Nieman’s 80th Anniversary Reunion Weekend

Deborah Schoch:  Hello. You’re about to hear from a Harvard health expert. Well published, well known, who thinks like a smart, enterprising journalist. Dr. Anupam Jena, best known as Bapu, asks the questions that aren’t being asked. He delves deep into science and economics. He comes up with some unexpected and provocative answers that speak to how we live today.

Bapu has both an MD and a PhD in economics. He’s a practicing internist, and he still sees patients at Mass General. He is a prolific writer. He produces much‑quoted articles in the most prominent medical journals. Yet, he’s very different than many of the health policy experts that I’ve talked to over my career. I’m a proud member of the Nieman class of 2000.

I was telling them last night how surprised I was at how he understood journalism. In fact, he’d surprised me by drawing a direct line between his own esteemed academic research and journalism itself.

He told me, “The key ingredient is to have the idea. A good journalist writes well, tells a good story, but most important knows what story to tell.” I’ll leave him to explain how that works. Thank you.

Dr. Anupam Jena:  I’d like to make this a little bit different and make it interactive. The success of this next 10 minutes will depend on you chiming in. You sound like a vocal bunch. I don’t think it’ll be difficult. I want to draw a line between the type of work I do and what I think makes for good journalism.

That’s about telling a story and coming up with questions that people want to know answers to. Either because they’ve always wanted to know the answer to a certain question, or they never even thought that that was a question that was worth answering.

That’s a lot of what I do is try to come up with these interesting types of questions. I’ll give you a couple of examples of those today. One of the most difficult things is coming up with these ideas in the first place. That’s what I want us to do together today. Does that sound OK?

Audience:  Yes.

Dr. Jena:  If you participate, I’ll give you a surprise at the end, OK?

Dr. Jena:  What is this a photo of?

That’s an easy question. Now, I’m going to turn it over to you. Tell me some ideas. What’s the first idea that comes to mind? You can say, “I want to see whether or not people who wear Nike shoes have faster times than people who are wearing Reebok shoes.” Literally anything.

In fact, just use the word, marathon. It could be what’s the effect of watching a Harry Potter marathon eating Ben and Jerry’s. I don’t know, anything. Look at this photo and give me an idea. Just yell it out.

Audience Member:  How quickly do your knees blow out?

Dr. Jena:  How quickly do your knees blow out? By the way, who’s run a marathon? Who’s thought about running a marathon but said, “This is a really bad idea”?

Good. How quickly do your knees blow out? Let me push that even further. What are the long‑term health effects of running a marathon? Turns out that if you look at marathon runners and you measure their blood when they run a marathon,the levels of cardiac enzymes go up just like in a heart attack.

Running a marathon, obviously, is not a normal thing to do. It’s definitely not a normal thing to do for your heart. You could then imagine, if I look at these people 20 years later, do they have higher rates of scar on their heart and higher rates of arrhythmias, abnormal heart rhythms? The problem with that kind of study is that you know what you’re going to find.

People who run marathons are healthier. They’re actually going to have better long‑term health outcomes. We’re not in a situation where we can really randomize people to run marathons and randomize people to not run marathons. Then follow them out over their lives and see how their health outcomes differ.

It turns out, in reality, you might do that. For instance, New York had a marathon that was canceled several years ago because of bad weather. If I knew the names of people who were registered to run that marathon and wanted to run that marathon but for this completely random event happened not to.

I could look at 60‑day or 6‑month mortality rates from heart attack for those people who were supposed to run the marathon but, again, for a random reason, didn’t. That’s what we call a natural experiment. That’s not the idea I had in mind. What else? One more.

Audience Member:  How many different t‑shirts are there?

Dr. Jena:  How many different t‑shirts are there? How about for the next eight minutes you count and tell me what happens at the end?

How many different t‑shirts are there? That’s a good question. Focus on this photo. Think about the streets that are adjacent to this photo. Are they going to look empty or busy?

Audience Member:  Empty.

Dr. Jena:  They might look empty. In fact, if no one said empty I would have said that I heard someone say empty.

The streets would look empty. What ideas come to mind now? There are streets that are empty next to this busy photo.

Audience Member:  Impact on business.

Dr. Jena:  Impact on business. Good. What else?

Audience Member:  Traffic.

Dr. Jena:  Who said traffic? Don’t be shy? What about traffic?

Audience Member:  Rerouted traffic area.

Dr. Jena:  Good, rerouted traffic. Remember, my background is in medicine and economics. Let’s not focus the ideas on rerouting of traffic.

Audience Member:  Exhausted missions.

Dr, Jena:  Say that again.

Audience Member:  Exhausted missions.

Dr. Jena:  Exhausted missions, OK. What else?

Audience Member:  Stress.

Dr. Jena:  Stress.

Audience Member:  People dying in ambulances.

Dr. Jena:  People dying. He’s so morbid. Come on, man.

A few years ago, my wife ran a race called the Race to Remember. It starts in the Seaport area. It goes through Beacon Hill which is where Mass General Hospital’s is located, which is where I work.

She asked me to come watch her along the race route because it was her first time running a race. I drove down Storrow Drive, was going to get off at the MGH exit to park there and watch her. I couldn’t get off the road. I couldn’t get off the exit. It was blocked because of the race. I drove back home.

A few hours later, she says, “Gee, what happens to people who need to get to the hospital during a marathon with all these roads closed?” I thought that’s a really interesting idea. We’ve got 26 miles or more of roads being closed.

Any of you who’ve lived in a city that’s hosting a marathon know it’s a real pain to get anywhere on marathon day, particularly in this city. What do you observe? We looked at data from Medicare. Medicare is insurance for the old.

We had data on all people in Medicare who were hospitalized in US hospitals. We looked at 10 cities that hosted marathons over 10 years. This was published in the “New England Journal of Medicine.” It’s a big medical journal. It came out just a few days before the Boston Marathon.

What we found was that if you were hospitalized with a heart attack or a cardiac arrest on the day of a major US marathon, your mortality rate goes up by about 15 percent relative to the surrounding days. That’s that orange line, relative to the surrounding weeks.

If Boston happens on Monday, we look at the subsequent three or four Mondays and the preceding three or four Mondays. You find a 15 percent increase. You don’t find any changes in areas that surround the marathon route but that are not directly on it. These are people whose homes are along the marathon route. These are not runners.

These are older Americans. These are people who are unlikely to be running marathons but happen to have a heart attack or cardiac arrest on the day of a marathon and, quite possibly, can’t make it to the hospital in time. In these kinds of conditions, time matters.

That’s what we find. How would you show that this is true, that this is about delays in care? What kind of data would you need to show that?

Audience Member:  Ambulance response times.

Dr. Jena:  Oh my goodness! You guys are doing my job for me. I feel like running a clinic here. Ambulance response times. We were lucky enough to be able to get ambulance response times from these cities over this period of time.

What you can see in this orange line is that ambulance transport times, they go up about 20 to 30 percent on the days of marathons. Do you expect them to be higher in the morning, or in the evening, or both?

Audience:  Morning.

Dr. Jena:  Morning because the roads are closed in the morning. Low and behold, if you look in the evening, the line looks flat. There’s no increase in ambulance transport times in the evening, only in the morning when the roads are closed. What we, basically, find is that during a major marathon mortality rates go up for people who are not running marathons.

The purpose of this study was not to say that we shouldn’t hold marathons. Who was it that mentioned Taylor Swift? Someone talked about Taylor Swift. That we shouldn’t hold Taylor Swift concerts because we worry about these negative spill‑over effects related to traffic and so forth.

Keep in mind here that the number of people who we would project to die in Boston in any given marathon because they can’t get to the hospital in a timely fashion actually exceeds the number of people who died in the Boston marathon bombings.

This is a public policy point about how do we think about public health and emergency preparedness around these events. We focus on people who are at the events.

We never think about the people who might live around those areas but can’t get to the hospital in time because of delays in care. That’s one example. We have one more and then I’ll stop. What is this a photo of?

It’s a conference. Conference for who?

Audience:  Cardiologists.

Dr. Jena:  You said cardiologists like you have something against cardiologists.

Are you a dermatologist or what?

What ideas come to mind?

Audience Member:  Predominantly male.

Dr. Jena:  They are predominantly male. That is correct. You could say what’s the fraction of cardiologists that are male and how has that changed over time? Right now, we’re about a third female. What else?

Audience:  They’re white.

Dr. Jena:  There are a lot of white people there, that’s correct. [laughs] That’s accurate. To be honest, this is quite surprising. I’m Indian and at MGH, probably, a third of the cardiologists are Indian. This is not a representative photo, clearly.

What else?

Audience:  Drug money.

Dr. Jena:  Drug money, OK. You see an ad there for Repatha. Repatha is a new drug to treat high cholesterol. You might say, “All right, let’s look at rates of prescribing of Repatha in the weeks after this meeting because clearly it’s being advertised very heavily.” We’ve actually looked at that. We don’t see any increases in rates of Repatha prescribing after the meeting. What else?

Audience Member:  Mortality rates when all the doctors are at the conference.

Dr. Jena:  Mortality…What do you say? OK.

Gentleman in the front says mortality rates when all the doctors are at the conference. Let me get a raise of hands. Do mortality rates go up or go down? Who says up?

Up because all the cardiologists are out of town and it’s difficult to get care. Makes sense, right?

Who says down? We have a representative sample of 10 people who voted, half‑half. It turns out that mortality rates fall. They fall.

These meetings are attended by a lot of people. About 15,000 people attend these meetings every year. About 8 to 10 thousand cardiologists attend these meetings every year. Patients don’t know when these meetings are happening.

Obviously, they don’t know when these meetings are happening. Why is that useful for me? It’s useful for me because this creates a very nice natural experiment. A person has a heart attack. They have cardiac arrest. They’re taken to the hospital. They don’t know what the environment in the hospital’s going to be that day.

It happens to be in the middle of November or middle of March when these major cardiology meetings are being held. What happens if you have cardiac arrest? The normal trajectory of someone with cardiac arrest is as follows.

If you go to the hospital and you had cardiac arrest at home or somewhere else, you have a 70 percent chance of being dead within 30 days of that hospitalization.

It’s a serious condition. Even heart attacks, high‑risk heart failure, we’re talking about 20 to 25 percent of these people are dead within 30 days of getting to the hospital. For cardiac arrest, if you are hospitalized on the dates of a major cardiology meeting, your mortality rates fall from 70 percent to 60 percent.

That’s a 10 percentage point increase. Is that large or small? I’ll tell you this. If you take all the interventions that cardiologists have at their disposal to date, that mortality effect, if you sum up all those things, is less than 10 percentage points.

Everything that we do now in cardiology where there be stenting, aspirin, beta blockers, new cholesterol medications, whatever it is, it’s not as impactful as this right here. We’re talking about a 10 percentage point reduction in mortality if you happen to go to the hospital when cardiologists are out of town.

Why would this be the case? You think this should be the reverse, right? Really good cardiologists, leaders in cardiology are at this meeting. We would expect outcomes to be worse. Why might they be better?

Audience Member:  Less risky interventions.

Dr. Jena:  Less risky interventions. One more question, yes?

Audience Member:  [off‑mic comment]

Dr. Jena:  Yes, it could be that the composition of people who is there is different. Maybe, the ones who remain behind are the ones who see a lot of patients as opposed to the people, like me, who don’t see patients as often. [laughs] Outcomes could be worse. Those are the ones who are going to the convention. That’s possible.

It could also be that there’s a change in the number of risky procedures that are performed. Suppose I told you that the rate of stenting of the heart fall by about a third, a third if you have a heart attack during the dates of these meetings. That’s a large reduction. How might that actually benefit patients?

I’ll give you two examples. One is a 40‑year‑old guy. He’s a contractor. He smokes but no other medical problems. He’s working on the site. He develops chest pain. He is brought into the hospital. In the ED they do a EKG, a real one, not on an Apple watch. They do a real EKG.

It shows a particular type of heart attack has occurred. He gets rushed to the catheterization lab. He gets a stent. He’s discharged in two days. He lives a fine, happy, normal life. Take a 90‑year‑old woman. Develops chest pain at her nursing home. She’s on 10 different medications. She gets taken to the Emergency Department.

She has the exact same EKG, exact same blood tests. She’s having a very similar looking heart attack. She gets taken to the cath lab. She gets a stent. She dies within two weeks because of complications with the procedure. That’s not a difficult story to imagine.

I imagine that some of you would have family members or friends who have fallen into that category where an unintended consequence happened from medical care. The idea here is that, often when we practice medicine, there are black and white cases where we know what the answer is.

There’s a lot of area where we’re in the gray. Where we think we know what we’re doing and we hope what we’re doing is right. We may not be doing what’s right.

This is a way to use big data and interesting natural experiment in an observation that I had as a resident of how care looked differently during the dates of these meetings. To try to understand better what works and what doesn’t work in medicine. Who are the patients who we are helping? Who are the patients that we might be harming?

Let me do this, actually. I’ll give you one more example, just from some work we’re doing now. Then I’ll start with something else. $3.99, why do stores charge $3.99 for an item instead of $4.00?

People are more likely to buy the item for $3.99 than they are to buy for $4.00. Their increase in their likelihood of buying that item is much more than that penny difference. That’s called left‑digit bias.

Behavioral economists refer to this as left‑digit bias. You focus on the left digit, the three versus the four. How might you apply that to a cardiologist who’s seeing a patient in the Emergency Department? Any ideas? Focus on the age of the patient.

Let’s say you got a guy who’s 69 years old and 50 weeks versus 70 years old and 2 weeks. Do you think the cardiologist might look at those patients differently? They’re only four weeks different in age. Why might a cardiologist look at those two patients differently?

Audience Member:  The bias.

Dr. Jena:  Yeah, the bias, the left-digit bias. They may focus on the six. They’ll say, “This 69.9‑year‑old man. This is a 60ish-year-old guy. The 70.1‑year‑old, that’s a 70ish-year-old person.” As it turns out, the older you are the less likely doctors want to do procedures on you.

If you look at people who are basically the same age but who are only separated by two weeks. One person has their age beginning with a six, one person has their age beginning with a seven. There’s a 10 to 15 percent difference in the likelihood of getting a stent. The 69.9-year-old person is about 10 to 15 percent more likely to get a stent than the 70.1-year-old person.

What’s the next question you want to know the answer to?

Audience Member:  Who lives?

Dr. Jena:  Who lives? What do you think? Is there a difference in mortality between these groups? Let’s say, who said there’s no difference in mortality? Who says that because the 70.1-year-old person is not getting a stent, they are actually more likely to die? It turns out there’s actually no difference in mortality.

Again, this is a natural experiment that relies on biases that we have as individuals that shows that, at least, 10 to 15 percent of what we’re doing in terms of stenting people isn’t actually impactful, doesn’t affect patient outcome.

I want to end with this. I told Deborah, depending on whether or not you participated, I would do something different at the end. Bear with me. I know I put you in the hot seat.

Does anybody know sign language? Good. Perfect.

I just narrated the Bible.

I put you in the hot seat, so I thought it was only right for me to do the same for you. We have two kids. About a year ago, our son, Aiden, was born deaf in both ears. My wife, who is also a physician, she’s a radiologist at Brigham and Women’s Hospital, which is one of the largest teaching hospitals in the Harvard system.

She’s really been a force for us to learn American Sign Language. Why? Our hope is that one day, through education and our support, he can be in a room like this, sitting with people like you, doing what he loved to do.

The title of this talk was to think differently, to be open‑minded, to be creative. My hope is that you all do that and to give people the chance who, otherwise, you wouldn’t think about giving a chance to. Thanks.