What Happens to Patients When Thousands of Cardiologists Leave Town?

Bapu JENA: Back in 2015, I did a study about what happens to patients when cardiologists flock to their big annual conferences. Because when thousands of healthcare professionals all pile into a convention center, it must have an effect, right? Well, the results surprised me. And it sparked the young Stephen Dubner’s interest too. He asked me to come talk about it on Freakonomics Radio. And in a lot of ways, you can draw a line from that first conversation with Stephen to where we are now.

From the Freakonomics Radio Network, welcome to Freakonomics, M.D. 

I’m Bapu Jena. I’m an economist but I’m also a medical doctor. In each episode, I dissect a fascinating question at the sweet spot between health and economics. Today: When a lot of cardiologists leave town and head to their big annual conference, what happens to the patients who need care? 

*      *      *

JENA: I thought it would be fun to hop in a time machine today and have you listen to that conversation from Freakonomics Radio about the cardiology meetings study. And after that, I want to tell you about some hot-off-the-presses research my colleagues and I just published that was inspired by it. So stick around for that. 

But for now, let’s go back to 2015, when this conversation aired on Freakonomics Radio. Remember 2015? Thanks to the pandemic, it actually all feels like a long time ago now. Let’s see. In 2015, Barack Obama was still president

Barack OBAMA: Welcome to the fourth quarter of my presidency.

and the primaries were in full swing. People watched movies in theaters. 

Chris PRATT: You just went and made a new dinosaur? Probably not a good idea.

Jurassic World broke all box office records in its opening weekend. 

Bruno Mars’ “Uptown Funk” was pretty hard to avoid anytime you found yourself on a packed dance floor at a wedding, which I certainly did. And if you were willing to sell your house, you might have been able to get a ticket to see Hamilton on Broadway. 

Alexander HAMILTON: I am not throwing away my shot.

As for me, I had recently become a young dad and my cardiology meetings study had just come out days before my daughter was born. The reason I remember it so clearly was a reporter called me to talk about it when my wife was in labor. And yes, I did make the mistake of stepping out of the room to take that call. Bad move, folks, bad move.

Not too long after that though, Stephen called me up. And he wanted to chat about the study and how it fit into some episodes he was doing on how healthcare is delivered and when it works and doesn’t work.

*      *      *

Stephen DUBNER: Our healthcare system generates an enormous amount of data. Innumerable inputs. Innumerable outputs. For us laypeople, that can seem like a nightmare; for an economist, it’s a dream. For there’s a lot to be learned from a clever and robust analysis of all that data. Consider the data collected through Medicare, the government-run program that provides coverage primarily to Americans 65 and older. 

JENA: So, anytime a Medicare beneficiary receives any care, whether they see a doctor in an office, or whether they’re hospitalized, a claim is filed to Medicare for billing purposes.

That’s Anupam Jena. He’s an assistant professor of healthcare policy and medicine at Harvard Medical School. But Jena isn’t just an M.D.; he also holds a Ph.D. in economics. That dual training has very much informed the way Jena thinks about his work. When he was doing his residency, for instance, in internal medicine, a question came to mind. How helpful were the medical procedures being carried out by his elders?

JENA: And in some instances, it was pretty clear, at least to the residents in training, that a procedure may not have been appropriate for that patient.

Jena also wondered about the differences between doctors in a given hospital.

JENA: And so that got me thinking, well, what happens to patients when certain doctors aren’t around, or when, let’s say, doctors go away to conferences? What happens to their patients during those dates? 

Yeah, what does happen to patients when doctors go to medical conferences? When the doctor’s away, does the patient pay? Let’s think about cardiology, heart medicine. Every year, there are two major conferences for cardiologists in the U.S.: the American Heart Association conference, usually held in the fall, and the American College of Cardiology, or A.C.C., conference, typically held early in the year. This year, it was in March, in San Diego, a nice place to be after a long cold winter back east. Out of the 30,000-plus cardiologists in the U.S., more than 7,000 of them are estimated to be at each of these two meetings.

JENA: So it’s not a small number of cardiologists that attend. And the purpose is to present new research findings, to hear about old research findings, to recertify, to get more information about what’s up to date in the field.

Jena thought he knew what might happen to heart patients during these conferences. There have been a number of studies, he says, about what happens to hospital patients during off-peak hours.

JENA: So, if you’re hospitalized after midnight, or if you’re hospitalized on the weekend, as a patient, do you have worse outcomes? And the answer in most studies has actually been yes. Not all, but most of these find that patients receive—

Stephen DUBNER: They have worse outcomes, yes?

JENA: Exactly. Worse outcomes if you’re hospitalized during off hours. And so that was kind of a natural stepping stone to say, what happens if you’re a patient and you happen to be hospitalized with a really acute condition when cardiologists are out of town at a national meeting?

To answer that question, Jena, along with three co-authors – Vinay Prasad, Dana Goldman and John Romley – turned to the Medicare data. 

JENA: So, when a Medicare beneficiary dies, that information is also reported back to Medicare. And what that kind of information allows us to do is to say, when are they hospitalized? Meaning what date are they hospitalized? And what happens to you after you leave the hospital? Do you go to a skilled nursing facility? Do you go home? Do you make it past 30 days? Are you dead within 30 days? So all of that information is available for researchers to analyze.

They began to overlay patient data with the dates of the cardiology conferences, covering a 10-year stretch. They looked at patients admitted during those conferences and, for comparison’s sake, patients admitted in the three weeks before and after those conferences. They narrowed their analysis to the patients who were in really bad shape.

JENA: The idea is we want to pick conditions where a patient doesn’t choose to not come to the hospital because their particular doctor is away or because they have some knowledge that cardiologists are away during this time. And so we wanted to pick three really acute conditions. The first was cardiac arrest. Cardiac arrest is a condition where your heart stops beating completely.

DUBNER: Oops, hang on one second. 

Let me just say we were speaking to Jena back in February. I was in New York, and he was in Boston, which had just been hit by yet another monster snowstorm.

DUBNER: Hang on one second. I was hearing some background sound like a truck or something. 

JENA: Yeah, that was a snow plow that just went by.

You may hear a few more snow plows before this conversation is over. Okay, back to Dr. Jena and cardiac arrest.

JENA: Cardiac arrest is a condition where your heart stops beating. It’s not a condition that someone chooses to have, it just happens to you. By definition, you die. And you’re brought back to life. So it is the most acute thing that you can imagine happening to you. 

Jena and his colleagues also looked at heart failure. 

JENA: And heart failure, as you might know, comes in a number of varieties, but we looked at patients who have really severe heart failure. So, to give you a sense of the numbers involved, about 30 percent of these patients are not alive within 30 days of hospitalization. So it’s a pretty acute condition, a high mortality condition. And the last condition that we looked at was heart attack. The medical term is acute myocardial infarction, but it’s basically when one or more of the arteries that supply your heart has an acute blockage, and so blood doesn’t go to your heart. And again, the same kind of mortality rates. Thirty percent mortality almost at about 30 days.

The Medicare data covered tens of thousands of hospitalizations for these three conditions over the 10 years’ worth of annual cardiology meetings. 

JENA: It’s about as close to a randomized control trial as you could ever hope to get. These patients are nearly identical on meeting and non-meeting days. They’re the same age, the same sex, the same race. We look at 10 different chronic conditions that they have. They have identical percentages of each one of them. So they’re basically the same.

So, what did Jena and his colleagues find? How did the cardiologists’ absence affect patient outcomes?

JENA: We just assumed that the decreased availability of doctors would imply that outcomes would be worse. And that was our initial hypothesis, but what we found was the opposite.

The opposite – meaning that patients were less likely to die while the doctors were away. But only, we should point out, for certain high-risk patients in certain kinds of hospitals. Jena and his colleagues looked at teaching hospitals and non-teaching hospitals. The assumption is that teaching hospitals have more of the type of cardiologists who are likely to attend conferences, but that’s only a hunch. And that is where Jena found the surprise in his data. In non-teaching hospitals, the conference didn’t seem to matter. But patients who were admitted for cardiac arrest to a teaching hospital during one of the cardiology conferences were roughly 10 percentage points more likely to survive than if they were admitted on non-conference dates. Patients with heart failure — again, at teaching hospitals — were 8 percentage points more likely to survive during a cardiology conference. 

JENA: High-risk heart failure. What we found is that if you’re hospitalized on a cardiology meeting date, your mortality is about 17 percent at 30 days, 17-18 percent. Whereas if you’re hospitalized just a few days before or a few days after, your mortality is closer to 25 percent. So that’s a very large difference.

DUBNER: Wow. It just sounds so absurd. I mean, I’ve read the paper; I know what you’re going to say, but it still sounds so absurd. So, basically, if I have a major heart condition and someone gets me to a good teaching hospital— I live in New York. My hospital is Columbia Presbyterian. That’s a great teaching hospital. You’re saying that I have a better chance of surviving if there’s a cardiology conference going on and some of the top cardiologists are not there?

JENA: That is correct.

DUBNER: Oof. Okay, and describe for me overall the magnitude of this effect compared to, let’s say, standard, you know, cardiology treatment, whether it’s beta blockers or statins, angioplasties, stents. How much better off are you by having those interventions than you are by just simply going to the hospital when the cardiologists are not there?

JENA: So, just to give you a sense, the mainstays of treatment for heart disease are beta blockers, statins, aspirin, for some individuals, a blood thinner like Plavix. If you were to combine all those therapies together, we’re probably talking about reducing your mortality by about 2 to 3 percentage points.

DUBNER: Wow, percentage points, okay. And here you’re talking about 8 percentage points in one case and 10 percentage points in another.

JENA: Exactly. So these treatments are very effective, but they’re not nearly as large in magnitude as what we’re finding here.

Another finding in Jena’s paper: while patients with cardiac arrest and heart failure were less likely to die during a cardiology conference, there was no difference in outcome for patients who were admitted with a heart attack. The study did find, however, that these patients received far less invasive treatment – stents and angioplasties, for instance – when many cardiologists were away.

JENA: So, an angioplasty is a procedure in which a balloon is inflated in one of the arteries that supply the heart. It basically opens up the blood vessel that was clogged. That is different than stenting. Stenting is a procedure in which an actual stent is placed in the heart and opens up the artery. And it keeps the artery open. There’s nearly a third reduction in rates of angioplasty/stenting during meeting days. And remember, in heart attacks we didn’t find any difference in mortality. So at the very least, what this would suggest is that, look, we’re able to reduce these procedures by about a third, and yet we see no difference in mortality, in heart attacks.

DUBNER: What does the empirical research say about the efficacy of, say, angioplasty? In other words, we hear about these invasive treatments – angioplasty and stenting and so on, and we the public like to think that if doctors have gone to the trouble and researchers have gone to the trouble to come up with these things, of course they work really well – not only to come up with these things, but to use these things and to build for these things and so on. But talk to me for a moment about what we actually know or maybe don’t know about the efficacy of such interventions.

JENA: These interventions, for example stenting or angioplasty, are extraordinarily effective. I think if you look at the interventions that have been developed in the last 30 years for heart disease, they rank at the highest in terms of their innovativeness and their effectiveness. And most of the patients for whom these interventions have been studied have been what I would characterize as average to moderate risk patients. There have been some studies with very severe coronary artery disease, very sick patients for whom these interventions have been studied. And even in those studies, they do find average benefits for these procedures. So by and large, I think if you were to look at this study without any information about the results, you would think to yourself, by lowering rates of these invasive procedures we are likely to harm patients.

But, as we know, that’s not what the study found. Jena admits that, given the data, it’s impossible to point to an exact cause.

JENA: The strongest limitation of the paper is that we can’t tell you exactly what’s going on. So, what I can tell you with as close to certainty is that something is happening in the hospitals that is responsible for the lower mortality on meeting days. 

 And keep in mind it’s not as if there are no cardiologists in hospitals during annual meetings. Some docs cover for others; more junior staff might take over for a few days. Jena suspects that the doctors who stay behind may be more cautious.

JENA: What we’re identifying is that group of patients for whom the cardiologists who are left behind may have thought to themselves, this person may not be appropriate for this procedure. And the clinical decisions that they’re making are different then. And they could be different in a way that actually improves outcomes because they’re restricting procedures for those who are at the margin, who would be least likely to benefit.

DUBNER: And your explanation would be that these are docs who are covering, who might be more junior or whatnot, and they are less likely to order up what a more senior, confident, experienced doctor might order up. Is that one explanation?

JENA: That’s one explanation. Another explanation could be that the covering doctor says to himself or herself, “Look, I don’t want to do this thing because I’d be better off not harming the patient by doing a procedure. Why don’t I just wait and see what happens, simply because this is not my patient, I’m covering the patient for somebody else?”

And this, Jena says, brings us to the “less is more” dictum in medicine – which, he notes, is not universally embraced. 

JENA: The perception of healthcare is that by doing more we can improve health. And what we need to recognize is that so much of healthcare, so much of the clinical decisions that we make operate in this gray zone. It’s not black and white. And it could very well be the case that in the gray, less may be more.

DUBNER: Could it be that, you know, the doctors who are most likely to attend these conferences are those who are involved in research and that they, perhaps, aren’t as good at clinical care, and the ones who are left behind are maybe better? Is that a possible explanation?

JENA: That’s a possible explanation. I think the reason that we find our results predominantly in teaching hospitals is because if you look at the fraction of academic cardiologists who attend these meetings, and you look at the fraction of community cardiologists who attend these meetings, I think that the share would be larger among academic cardiologists. We haven’t been able to get any great data from the American College of Cardiology or the American Heart Association. But whatever data they do publish online would support that. And so that’s why it’s not surprising that we saw the effects there.

DUBNER: I’m curious, Dr. Jena, have you heard from either the American Heart Association or the American College of Cardiology, the two conferences that you measured?

JENA: Not formally. I think the American College of Cardiology released a statement, which was very well-worded. It basically said, it’s reassuring to know that during dates of national cardiology conferences, our patients receive no worse care, which is technically true.

DUBNER: Technically true. But I could see this being— I could see them playing it either way. They could take it as an indictment that they are, you know, they represent a bunch of people who don’t contribute to better care. On the other hand, they could say, “Hey, you know, our conference is a major life-saver. We draw away all these people who are doing too many procedures, and that’s saving a lot of lives.” But I can’t imagine they’ll turn that into their slogan, would you?

JENA: Exactly. I’ve tried to advocate that to my own chair — to let me go to more conferences — but that hasn’t worked.

*      *      *

DUBNER: We did ask the two organizations to respond to Jena’s study. Richard Chazal is the American College of Cardiology or A.C.C.’s vice president and a practicing cardiologist.

Richard CHAZAL: I think many of us were— actually, were reassured that there wasn’t an increase in risk during these meetings, because many people had hypothesized that might be the case, that when there was a departure of some physicians, that the staffing levels could result in an increased risk. So I do think it’s reassuring to the public and to you and to me that we can safely get our care at a teaching hospital during periods of time of meetings. So that part is encouraging.

 Pretty much as expected. Dr. Chazal did add this however:

CHAZAL: This is important and interesting information. The biggest concern here, and the one that we have to look at, is this finding, and to try to tease out, if we can in the future, why that is and how we modify our behavior so that we have a lower risk at all times.

The American Heart Association, meanwhile, sent us a recorded response from its president, Dr. Elliott Antman; we weren’t allowed to interview him. Dr. Antman says there is no evidence of cause and effect in Jena’s study, and that we should essentially think of it as a “calendar analysis.”

Elliott ANTMAN: So, the investigators happened to analyze the period of time when cardiology meetings were occurring. They could have picked Christmas. They could have picked New Year’s or Easter. We know that there are changes in the staffing schedule when there are holidays, when there are important national cardiology meetings. Bottom line for us at the American Heart Association, there’s nothing in this study that we see that would lead us to recommend a change in clinical practice. 

In my interview with Jena, I had raised the same point:

DUBNER: Okay. Is it possible that these cardiology conferences are perhaps typically held at the same time of year every year, which might be the time of the year that coincides with lower mortality?

JENA: That is definitely possible, but you know, as it turns out, over the nearly 10 years that we looked at the data, the cardiology conferences actually varied in their time. So it wasn’t that they were always at the same time of the year. They actually do vary slightly from week to week across the 10 years.

DUBNER: Right. I’m curious, has there been any similar research with psychologists or other mental health professionals? I’m wondering if anyone has ever looked at suicide and/or hospitalization or even depression outbreaks, if such a thing could be measured, during psychiatric conferences or maybe even just during August when all the shrinks in New York go to Cape Cod.

JENA: No, that’s a great question. I’m not aware of any. And I actually looked into this to see whether or not there was anything that was done. It’s something we’re going to do if it hasn’t been done after an exhaustive search, but it’s a great example. I’ll give you another example of something we have actually looked at, which we do know data about and I’m sure you’ve heard about, is this July effect. It’s the idea that patients who are hospitalized in academic medical centers in July have worse outcomes because the residents who are there are inexperienced. And by and large, what this literature has found is that the July effects, if they happen, are very small. And the question for clinicians was always, “Well, how is that possible?” How is it possible that something that every clinician thinks to be a big issue turns out to not be a big issue when you look at the data? And what we— the insight that we had is that every patient who comes into the hospital is different. And for most patients, it really does take a lot to lead to an adverse outcome. And so what happens if you focus on those patients who are the most sick? And what we did is look at the same kind of group of patients. We looked at patients who had heart attacks who are in the top third of predicted mortality. And these patients have a mortality rate of about, as I said, 25-30 percent. And what you see is that for those patients there actually is a July effect. So if you’re hospitalized in a teaching hospital with a very severe heart attack, you are 5 percentage points more likely to die if you’re hospitalized in July versus May. So, basically, 25 percent mortality versus 20 percent mortality. So it’s a large effect. Of course, it goes in the opposite direction of what we’re saying here in the sense that less is worse, but you know, that’s one example of people looking at what happens around specific times of year or specific types of practices.

DUBNER: How do you account for that contradiction?

JENA: I think the contradiction is going to be what is the clinical decision that’s being made. So, I keep on coming back to this in my mind. Why is this happening? And all I can come down to is I think that cardiologists are just making different clinical decisions during non-meeting dates. And I think the decision that’s being made is, is this patient appropriate for a procedure? Because whenever you do something, you have to have in your mind, is this person good or bad for it? And let me give you another example. If a cardiologist had 100 procedures to allocate, and that’s all they could do, I’m fairly confident that each one of the people that they allocated those 100 procedures to would do extraordinarily well. They’d have a beneficial outcome. But there is no constraint like that in reality. So you can go to the 101st, the 102nd.

DUBNER: That’s such an interesting way to think about it. And it makes me think that the economist part of your brain is intruding on the medical part of your brain there, and I wonder if you’re leading to some kind of relationship between cost and treatment and availability and supply and demand and so on. So, do you think that’s— forget about just cardiology for a moment— do you think that’s a major component of adverse outcomes that we’re looking at generally, including just, you know, the fact that we spend more— a larger share of our G.D.P. than any other country on earth I believe for healthcare, and yet our outcomes are super sub-optimal. How much do you think that is due to a kind of endless— almost an endless supply and a relatedly very high demand?

JENA: You know, I think the common concern is that, because doctors are paid fee for service, meaning they get paid for every procedure that they do, that it leads to over-incentives for them to do procedures. I’m actually not convinced that that is really the root of all the quality problems that we find. My hunch is that the reason that physicians may be doing more procedures than is clinically optimal is that they just don’t know any differently. If you think about what is it that impacts a physician’s decision, well sure, what they get paid impacts it, just like it would impact anyone’s decision. But what about where they went to medical school, what they learned during residency, whether or not they’ve been sued before? And most importantly, what is their level of risk aversion? I’ve got to imagine that that would translate somehow into clinical practice. I think that there are underlying differences in how people think, how cardiologists— how doctors think that drive these decisions.

*      *      *

JENA: That was an excerpt from the 2015 Freakonomics Radio episode number 202 called “How Many Doctors Does It Take To Start a Healthcare Revolution?” Or Freakonomics, M.D. Episode Minus 1, I guess. The Prequel. There’s some other great interviews in it if you want to check it out. We’ll put the link in the show notes. 

When we did that episode, the cardiology study had just come out and the findings were so strange that a lot of people, including myself, weren’t sure that it could be replicated. 

There was also the thorny problem that the study left a lot of questions unanswered. Like, who exactly was attending these meetings? Were they people who spent most of their time doing scientific research and not seeing patients? If so, maybe that could explain why replacing them for just a few days could improve patient outcomes. 

Since it was published, I have gotten some answers from two different studies my colleagues and I conducted. In 2018, we studied what happened to patients who have heart attacks during the dates of a different, but very specific cardiology meeting that is attended by doctors called interventional cardiologists. These are the types of doctors who specialize in procedures to treat heart attack patients. The general cardiology meetings that I’d studied initially are attended by all sorts of different cardiologists. Studying the effect of this specialized event kind of gave us a way to zoom in.

And we still found that patients with heart attacks had lower mortality rates if hospitalized during the dates of an interventional cardiology meeting. 

We didn’t see a difference in age or gender between the two groups of doctors. But here’s what we could measure: The people who were going to conferences were more likely to have attended a top medical school, more likely to have received research funding from the National Institutes of Health or led a clinical trial, and they had nearly three times as many publications, on average. 

We can infer that since the doctors who remained behind to treat patients tended to be less specialized in research, they may be more experienced and specialized in patient care. 

And I thought, maybe that’s what was happening in both studies. Patients may do better during these big conferences because it means they’re even more likely to be treated by those cardiologists who are oriented towards hands-on patient care. 

And here’s where that new study I mentioned comes in. It was just released this week, actually. My colleagues Hirotaka Kato, Yusuke Tsugawa, Jose Figueroa and I, we examined whether doctors who treat fewer patients per year have worse outcomes than doctors who treat more. We studied hospitalized medical patients because those patients typically don’t choose their doctors and vice versa, which means it’s a good natural experiment. And we found that those doctors who split time between seeing patients and on research, administrative roles, or other work — and so they treat fewer patients — those patients had higher hospital mortality rates than those treated by doctors who were clinically more active. 

Now, those of you who’ve been following this show from the beginning know that I am a doctor who splits time between seeing patients and doing economics research. So, I think there’s value to having people working and learning across disciplines and in hybrid roles, but there may also be costs. And these follow up studies point to something that economists, and probably all of us, have known for a long time: that specialization matters. Practice matters.

All right folks, that’s it for Freakonomics, M.D. this week. You can find links to all the studies we mentioned at freakonomics.com.

Also, please consider following the show. If you’re enjoying the show and leave us a review, you’ll help introduce the show to more people. 

If you have thoughts on the show, shoot me an email. That’s bapu@freakonomics.com.

Thanks for listening. 

*      *      *

Freakonomics, M.D. is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, and People I (Mostly) Admire. This show is produced by Stitcher and Renbud Radio. You can find us on Twitter and Instagram at @drbapupod. This episode was produced by Tricia Bobeda and mixed by Eleanor Osborne. Original music by Andrew Edwards. Our staff also includes Alison Craiglow, Greg Rippin, Joel Meyer, Jasmin Klinger, Emma Tyrrell, Lyric Bowditch, Jacob Clemente and Stephen Dubner. If you like this show or any other show in the Freakonomics Radio Network, please recommend it to your family and friends. That’s the best way to support the podcasts you love. As always, thanks for listening.