Search the Site

Episode Transcript

Many of you listening are probably used to an eight- to ten-hour workday. That’s standard for lots of Americans. You’re also probably aware that doctors tend to work longer shifts. Especially during residency, that training period that comes right after medical school and which can last up to seven years.

Dr. BAILEY: Your first shift starts now and lasts 48 hours. You’re interns, grunts, nobodies, the bottom of the surgical food chain. You run labs, write orders, work every second night until you drop, and don’t complain!

Forty-eight hours might be a slight exaggeration. But until work hour reforms were passed in 2003, it was really normal for residents to have 36-hour shifts. The reforms were tightened again in 2011, and capped workweeks at 80 hours. They also capped shift length based on seniority. Interns, the lowest-level residents, couldn’t work longer than 16 hours. But upper-level residents could still work up to 24 hours straight — with up to four additional hours for transitions in care. So, even with those restrictions, some residents could still work 28 hours in a row as part of their 80 hours each week. Now, it’s not hard to imagine that those crazy long shifts could have downsides — not just for sleep-deprived doctors, but for patients.

Jeffrey ROTHSCHILD: In other industries, work shifts of that duration are associated with fatigue and drop-offs in performance. And so, we did a study that looked at a schedule where some residents worked as long as 36 hours, compared to a schedule in which doctors worked less than that, with a maximum of 16 hours a shift and more commonly, a 12-hour shift.

That’s Jeffrey Rothschild. He’s a physician at Newton Wellesley Hospital, just outside of Boston, and he’s a professor at Harvard Medical School. The study he’s describing is from 2014. And the findings are pretty intuitive.

ROTHSCHILD: We showed that the shorter duration shifts resulted in reduction in medical errors. So, that was the good side of that study. But there were some unintended consequences.

Solving real-life problems is often more complicated than it looks on paper. Implementing what seems like a simple solution can end up creating new problems. In this case —

ROTHSCHILD: When you have to cut back on the amount of hours in a shift, you still need the same person power. So, we had a lot more handoffs as a result of that study.

In other words, instead of having one doctor for 36 hours, you might have two or three doctors covering that same 36-hour period. And why is that a problem?

ROTHSCHILD: Poorly done handoffs resulted in more errors and adverse events.

From the Freakonomics Radio Network, this is Freakonomics, M.D.

I’m Bapu Jena. I’m an economist and I’m also a medical doctor. Each episode, I dissect an interesting question at the sweet spot between health and economics.

Today on the show: How do handoffs between healthcare workers affect patients? Affect outcomes? And what can we do about it?

Adrienne SABETY: There are a number of additional frictions that we’ve set up in this health system that make it even harder for these transitions to occur seamlessly. 

Jeffrey Rothschild is what we call a hospitalist, a doctor who works exclusively on inpatient care within a hospital, typically in shifts. That’s important because end-of-shift transitions in patient care happen all the time in a hospital. and they can be dicey. There’s always a risk of some information being lost — something being missed in translation. Studies have suggested, for example, that aside from medical errors, mortality itself could be higher for patients whose care is handed off from one medical team to another—not because those patients are sicker, but because they happened to be admitted to the hospital in the final days that a team was scheduled to work.

ROTHSCHILD: I certainly can give anecdotes of my own experience. I’ve had a patient in which they were slowly bleeding and I had to make sure their future blood count did not drop more precipitously. That would require either a transfusion or an urgent procedure done during the night. My shift ended around 7:00 p.m. I was there during the day. And I asked my colleague, who was going to be working overnight from 7:00 p.m. to 7:00 a.m., to check on that blood count, which would have returned maybe at 9:00 p.m. or 10:00 p.m., and compare that to the previous results to make sure the patient was stable and not bleeding any further.

Thankfully, in this case, the patient did okay — the bleeding stopped, and the handoff didn’t cause any problems because Jeffrey communicated with his colleagues effectively. But, what if he hadn’t? What if he was rushing to get home at the end of his shift and forgot to tell his replacement to compare the patient’s new blood count with the previous results —  and the patient’s bleeding didn’t stop on its own?

ROTHSCHILD: Communication failures are one of the more common causes of harm to patients. Whether it’s between physicians or nurses or other healthcare workers, there’s a tremendous amount of inconsistency on the information that’s shared when patient care is handed off from one person to another.

So, Jeffrey and his colleagues came up with an intervention — a sort of tool — to reduce hand-off related medical errors. Their goal was to make the information that’s shared during transitions more consistent. And they designed a study to test its efficacy. But before we talk about Jeffrey’s work, I wanted to get a broader view of the effects of end-of-shift transitions in hospitals.

So, I sat down with another researcher, who’s looked at this question in a different — and unique — hospital context: the emergency department. There, he found that end-of-shift transitions also had costs, but for other reasons.

David CHAN: My name is David Chan. I am a professor of health policy here at Stanford University. And I am a hospitalist at the V.A. Palo Alto Medical Center.

Bapu JENA: You’re also an economist, right?

CHAN: That’s right.

JENA: You embarrassed to say that in public or what?

CHAN: No, not at all. No, I’m in good company.

JENA: All right. How did you get interested in this? What led you to look at this question in the first place?

CHAN: I’m an internal medicine doctor, and we’re on the receiving end of getting admissions from the emergency department. As an internist, when you get to go home is a function of what the emergency department doctor does. And, you know, there’s certain patterns that you could just notice descriptively. Like around seven o’clock, you get a whole bunch of admissions from the emergency room. That makes you wonder. And when you start looking into the data, you see when patients arrive near an emergency room doctor’s end of shift, versus in the beginning or middle of shift, they’re treated differently, even though these patients are as good as randomly arriving at different times relative to the emergency department doctor’s shift. The key finding was that if you arrive near the emergency doctor’s end of shift, you’ll end up spending 67 percent less time in the emergency department. You’re more likely to be transferred upstairs and admitted. And as a result, much more money is going to be spent on you, more tests are going to be done. The emergency department doctor could have just watched you longer, but instead appears to be rushing patients out the door as they’re nearing their end of shift.

JENA: If patients come towards the end of a doctor’s shift, they get tests and procedures done more quickly, perhaps more of them done, they get admitted to the hospital — that wouldn’t necessarily be a bad thing, right? So, I presume you’ve looked at the outcomes for the patient?

CHAN: So, you’re shortening the time that the patient spends in the emergency department, but you saw a shifting from emergency department resources onto broader hospital resources, and there was no change in overall patient outcomes such as mortality, or whether a patient revisits the hospital. So, it’s not very efficient. You would end up spending quite a bit more money whenever you admit a patient and from the patient’s point of view, you might spend a night at the hospital, whereas you could have been sent home.

JENA: So, in other words, no better outcomes, but you’re spending more resources to get the same thing.

CHAN: Exactly.

JENA: So, what is it about the end of shift that leads an emergency room doctor to hurry things and maybe order tests and procedures that they otherwise would not order, if that same patient had come in earlier? Because they can obviously just hand that patient off to another doctor who will overlap with them.

CHAN: That could happen, but we, in practice, don’t see patients being handed off between emergency department doctors. And I think that it has something to do with imposing work on someone else. If you’re handing work off to another emergency doctor, that is somebody that you know, and work with, as opposed to admitting the patient to another part of the hospital — that’s handing off work to somebody that you actually don’t know.

JENA: So, let’s say that, like, I’m an emergency room doctor. I, by chance, have some patients that arrived in the emergency department towards the end of my shift. If the doctor who I’m going to hand those patients off to just got there, that feels like a different scenario than if the doctor I’m going to ultimately hand that patient off to maybe has been there for a few hours. Does it matter the degree of overlap?

CHAN: Yeah, absolutely. If you have a lot of overlap between those two doctors, you can slowly taper your work as you’re nearing the end of shift. And you might think that it would be inefficient — you know, if you’re hired to spend eight hours on shift, you shouldn’t be tapering your work as you’re nearing the end of those eight hours. But we find that if doctors have the three hours overlap, then they can have it be a more gradual process, as opposed to a sudden process of offloading all your work onto somebody else. The title of this paper was actually called “The Efficiency of Slacking Off.” There’s a trade-off here.

JENA: When you say slacking, what exactly are the E.R. doctors doing that you would characterize as slacking?

CHAN: It’s a bit of a tongue-in-cheek — usually we think that slacking is being lazy and not doing work when you should be doing work. So, there’s a bit of an irony that slacking is actually efficient when you have the incentive to go home. That’s kind of the fundamental misalignment of economic incentives. The emergency department doctor really values going home and the hospital wants to make sure that patients are well taken care of with a good use of resources. If you don’t have that much work to do near the end of your shift, then you don’t have as much of a conflict that you otherwise would if you’re really busy near the end of shift.

JENA: So, then one prediction would be that you might see more shirking behavior at the end of the shift on those occasions where the value to you as a doctor of not being in the hospital is greater, if it’s like your kid’s birthday or a nice day outside. Is that accurate?

CHAN: Exactly. As part of the publication process of this paper, we did a heterogeneity analysis where we looked into times when your end of shift was like 3:00 a.m. — so, if you go home early, everybody else is sleeping, there’s no restaurant to go to — versus if your end of shift is kind of prime-time hours, like getting home in time for dinner. And we find that, indeed, the end of shift effect is stronger when your end of shift is during a time when other people are awake and there’s other things for you to do.

JENA: So, what happens at the end of the shift when the season premiere of The Bachelor is coming on? It’s like a nightmare, right? Every patient is going to be admitted to the hospital within three minutes.

CHAN: You know those E.D. doctors.

JENA: So, is that sort of tapering common? Do you typically see a lot of overlap between doctors in the emergency department?

CHAN: You do.

JENA: You’re telling me that they didn’t need an economist to tell them how to run the hospital?

CHAN: Yeah. You know, I think they figured it out to some extent. They might not be running regressions, but they might have anecdotal stories about how, if you don’t have very much overlap, that transition was just brutal.

JENA: You know, You and I are both economists and physicians, and I’ve always felt that a lot of the stuff that I do comes from things that I see in the hospital that most people probably wouldn’t pay much attention to. And I bet if you were to talk to an emergency room doctor or another internist, they’d be like, “Yeah, of course that happens.” And people say that about my research, but nonetheless, we can both make careers out of stating what’s obvious to other people.

CHAN: It’s intuitive. It makes sense, but actually showing it is kind of cool and neat.

So, two different hospital contexts, two different issues caused by end-of-shift transitions.

After the break, we’ll get back to this question of information loss during handoffs and, as promised, we’ll discuss the tool that Jeffrey Rothschild designed to mitigate it. But before that, we’ll hear from another researcher who studies the impact of a very different kind of handoff between doctors — one that happens outside of a hospital, and on a very different timescale.

Adrienne SABETY: “Something like shift changes, those are very short relationships. That’s just someone that you’ve met for that day or that weekBut in our study, it’s really thinking more about these longer-term relationships that have taken a lot of time to build.”

I’m Bapu Jena, this is Freakonomics M.D., and we’ll be right back.

*          *          *

Before the break, we heard about two types of transition costs that can happen at the end of a doctor’s shift in the hospital: preventable medical errors due to communication failure, and unnecessary medical care.

But for most of us, if we’re lucky, the majority of our interactions with physicians don’t take place in a hospital. They take place in a doctor’s office. With set appointment times, shift ends don’t matter that much. There’s another type of transition that can occur though.

SABETY: The key nugget here is that many patients, myself included, value having a relationship with their physician.

That’s Adrienne Sabety. She’s an economist at the University of Notre Dame, where she studies the social determinants of health, like access to and quality of healthcare.

SABETY: You know, my grandma loved her P.C.P. — primary care physician — had a relationship with this physician for over a decade, and would do anything she asked. I also think the physician, because of that relationship, was really skilled at navigating the different specialists that my grandmother saw. And this enables potentially more continuity of care.

Are you close with your Primary Care Physician? If so, then you know that an ongoing relationship helps to build trust. And it might also allow your doctor to develop a better understanding of you as their patient, so they have a better sense of when things are out of the norm.

SABETY: And so, it seems like, from the patient perspective, there is value in this relationship. From the physician perspective, that’s true as well. But as researchers and policymakers, and even practitioners, we don’t actually have the numbers to say is this relationship valuable? And so, we decided to try to answer this question by using this transition of losing your relationship to see what happens.

And there’s a reason this question — what happens when you lose your relationship with your Primary Care Physician — is more pressing now than ever.

SABETY: In terms of the labor force, we’re about to have a huge number of baby boomer physicians retire. And so, the number of transitions, the number of relationships that are going to be lost, is going to occur at an increasing rate over the coming decade. And it’s actually even more important because older physicians tend to have older patients who are probably more vulnerable to the loss of these relationships.

Patients like Adrienne’s grandmother. But retirement isn’t the only thing that can sever longstanding patient-physician relationships.

SABETY: There’s two additional big pieces here. The first is that, actually, 30 percent of patients change insurance plans in a given year. And when you change insurance plans, often your network changes, and that changes the physicians that are covered by your insurance plan. And so, that naturally severs these relationships. And the other thing is that many states — about 36 — enforce non-compete agreements, which means that when a physician moves to a different clinic or to a nearby town, they actually are legally prohibited from taking their patients with them.

In fact, the New York Times ran a piece in 2019 called “Did Your Doctor Disappear Without a Word? A Non-Compete Clause Could be the Reason.” According to this article, a survey of nearly 2,000 primary care physicians in five states found that roughly 45 percent were bound by such clauses.

Clearly, losing a longstanding physician relationship could impact a patient. What Adrienne wanted to know is: by how much?

 SABETY: What we do here is very different from other studies in the literature in the sense that these patients have already had relationships of about three years with these doctors. The average patient in our sample sees the primary care physician about six times a year. That’s quite substantial. Something like shift changes — those are very short relationships. That’s just someone that you’ve met for that day or that week, whatever your hospital stay is. But in our study, it’s really thinking more about these longer-term relationships that have taken a lot of time to build.

And when a relationship like that is dismantled, it can reverberate across the medical field.

SABETY: When a patient loses this long-standing physician who they’ve had a relationship with, they use the emergency department more. They use the inpatient setting more. In follow-up work, I show that patients are more likely to go to a skilled nursing facility, more likely to use hospice care, more likely to die.

So, why is that? Once their doctor leaves, what happens to patients?

SABETY: When they lose that main primary care physician, they visit the other providers that they already had a relationship with a little bit more. So, they start going to specialists to get their flu vaccine, to get their preventive screens, to get their annual exam. But those are things that we typically think of as being in the primary care physician specialty. And so, as a result, these patients get slightly lower levels of primary care. These adverse events actually increase across the length of the relationship. So, a patient who had a relationship with that physician for 15 years versus a patient who had a relationship for three years is way more likely to have an adverse event.

So, how can we change the way transitions occur between Primary Care Physicians to reduce adverse outcomes for patients?

SABETY: The key here is that you need your replacement physician to start taking over much sooner. Typically, what happens in our health economy is that when a physician leaves, they just mail out letters to all their patients saying that the physician is leaving at this date, that they will be transitioned to another physician, and that is all the patient gets. And what I show is that these patients are visiting the exiting physician until the day that physician exits. And then they try to figure out what’s next. Then they try to make that new appointment with the replacement physician. Instead, what I think would be way better for patients and even the exiting physician is, as soon as that physician realizes they’re leaving, that they have that replacement physician incorporated in the patient’s care. The leaving physician and the replacement physician should see the patient together for a couple of visits. That will help that new physician incorporate some of the nuance that the leaving physician has taken years to build. And then from there, you can build the relationship.

It’s an interesting idea, in theory, at least for patients who are high risk. But could we really make this happen? Who could help make this work?

SABETY: In commercial insurers, like the Aetnas, the Blue Cross Blue Shields, you can only have one P.C.P. But if the physician files with the insurer that they’re going to be leaving in six months, the insurer can make a determination where they say, “Okay, we are going to allow you to see multiple physicians for this period.” I think it’s really on the insurer to reach out to the patient and say, “Hey, we want you to start transitioning your care. We highly suggest you do it.” There are a number of additional frictions that we’ve set up in this health system that make it even harder for these transitions to occur seamlessly. One being different practices have different electronic health record systems that don’t speak to each other. Because of the way that these transitions are structured right now, I think they’re actually quite detrimental to patients and their outcomes. And we could do a lot better.

This idea of increasing the overlap between the old and new doctors reminded me of what David Chan observed in his emergency department work — even though he was looking at some of the shortest-term doctor-patient relationships.

CHAN: The more overlap that you have between shifts the less doctors are subject to this wedge in terms of wanting to go home as opposed to providing the best care for their patients. And the easiest way to do this is just to make sure that you’re properly staffed, especially at the near the end of doctor’s shifts. I think more generally, this kind of points to how you would staff the hospital or how you would have resources to provide care for patients in a way that is not directly mechanical because the providers are not robots. They’re not working without their own concerns of going home in time for dinner.

And remember Jeffrey Rothschild? The setting of his study was a sort of middle ground between Adrienne and David’s — not the space for long-term relationships, but where a physician has enough familiarity with their patients that they have crucial information to pass on to the next doctor. As I mentioned before, Jeffrey’s study tested a tool that was designed to reduce errors by improving communication during those handoffs.

ROTHSCHILD: The tool we used builds off the acronym IPASS. The “I” stands for illness severity. The “P” is a patient summary. The “A’s” an action list. The “S” is situation awareness and contingency planning. And then finally, the last “S” is a synthesis by the receiver, which is that the person receiving the handoff confirms they received the information and they are in agreement that they understand what’s going on. So, the study was done at nine different pediatric residency programs in children’s hospitals. And what we did was a pre- and post- intervention, which means you study the usual care, so there was six months baseline data collected. Then it took several months to implement the intervention, which in this case required education, training, implementation of the handoff tool. And then we did another six months of observation using the tool.

To be successful, a tool like this has to be simple and easy to use, so when residents hand off multiple patients at the end of their shifts, it doesn’t add to their burden of time. Jeffrey’s study showed that his tool checked this box — the time it took residents to do good handoffs wasn’t any longer than it took them to do handoffs before IPASS was introduced. But that was far from the most exciting finding.

ROTHSCHILD: The predominant finding is that the intervention had a significant reduction in medical errors from the baseline to the intervention period. There was about a 23 percent relative reduction in medical errors and there was a 30 percent reduction in preventable adverse events — those errors that are harmful or potentially harmful.

Not only is IPASS still used in those nine pediatric programs, but other hospitals and medical settings have implemented it as well. It’s used for adult inpatients, it’s used between nurses and physicians, and between different locations in the hospital where other transitions occur.

ROTHSCHILD: For example, when patients are transferred, say from the, O.R. to the post-operative care unit or from the I.C.U. to the floor, this tool has been useful in those scenarios, too. Through the years, this tool has been well accepted by tens of thousands of physicians and nurses nationally.

As David said, physicians are humans, not robots. This means that transitions in care are inevitable — and those transitions inevitably have costs. But the good news is that there are research-backed things we can and should do to improve the transitions. And in some cases, we’re already doing them.

There you have it — that’s it for today’s show. I hope you enjoyed it. Coming up next week on Freakonomics, M.D: The last two years haven’t really been funny for health care providers, but there have been a few bright spots.

Will FLANARY: I have patients call my clinic asking to make an appointment with Dr. Glaucomflecken.

Some research has indicated that humor could be good for our health. That laughter might be the best medicine.

FLANARY: I would get comments from people saying, “I have no idea what’s going on in this video, but for some reason it’s making me laugh.”

But does that mean your doctor should be funny?

KELLY: If you think about the physician-patient context, it is not a physician’s job to be a comedian.

That’s next week on Freakonomics, M.D. I want to thank my guests, Jeffrey Rothschild, Adrienne Sabety, and David Chan. And thanks to all of you, for listening, writing in, and supporting the show. If you can, leave a review for Freakonomics, M.D. wherever you get your podcasts. It really helps us out. And if your friends and family aren’t listening, talk to them, be with them, help them see the light.

*          *          *

Freakonomics, M.D. is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, People I (Mostly) Admire, and Off Leash. All our shows are produced by Stitcher and Renbud Radio. You can find us on Twitter and Instagram at @drbapupod. This episode was produced by Jessica Wapner and Lyric Bowditch, and mixed by Eleanor Osborne. Our senior producer is Julie Kanfer. Our staff also includes Neal Carruth, Gabriel Roth, Greg Rippin, Rebecca Lee Douglas, Zack Lapinski, Morgan Levey, Ryan Kelley, Jasmin Klinger, Emma Tyrrell, Jacob Clemente, Alina Kulman, and Stephen Dubner. Original music composed by Luis Guerra. 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.

*          *          *

JENA: David, if you had to be an economist or a doctor, you could pick one path, what would it be?

CHAN: I mean, can I just pick your path?

JENA: Yeah, you could definitely host a podcast. It’s not that hard.

Read full Transcript


  • David Chan, professor of health policy at the Stanford School of Medicine.
  • Jeffrey Rothschild, professor of medicine at Harvard Medical School.
  • Adrienne Sabety, professor of economics at the University of Notre Dame.


Episode Video