Stephen DUBNER: Hey Bapu!
Bapu JENA: How’s it going?
DUBNER: I like your podcast!
JENA: Thank you.
DUBNER: So why don’t you just say your name and what you do?
JENA: My name is Bapu Jena. I’m an economist and a physician at Harvard. I teach healthcare policy and health economics. I see patients at Massachusetts General Hospital, and I’m a professor at Harvard Medical School.
DUBNER: As if you need another job, you’re getting ready to host a new podcast for the Freakonomics Radio Network. We’re about to play, now, for our listeners a pilot episode of your new show — which I’m incredibly excited about. But first, let me just ask — “Bapu” isn’t the name on your birth certificate, is it?
JENA: Is this being recorded for legal purposes? No. My first name is Anupam. Bapu is my middle name. It’s really more of a nickname. It’s probably on every legal document except for my birth certificate. So Bapu means “father” in a variety of different Indian languages. It’s what they used to actually call Mahatma Gandhi. Not that I’m trying to draw any — there’re important distinctions to be made.
DUBNER: Bapu, how many people are there in the world like you that have both an M.D. and a Ph.D. in economics?
JENA: Oh, in the world, I’d say, I don’t know, maybe 10 to 20.
DUBNER: So, I guess you could look at that two ways. One is you’re a very, very, very, very rare bird, and that’s awesome. Or you could look at it from the demand side and say, if there’s so little demand for that, it must be a waste of time.
JENA: I know, exactly, when you see so few people going into something, you’ve to wonder why, unless there’s some market power or something, that’s the only thing, unifying explanation.
DUBNER: So maybe that’s what you are after then.
JENA: Yeah, exactly. I’m on a quest for market power.
DUBNER: You’ve a bit of history with Freakonomics Radio. Can you recall what you have told our audience in the past?
JENA: So, I’ve been on a couple times. The first time was about a study that I had with some others looking at what happens to patients who’re hospitalized during the dates of national cardiology conferences, when cardiologists are away, they’re out of town at these meetings. We found that patients actually do better. Their mortality rates fall during the dates of those meetings. So that was the first time I was on the show. And then you did a really nice series on “bad medicine,” which I was on a couple times. And I presume you did not ask me to join because I’m reflective of bad medicine.
DUBNER: No, we did not. So Bapu, as you know, we love economists around here and we also love doctors. They each seem to have their own intellectual superpowers, so is it fair to say that you’re both Superman and Batman?
DUBNER: Although Batman does not actually have superpowers, does he? Other than common decency, which does seem like a superpower these days. The way that you ended up blending the tools of medicine and economics in your research, where did those curiosities come from?
JENA: A lot of the work that I do is sort of like Freakonomics meets medicine. It’s questions that relate to medicine and healthcare. That’s criteria one. Criteria two is that it often requires large amounts of data like economists are very familiar with. Criteria three is that there has to be an approach that is causally valid. I’m not interested in associations for the most part in my research. I really want to know whether X causes Y and a lot of those tools that economists use, I implement in my work.
DUBNER: Some of your research highlights the fallibility of doctors or at least the fact that they’re as human as the rest of us. I’m really curious to know how that’s received in the field — when you write a paper that shows that older doctors, for instance, that their skill set seems to deteriorate.
JENA: It’s certainly the case that when that study came out, I got emails from doctors who had tons of clinical experience who said, “This just absolutely can’t be true.” To which I responded and I said, “It may not be true for you, but that’s not really the question I’m trying to answer. What I’m trying to answer is if we took 1,000 doctors above the age of 65 and 1,000 doctors between 35 and 40, and we randomized patients to those two groups of doctors, where would we expect to see better outcomes? And we’d expect to see better outcomes in the doctors who are younger.” Now, there are certainly going to be doctors who are older, who have a lot of experience, who maintain high volumes in their clinical practices that would have superior outcomes. And maybe the doctors who emailed me would fit into that category. But it’s certainly possible that some of the people who emailed me didn’t. And maybe that’s why they had time to email me.
DUBNER: Why did you want to make this podcast, other than the fact that I called you and begged you to do it?
JENA: Yeah. I don’t have anything else to do! You know, I think that there is a real interest in the intersection between economics, human behavior, and medicine — and that is exactly my sweet spot.
DUBNER: So, the idea here is that every episode, you’ll dive into a single medical study — including some studies that were done by you and your colleagues and others done by other researchers. Why’s that the right way to talk through an issue?
JENA: I think the structure of a study, it follows the structure of a podcast. It starts with a question. And then, what would you need to answer that question? What would the data need to be to answer that question? All right, now you’ve the data. How would you answer it? What approach would you take? All right, you’ve an approach. You’ve an answer that’s tentative. Now, how do you know that that’s the right answer? What sorts of ways do you’ve to interrogate your approach to make sure that you’ve gotten to the right conclusion? That’s exactly what researchers try to do in every study that they publish. And then the last thing is: So what? What do we do about this? And that’s the most interesting part, coming up with those implications.
DUBNER: Now that you’re a medical doctor and a Ph.D. economist and a podcast host, how much harder would you say it is to be a podcast host than to be just a doctor or an economist?
JENA: Oh, I mean, I think being a podcast host is the most incredibly difficult thing in the world. It requires immense skill, intelligence, and really defining good looks. Is that, is that correct?
DUBNER: That’s pretty much the answer I was going for, yeah. And humility.
DUBNER: Glad we’ re on the same page there.
* * *
I’m going to call today’s episode: On your mark, get set, croak!
In Boston, where I live, there’s a big annual event that I love to watch every spring. The Boston marathon. It’s the oldest annual marathon in the world. And they’ve run it every year since 1897 — except last year, because of the pandemic. This year, it’s been delayed until October. But it still feels like marathon season here in Boston.
Every year, the marathon draws about a half a million spectators along its 26-mile route that starts in the suburbs and moves its way all the way into downtown Boston.
Now, I’ve never run a marathon. In fact, the closest I have probably come to it is watching a Harry Potter marathon on T.V., which I recognize is not quite the same. But my wife runs. And although she has never run a marathon, about five years ago, she ran a 5K charity race here in Boston. The route for that race went right by Massachusetts General Hospital, which is the big teaching hospital where I work. I’ve got a parking spot there, so on the day of the race I decided to drive into the city and park at Mass General so I would’ve a good spot close to the route.
It was my wife’s first time doing a race like this and it was important to both of us that I be there to cheer her on. But as I drove to the hospital, I hit a snag: It turns out that one of the main roads was blocked off because of the race. The hospital was just a few blocks away, but there was no way for me to detour around that roadblock. I was cut off from the hospital — and from my perfect spot — because of the race.
So, I was cursing my decision not to take public transportation, and I turned around and headed back home. Hours later, my wife got home after finishing the race. And understandably, she was a little bit surprised that I hadn’t shown up and I felt horrible that I had missed the race.
But when I explained to her what happened — and she’s a radiologist, by the way — she thought about it for a minute. And then she said: I wonder what happened to all the other people who needed to get to Mass General this morning? And I thought: What did happen to all those people who needed to get to the hospital? Especially the patients. What if someone had an emergency?
I’m a doctor, but I’m also an economist. And nothing gives me more of a thrill than a question that really gets both sides of my brain going. That doesn’t happen every day. Here, one of those questions had fallen into my lap. And I only had to miss my wife’s race to get it.
There’re thousands of marathons held every year around the world. And anyone who has lived in a city hosting a marathon knows how disruptive they can be — like my wife’s 5K, but in the case of marathons, it’s obviously much worse. Blocked roads for miles. Impossible to get around. Traffic grinding to a halt on the morning of a race. For patients who need to get to the hospital quickly, those sorts of disruptions could be a recipe for disaster.
So I wondered: would it be possible to measure whether marathons cause delays in treatment? Because if I could measure that, I would have a way to measure something else: How much of a difference does that delay make in the survival of those patients.
So I put together a small team of researchers at Harvard to analyze the data. We made a list of the 11 biggest marathons in the United States. We included cities like New York, L.A., Honolulu, and Chicago. And we mapped their routes. Then we looked at more than a decade of Medicare data to find people who lived in ZIP codes along those race routes who happened to have either a heart attack or a cardiac arrest on the day a marathon was held in their city.
We looked specifically at data from Medicare patients because that automatically narrows it down to the over-65 crowd. And the over-65 crowd, simply because they’re older, are the group that’s most likely to have heart attacks and cardiac arrests.
Now, why did we focus on heart attack and cardiac arrest as opposed to any other sort of medical care? Well, for one, they are emergencies, and that means they are random — people don’t choose when they happen.
The second reason we looked at heart attacks and cardiac arrest is that these are serious business. A heart attack is when a blood vessel that supplies your heart is blocked off. There is a clot that forms and blood can’t get to the heart, and your heart muscle dies. So it’s a big deal.
A cardiac arrest is an even bigger deal. It’s when your heart stops pumping blood to the body. It’s the closest you can come to being dead. Sometimes we’re lucky to bring people back to life from a cardiac arrest. So these’re both really serious conditions, and the treatment is really time-sensitive. Every minute counts.
Just like me trying to get to my parking spot to watch my wife’s race, patients who live near a marathon could get cut off from the nearest hospital by blocked roads and detours. So, here was our goal: We wanted to see if living near the marathon route made it more likely that you would die if you had a heart attack or a cardiac arrest on the day of a race.
We looked at data from marathon days. And because heart attack mortality can actually vary day by day, we compared mortality on marathon days to data from the same day of the week as the marathon, but in the five weeks before and the five weeks after the race. The non-race days basically served as a control group. So, if the race fell on a Sunday, we looked at the five Sundays before and five Sundays after the marathon.
Now, it is possible, though really unlikely, that heart attack mortality could be higher on race days than non-race days for reasons that’re unrelated to marathons and ambulance delays. So, to address this possibility, we had a second control group as well. We looked at similar patients on marathon and non-marathon days who lived in ZIP codes just outside areas affected by the race routes. These are patients who shouldn’t have been affected by any marathon-related delays.
All in all, we looked at 1,145 patients who were hospitalized for a heart attack or cardiac arrest in the cities that we studied. And we compared them to just over 11,000 hospitalizations for patients on non-race days.
What did we find? For patients in areas near marathon routes, the percentage who died within 30 days of being hospitalized — the 30-day mortality rate — was 13 percent higher if they had a heart attack or cardiac arrest on race day than if they had either one of those conditions on a non-race day.
We didn’t find any increase in mortality on marathon days in those people who lived in nearby ZIP codes that were unaffected by the race route, which makes a lot of sense. Their trip to the hospital shouldn’t have been affected by blocked roads.
You might be thinking — correlation is not causation.
When we do a study like this, we’ve to be sure to eliminate all the other possible reasons for the effect that we are seeing. So what’re some of those other possible explanations?
What if the people having heart attacks were actually running in the race? Well, we studied patients aged 65 years and older — and, okay, there’re lots of runners who’re over 65. So we looked at people with multiple medical conditions — people who were chronically ill and therefore were really unlikely to be running a marathon.
Or what if, for some reason, the patients having heart attacks on marathon days were just different from patients on any other day? It doesn’t seem plausible, at least not to me, but to double-check, we compared patient characteristics — like their age, or other cardiac problems. And we found that those characteristics were about the same on race days vs. non-race days.
What if hospitals are short-staffed on marathon day? That didn’t explain it either, because hospitals were performing all of their typical cardiac procedures on marathon days. Which suggests that there were plenty of people on hand to care for heart attack patients.
It occurs to me that those folks must have gotten an early start on their commute so they didn’t get stuck in traffic!
What if ambulances took patients to hospitals that were further away to avoid roadblocks. We found that the hospitals that patients were taken to were actually the exact same on marathon days and non-marathon days. It just took patients longer to get there.
All we were left to explain [that] the difference in mortality was a delay in treatment.
So what kind of delays are we talking about here? On a typical, non-marathon day, the average travel time in an ambulance for patients with a heart attack or cardiac arrest was 13.7 minutes. On a race day? That went up to 18.1 minutes.
That means, on average, it took about four and a half minutes longer — 32 percent longer — for patients to get treatment on the day of a marathon. That may not seem like a long delay for your average commuter, but even small delays in care can lead to significant heart damage — which, by the way, is why we say in medicine that when it comes to heart attacks, “time is tissue.”
It’s also worth noting that we showed that ambulances were delayed. But in our data, about 50 percent of people who had heart attacks did not arrive by ambulance. Someone drove them.
Those people likely faced even larger delays because private cars can’t do the same things ambulances can do, like going through red lights, or breaking the speed limit.
Just to recap: what we found was that even small delays in care lead to 13 percent higher mortality in these patients. And that finding relates to one of the fundamental questions we ask ourselves in medicine: how fast do we need to act when someone is sick? Do we have days, hours, or just minutes?
The gold standard for experiments in medicine is the randomized controlled trial. In that kind of trial, we identify a big group of similar patients, and some of them get one treatment while others get a different treatment, or in some cases, even a placebo. The patients don’t know which group they are in, and neither do the doctors.
Now, we could never have conducted a randomized trial where some heart attack patients were told, “Hey, hang on a couple of minutes,” and others were given treatment immediately and then we saw who did better. That wouldn’t be ethical, because we already know that acting fast can save lives — like I said, “time is tissue” when the heart is under this kind of stress. But we just don’t know exactly how fast we really need to be.
This study actually gave us what economists call a natural experiment to help us answer that question. In a natural experiment, conditions out in the world, that we have no control over, randomize patients for us, if only by accident.
In our study, the combination of randomly timed emergency events, and highly time-sensitive medical conditions, allowed us to answer the question of just how time-sensitive this care is.
So, if there’s one thing I want you to take from this study, it’s this: If you’re experiencing chest pain, don’t hesitate. Call an ambulance.
For me, this study was right in the sweet spot where economics and medicine converge. And it was all prompted by my wife’s observation. Sometimes it just takes a random thought — or a random traffic snag — to get the mind going on a bigger question.
And it’s questions like these that we’re going to explore in this podcast. Like, what can supermarket-pricing tactics teach us about which patients get cardiac surgery? And what does that tell us about the problem of overuse in medical care?
Questions like: why children with summer birthdays are more likely to get the flu. And what that finding shows us about the importance of making healthcare more convenient.
I’m hoping that this podcast will give you some of the analytical tools that will help you look under the hood of the latest medical headlines, distinguish correlation from causation, and weigh costs and benefits in your own healthcare decisions.
I’ll end this pilot with a public service message: If there’s a race happening in your city — and someone you love happens to be running in it and is really counting on you to be on the sideline, even though it’s only a 5K and it’s over in about half an hour — don’t count on your normal parking spot. Go really early. Or take public transportation, okay?
* * *
Freakonomics Radio is produced by Stitcher and Renbud Radio. This episode was produced by Matt Frassica. The Freakonomics Radio Network staff also includes: Alison Craiglow, Greg Rippin, Joel Meyer, Tricia Bobeda, Mark McClusky, Rebecca Lee Douglas, Zack Lapinski, Mary Diduch, Morgan Levey, Brent Katz, Jasmin Klinger, Emma Tyrrell, Lyric Bowditch, and Jacob Clemente. Our theme song is “Mr. Fortune,” by the Hitchhikers; the rest of the music was composed by Luis Guerra. You can get all the Freakonomics Radio Network shows on Apple Podcasts, Spotify, Stitcher, or wherever you get your podcasts.
- Bapu Jena, professor of healthcare economics and healthcare policy at Harvard Medical School, physician at Massachusetts General Hospital, and host of the Freakonomics Radio Network’s newest podcast series.
- “Physician Age and Outcomes in Elderly Patients in Hospital in the US: Observational Study,” by Yusuke Tsugawa, Joseph P. Newhouse, Alan M. Zaslavsky, Daniel M. Blumenthal, and Anupam B. Jena (BMJ, 2017).
- “Delays in Emergency Care and Mortality during Major U.S. Marathons,” by Anupam B. Jena, N. Clay Mann, Leia N. Wedlund, and Andrew Olenski (The New England Journal of Medicine, 2017).
- “Mortality and Treatment Patterns Among Patients Hospitalized With Acute Cardiovascular Conditions During Dates of National Cardiology Meetings,” by Anupam B Jena, Vinay Prasad, Dana P Goldman, John Romley (JAMA Internal Medicine, 2015).
- “Bad Medicine, Part 3: Death by Diagnosis (Ep. 270),” by Freakonomics Radio (2016).
- “Bad Medicine, Part 2: (Drug) Trials and Tribulations (Ep. 269),” by Freakonomics Radio (2016).
- “Bad Medicine, Part 1: The Story of 98.6 (Ep. 268),” by Freakonomics Radio (2016).
- “How Many Doctors Does It Take to Start a Healthcare Revolution? (Ep. 202),” by Freakonomics Radio (2015).