and usually the patients don’t call me but I got feedback. Once I think my first year of practice from the office manager. Someone was a little perturbed that when they called me at 11pm, for something that, in my opinion was not urgent was not a new question, something that should have been addressed during several days when our office was open. The feedback was that, you know, I wasn’t as pleasant at 11pm as, as I was in the office, there was no small talk, there was no pleasantries. It was just the facts in just the response. And that’s it. And this mom was was unhappy with that. And my response was, Well, I don’t feel I’m gonna change very much, but I I can imagine, you know, answering after message after message after message you said you kind of just answered it and then put the pleasantries in. I think a lot of doctors just skip the pleasantries.
Nick van Terheyden
Well, and I think, you know, the MLMs allow you to create something that’s a little bit variable, because, you know, if you just cut and paste, it’s not quite the same.
Craig Joseph
No, no, of course not. But so that was a that was an unexpected kind of positive that the physicians even though they didn’t save much time, at least at the beginning, were, were pleased with the pleasantries, and the patients are pleased and reading the pleasantries. And so it’s, uh, I’ll chalk that up as a win.
Nick van Terheyden
Well, I think that’s good news. You, you mentioned change. And since you mentioned it, we’ll talk about change healthcare, I think that’s still an ongoing challenge for folks and the cyber attack that has debilitating debilitated large numbers of organizations, lots unfolding, but it brings us to the whole issue of paywalls, which I gotta say, I got a bee in my bonnet with all of the investments made with research that is funded by the government, the NIH, and you can’t get to most of these articles that paywalled there are ways around it. But you went to look at an article and couldn’t get to it.
Craig Joseph
Well, I was I was a little perturbed because I couldn’t get to an article that I feel I paid for. Now, I couldn’t get to a journal you subscribe to I subscribe to it. And so I thought I could get everything in it, but they’re like, oh, no, some of this content is cost extra.
Nick van Terheyden
So that’s the Ferrari level. You didn’t You didn’t I didn’t
Craig Joseph
I was only at the Porsche level for this for this particular website. But I as I was mentioning to you, as we were preparing for the show, I have managed to get around it two different ways. One is there was a, there was an option to have it read it to me. Yeah. So, hey, read the article out loud, well read the whole article, which is great. It just, when I looked at it, I could only see the first three sentences. But when I asked it to read it for me, it read the whole thing. So I assume they’ll fix that at some point. But they haven’t fixed that yet. And, and I tried another thing I could I can gift an article to people who don’t subscribe. And I, you get a certain number of those options every month, I think. And so I tried gifting to myself. At first it didn’t work. But then I said, Oh, I wonder what will happen if I, if I use a private browser tab, which, you know, removes all the cookies. And so I’m showing to you even knew what private browser? I do. Okay, I’m an expert. I’m gonna read something. I’m not sure. Yeah, that’s fine. It’s a free country so far. Yeah. So I went to I went to a private browser tab and pasted that gift article. And then again, let me emphasize, I could not read. But when I gifted it to myself, like, I could read it again. So it was so yeah, things are, things are, yeah, it’s good. It’s the scientific method, Dr. Nick, just applying the scientific method, and
Nick van Terheyden
you were jumping through all the necessary hoops. And thank goodness, your tech chops were sufficiently good enough. Um, just
Craig Joseph
let me emphasize Dr. Nick, that I do subscribe to this particular periodical. So I felt like I should have had access to it. And I don’t want to hear from
Nick van Terheyden
you did, and we haven’t revealed it. So we’re not
Craig Joseph
we have not shown that? No.
Nick van Terheyden
Okay, so, recent study, I think this was another epic, which is one of your favorites. Right? This is the obstructive sleep apnea of somewhere that I spent a lot of time you know, sleep. And there was a study just out of Gallup that published that we were getting less than 25% of what we want in terms of sleep, it’s worse for women. 1942, when they first did the survey, 60% of people got more than eight hours, and we’re very happy with the manner. So anyway, um, obstructive sleep apnea, which is a clinical condition to be clear, but there was some research tell us what was going on there. Yeah, this is,
Craig Joseph
again, I think we’ve referenced this, a number of times Epic is a big electronic health record vendor, globally now. And they have access to a lot of data. And so they’ve, they’ve to try to spur people to use some of these data they have done. I’m gonna say research, but I’m gonna say research with air quotes, right? So it’s not a peer reviewed study. And there goes,
Nick van Terheyden
Hey, Oh, I’m sorry. No, it wasn’t those quotes. Sorry. Yeah,
Craig Joseph
no, no, not be. Now. Their goal is not to change. Perceptions or to change what we do their goal is to push people to say like, Oh, that’s interesting, we should do a more formal study of that. And so they’re kind of just very preliminary, but what they did, because they have access to so much data, they looked at 3.9 million atrial fibrillation, patients afib, or a three, almost 4 million AF patients, and said, Hey, are there more patients with who have afib. And we know that if you have afib, you’re at higher risk of having a stroke, does, how does obstructive sleep apnea play with that? And what they found was, yeah, if you have AFib as an underlying condition, and then develop obstructive sleep apnea, your risk of having a stroke is higher. So that was something that I think, has not been clear, but it’s probably now more clear, at least than than it was before. Again, these are all for very preliminary and not peer reviewed. I want to be clear about that. Another interesting aspect of the study, though, was that they looked at the the major model about predicting someone the risk of having stroke if they if they have some problems, like atrial fibrillation. And what they found was actually the model does a pretty good job of predicting your risk of stroke, whether you have obstructive sleep apnea or not, which was not expected. One would expect that if you don’t even ask that question, and the model doesn’t take into account you having obstructive sleep apnea or OSA than the you know, the model will give you a lower risk than you actually have and they did not find that to be the case. So, again, I think it’s just an interesting use of a very, very large data set. I think Epic is is trying to use it for good and not not for evil, then they don’t, they don’t seek to sell any of these data sets.
Nick van Terheyden
And public there’s no paywall.
Craig Joseph
There’s no paywall to their, to their little mini studies Exactly. And all of the data are, are volunteered by the their customers, the hospital systems. And so if a hospital system is uncomfortable with this, they don’t, they don’t have to share their data. But if they do share their data, then they their researchers have access to it in a de identified way. And so, again, in this case, you know, they found 3.9 million AFib patients, they couldn’t give you their names of those patients. But they have lots of demographic information about
Nick van Terheyden
I think what it says is that, you know, these are the data, you know, large data blocks with all of the information provides really good content for generating insights through these machine learning artificial intelligence tools, that can essentially see signals in there that we, as humans, certainly the ordinary ones, I knew a few consultants that I worked for, that seemed to have that artificial intelligence that could walk into a room and just grabbed a patient just like that. But, you know, for the rest of us, they elevate us all by finding some of these things. I mean, that’s certainly certainly what I’ve been seeing, which, you know, makes for more interest, more excitement. Anyway, moving on. I, I came across an article that, you know, it stood out to me, let’s just be clear. It was a essentially a user device design issue as user interface design. And in this particular instance, it was the law court system for lawyers. And the user design had been so confusing, that a lawyer had clicked the wrong button and ended up divorcing his clients by mistake. And they were committed to fixing this. And it wasn’t a huge piece. But what was really interesting to me was, you know, that’s just divorce. And I don’t mean to diminish divorce as a thing. But you know, for those unfortunate enough, it’s, you know, it’s not life or death. But use a designer is like, a critical issue in healthcare. And this was just one example where, hey, you clicked the wrong button. We’ve seen some examples. I mean, you’re knee deep in this. So knee deep, you decided you’d write a book about it.
Craig Joseph
I did. I co wrote a book called Designing for health, the human centered approach came out last year. And we in it, we talked about a colleague of mine, and I talked about how easy a lot of this stuff is really, it’s not I mentioned a few minutes.
Nick van Terheyden
When you say easy, what do you mean, easy to get? Right? Is it easy to
Craig Joseph
get less wrong? Okay. Yeah, yeah, you don’t need a years of training and master’s degrees, and to be a brain surgeon, to figure out a lot of these things when you just look at them and understand some of the basic principles of human centered design and user centered design. It’s part of the problem is intentionality, right? You have to kind of sit down and think, Hey, who’s my who’s my target audience who’s going to be using, in this case, I think we were talking about some webpage or some sort of technology. So you know, who’s going to be using it the, if you’re 20 years old, the target for clicking a button is much smaller, you can make it much smaller safely, then if you’re 60 years old, or if you’re 80 years old, right? We can’t see as well when we get older, and our dexterity decreases and get a little shaky. And so the risk of those things happening are greater, but even that, you know, one of the very first things I talked about, that we talked about in the book is stealing from one of the one of the founders of human centered design, Don Norman, and he went and looked at doors, doors as in the things that you walk through to go from one room or one building to another.
Nick van Terheyden
Do you know, I just I’m so glad you clarified that I would have been confused. Otherwise, listen,
Craig Joseph
you just talked about grokking. And you also talked about having a B and a botnet, which I’m assuming there’s another phrase, okay, you know, most of us Yeah, so don’t don’t question me and trying to over communicate with you, Dr. Nick. I won’t have it. I won’t be part of it. adore and Don Norman as a diplomatic person, and he would never put it this way, but I’ll put it this way. If your door requires instructions for use, you have designed a bad door are right, okay, right. And so by instructions of us, if I if you need to tell me how to use your doors, if I need that be told that I should push on this side and pull on that side, that’s a bad door.
Nick van Terheyden
Can I just as a quick side note, then all toilet doors in aircraft need to be redesigned because I actually have a true story of somebody that and this happens all the time that got stuck in the thing, and literally took the whole door off the hinges because they got panicked and handed it to the flight crew.
Craig Joseph
I’ve explained myself with this doctor Nick and I was very, I was very anxious and I wanted out and I kept pulling, but I guess I was supposed to push. Alright, there are there may be are exceptions to that rule. But in general doors do not require instructions or should not require instructions. They’re just poorly designed and, and kind of once you have a very basic or elementary understanding of a lot of this, you can avoid accidentally causing your patients to become divorced. See how I did that? I just turned that right back into it
Nick van Terheyden
right back.
Craig Joseph
Right. It was perfect segue and I said patients, I think I met clients. That’s just where my my head is
Nick van Terheyden
used to design. All right, remaining time that we’ve got, wanted to talk about the impact the national physician advisor Conference, which I attended. And my first time I was not aware, although I feel like I’ve been part of this group for a time, certainly a period of time. Absolutely fascinating, hands down one of the best conferences from a content standpoint, it is essentially all the physician advisors. So this is a new group. And, you know, to give you a sense of it, they’ve only been in existence for 10 years, it’s you know, really new, and it is the clinical folks who perform the function in many facilities sometimes outsourced a variety of these approaches to the denials, the appeals that go on when a hospital bills for something and the insurance company or Medicare, Medicaid deny it based on failing to fulfill the necessary elements that are required to be able to be paid appropriately. And truly fascinating. I mean, all sorts of tremendous insights. I I felt like I was with like minded people and and a lot of them it didn’t hurt that it was in Coronado beach. And it was a little bit chilly. I’m just gonna say when when we did the sundown is on the beach, and the sun went down. Everybody scurried. scurry back to the hotel, because it was a wee bit chilly, as they say, in Scotland, but it was a tremendous, highly recommended content. And I learned a lot. I was a little bit concerned that, you know, this was revealing a number of secrets potentially. And you know, this seems it is almost competitive. I know it shouldn’t be. But if if we could work more together, it would be a better thing. But that wasn’t what was going on. So
Craig Joseph
these are the these are the doctors or others who are who are don’t work for the big insurance companies. Right? That’s what Yes, yes, exactly.
Nick van Terheyden
And, you know, I’m imagining that there’s probably a corresponding conference like there is in all these things where the other side of the coin I gotta say I would, I would be interested to attend that. Wondering if some of those folks who are attending this one, but
Craig Joseph
don’t know if that’s the case? Well, if we have any listeners who want to invite send out an invitation to you or myself. We are open to we are open to that.
Nick van Terheyden
Yes, absolutely. I would be delighted to attend and it doesn’t have to be in a fancy shmancy place. Although this was I pay Coronado. Fantastic. It wasn’t the whatever the ark, what’s the Coronado hotel I think, which is the one that’s most famous. It wasn’t actually at that but it was down the beach. I drove past the Navy SEAL training Beach, which is all fenced off and I asked my Uber driver, I said, you know, does it You ever see it? Gabby? Oh, he says only at night, he says comes alive. I’m thinking well, you know, hat hat tip to them. So anyway, we find ourselves at the end of another episode, exploring health care’s mysteries before they become your emergencies. Until next time, I’m Dr Nick and
Craig Joseph
I’m Dr Craig
