Robert WACHTER: I came out to U.C.S.F. as an intern, I’d been a medical school student at Penn in Philadelphia, and, uh, kind of thought they made a mistake matching me. It’s one of the great programs in the country. And I was flabbergasted to get in.
That’s Dr. Bob Wachter. Today he’s professor and chair of the department of medicine at the University of California, San Francisco. Back in 1983, he was a newly minted white coat in the very same program.
WACHTER: And I remember feeling that imposter syndrome feeling that a lot of new interns and probably new people in any job feel. And my first day, I showed up and one of the third-year residents was finishing his year and handed me the beeper. And said, “Good luck, sucker!”
Things have changed a lot over the last 40 years — but this first day that Bob’s describing — where medical students graduate and, within a few days, start treating patients — that’s still happening. About 28,000 medical students become doctors each summer, most of them starting sometime between June and July.
WACHTER: You were thrown into the deep end pretty quickly. And all of a sudden, people are calling you “doctor” and expecting you to make decisions.
When patients arrive at the hospital, they expect their doctor to make decisions — the right decisions. And as we all know, sometimes, on your first day at a new job, you don’t always know what to do. That’s not a huge problem in some professions. But in a hospital, it can be the difference between life and death.
WACHTER: There’s no way to have a system where everybody’s experienced without going through a stage where they’re not. That is the July conundrum that we face every year.
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: If you’re a patient, how safe is it to go to a teaching hospital in July?
Robert HUCKMAN: Somehow, the system has to train new physicians. And anytime you have a new physician, there’s going to be some learning curve.
And how can hospitals make sure that patients receive the best possible care, whether it’s July or any other month of the year?
Vineet ARORA: The question about July is always a tricky one because somebody always has to be the first and that’s where the supervision is critical.
WACHTER: Frankly, I am more vigilant as an attending in July than I will be when I’m on in March. I always attend in July and have done that for the last 20 or 30 years, because I actually love being there with new trainees.
As an attending, Dr. Bob Wachter oversees new trainees — the first-year residents, called interns, who start each July — as well as the more seasoned second- and third-year residents. Everyone changes roles at the same time, so Bob keeps an especially watchful eye during this turnover.
WACHTER: My bias when I attend in March or April is that the team’s probably / going to be quite good. And I’m a little surprised when I have to really get down and dirty and really kind of oversee basic stuff. In July, I assume that I’m going to have to.
Paying more attention in July makes sense, right? These are brand-new doctors, making high-stakes decisions. Even I have to admit: I’d be a little nervous about being treated by a doctor on their first day! As an attending, Bob does his best to make sure mistakes are caught, ideally before they even happen.
WACHTER: You wonder and worry about people, sometimes making decisions that they’re not really prepared to make. And our job as program administrators is: can we create layers of oversight and create a culture where it’s actually lauded for them to say, “I don’t know this. I need some help.”
Even with those extra eyes, those extra layers of protection, you’d think having so many people switch positions at once would lead to some changes in patient care, at least in that first month. This is the phenomenon known as the July Effect.
WACHTER: The effect means that patients get worse care when they’re being cared for by people who are new at their jobs. And that makes some sense, that you’re dealing with people with relatively less experience.
It makes sense — and yet:
WACHTER: It’s not 100 percent clear that there is such a thing. You can find studies that support that the July Effect is real. You can find studies that find not much effect. You can find it in surgical specialties and procedural specialties and other studies that don’t show it so much. But I say, when you look at the studies — and there have been dozens and dozens of them — I think you come away with the general feeling that, if you had to put your nickel down, you would say there is a July Effect.
Bob and a colleague published a review in the Annals of Internal Medicine in 2011. They found that mortality increases and efficiency — basically, how much time and money doctors spend on care — decreases in teaching hospitals in the month of July.
So yes, there’s a July effect — but that doesn’t mean it impacts everyone equally. In 2013, I co-authored a study looking at whether the July effect is larger for really sick patients. Our thought was that maybe some of the previous studies of the July effect were mixed because they didn’t typically focus on the sickest patients in the hospital — the kinds of patients whose outcomes could be affected by even the smallest errors. So, we focused on heart attack patients and studied whether their outcomes were worse in July compared to earlier months, like May. We looked specifically at teaching hospitals, because this is where the residents work. And we used non-teaching hospitals as a control group since in those hospitals there isn’t any turnover of junior doctors in July.
Our research suggested that the July effect is actually quite a bit larger for these high-risk patients. An obvious response might then be: “Okay, if there is a July effect at teaching hospitals, I’ll just go to a non-teaching hospital in July.” But this might be a mistake. That same study showed that heart-attack patients have better outcomes in teaching hospitals most of the year – and in July, the outcomes just become equal to those in non-teaching hospitals. In other words, the July effect reduces the benefit of being treated at a teaching hospital — but it doesn’t eliminate it.
In reviewing studies of the July effect, Bob has also found that it’s challenging to pin down exactly why mortality is higher in teaching hospitals in July. One reason is that the junior doctors turnover. But there could be other reasons, too.
WACHTER: So, for example, the rate of hospital admissions goes way up over the winter months. So, the hospital is fuller, but it’s fuller mostly of people with seasonal respiratory illnesses, relatively few of whom do worse. The hospital in the summer doesn’t have flu patients in it. And so, the relative severity may actually be higher. So, there are a lot of moving pieces, and isolating the July Effect. It’s a pretty tricky thing to do.
In his work looking at the quality of studies on the July Effect, Bob concluded that the better studies suggest it exists, and I tend to agree with him.
WACHTER: You might say, “It’s a no-brainer. Of course it’s because these people are new and they’re not very experienced in their job and not very good at it.” Almost certainly there is that. There’s a learning curve by the individual, but there’s probably an institutional learning curve piece, which means that it’s sort of almost independent of how good the doctor is.
As in, new hires or transfers aren’t only learning new medical procedures – they also have to learn how their new hospital works.
Think about something as simple as ordering a test. Maybe you read the guidance provided to you on your first day — but then you order the test, and someone standing next to you says —
WACHTER: “No, that’s not what you do. You do this. Everybody knows that.” And so, this sort of “Everybody knows that” — the culture being, you know, “This is the way we do things around here” — it takes a little while of being around here before you get that. And so, there have been studies that looked at this July Effect and showed that it was as powerful in more senior people who moved sites as it was in where the issue was simply experience.
HUCKMAN: You’ve got a situation where all of the talent at a given level moves.
That’s economist Rob Huckman.
HUCKMAN: And that becomes a point of vulnerability for an organization if they don’t have the appropriate backstop measures in place to anticipate and account for that.
Rob’s a professor at Harvard Business School. His work focuses on how to improve the quality and lower the cost of healthcare. What he’s talking about — all the talent at a given level moving up — is called cohort turnover. In medicine, it amounts to about 110,000 doctors across the country changing positions all at once. A few years ago, Rob started studying the impact of the July Effect on hospital operations. He and his co-authors initially questioned if changing resident start dates might blunt the impact
HUCKMAN: We were never able to be that precise in identifying when this shock actually occurs. But what we did notice is the impact of this turnover on variables such as the length of stay in the hospital or mortality of the patient begin earlier than July or even June. It may begin as early as, say, March or April. So, the impacts of this July phenomenon are really much broader. And perhaps making it a spring phenomenon.
In their paper, Rob and his co-authors compared major teaching hospitals to minor and non-teaching hospitals. They found that in the months leading up to the July turnover, operational performance gradually decreased in major teaching hospitals. They offered several possible explanations.
HUCKMAN: When you’re doing something new for the first time – those who are supervising you may be in a position where they’re willing to let you take a little more time to do something.
For instance, in the spring, before the cohort turns over, an attending doctor might allow a first-year resident to handle the responsibilities of a second-year resident, to simulate what they’ll be doing when they move up the ladder in a few months. And you’d think this would be a worthwhile investment because it’s helping to train doctors ahead of the July transition.
HUCKMAN: Other pieces of it, however, might be inefficiency, and that’s really the tough part for us to parse out. But what we can say in our study is that the effects of this turnover, at least in our opinion, seem to be occurring before the turnover itself.
If you’re listening and starting to think wait — you’re telling me July isn’t the only time I should be a little more concerned with my care? Well, Rob’s team found something encouraging:
HUCKMAN: Our results suggest that the July phenomenon is most pronounced for length of stay in the hospital, which is our proxy for resource utilization. And are significantly less pronounced for risk–adjusted mortality, which is our measure of patient outcome.
In other words, Rob’s study found that the July effect mostly meant that patients ended up staying in the hospital for longer — but they weren’t significantly more likely to die in the hospital. Those longer lengths of stay could mean two things. It could mean that care is less efficient because doctors keep patients in the hospital for longer. But it could also explain why mortality may not have been as affected — because doctors were more careful, they kept patients in the hospital longer. Either way, the costs associated with the July effect could be large. But that’s unfortunately not a simple calculation.
HUCKMAN: Somehow the system has to train new physicians. And anytime you have a new physician, there’s going to be some learning curve. Whether you do it in July, March, or August. It doesn’t matter.
Those extra costs, which are maybe driven by longer lengths of stay, are actually one way to keep mortality from going up in July. You spend something, you get something. And if you encourage teaching hospitals to be more efficient in July, to spend less on care, that might then come at a price for patients: higher mortality. There’s also the question of how you would even begin to reduce those costs in July in the first place.
HUCKMAN: When you think about, for example, the cost of patients spending an extra day in the hospital and using a few more resources to get the same quality outcome —if we were to take steps to reduce that cost, how much would we be willing to spend to do that? And I think that’s a place where one could start to think about the relative cost of the supervision you’d need to put in place to counteract the July phenomenon, versus the direct cost of the July phenomenon itself. But the reality is that all turnover is costly. So, the question is: is this particular approach to turnover any more costly for the system as a whole than the alternate? In this case, something that would be more continuous during the year. And the key for policymakers and for managers is to figure out: what is that cost if we use these alternate approaches?
Changing the system would of course bring its own challenges. And it’s one thing for an economist like me or Rob to make these suggestions, and an entirely different thing to actually put it into practice. But that’s why we have economists.
Coming up: The July effect is actually part of a much larger question that we face in medicine, a question that goes well beyond July. How do we strike the balance between training doctors, giving them the autonomy they need to learn, and making sure patients get high-quality care — all at the same time?
ARORA: There’s a great variability of how people can approach this in a thoughtful way so that the residents have a positive experience and that patients are not harmed.
I’m Bapu Jena, and this is Freakonomics, M.D.
* * *
JENA: Do you think that there is a July effect?
ARORA: Yeah, I do think there is a July effect. I think we mitigate it as best as we can, but I do think there is definitely what medical educators would call a July effect.
That’s Dr. Vinny Arora, Dean for Medical Education at the University of Chicago’s Pritzker School of Medicine. You heard from her on an earlier episode titled “The Most ‘Unique, Excellent, and Promising’ Episode.” As dean – and as a physician – Vinny spends a LOT of time thinking about medical training, and how to make sure the university is best preparing students to begin their medical careers.
ARORA: We start this process even before July. So, in May and June, we have an expanded orientation where we offer a bootcamp and this uses simulation. And so, the idea is before you put new residency trainees out on the ward with actual patients, can you recreate some of those scenarios in a high-fidelity environment so that they can practice?
These boot camps are now common at teaching hospitals across the country — along with a few other noticeable changes.
ARORA: Some institutions might, for example, have lower capacity limits for the residents early on, meaning they are admitting fewer patients and that as they get better, they admit more patients, right? And so, everyone knows in July, you’ve got your guard up. Let’s make sure that we do the right things, but also train and, um, make sure the residents are learning what they need to learn.
Doctors and hospital administrators may not be the only ones with their guard up.
JENA: How comfortable would you be, uh, if an intern was taking care of you or a family member? Just in general and then also suppose that you had a procedure that had to be performed and the intern was going to be the one, performing it with some supervision, but it would be their first time doing it because you happen to be hospitalized in July. How comfortable would you feel in those situations?
ARORA: For things like taking a history physical exam, I’d be very comfortable with an intern doing those things because I think what people forget is that our medical students and residents actually are much more well-versed with sort of extracting information from the electronic health record and putting it together into a story and presenting the history and physical, than I think some of our attendings would be. And so, I would have no qualms about that. So, when people come to a teaching hospital and they’re like, “Well, I don’t want to have a medical student on my case.” I’m like, “You may want to rethink that. This person may have a lot of time for you. And they might be the person who actually breaks the case. I’ve had that happen.” There’s sometimes some bias towards residents and students in cases. And I think people need to understand that a lot of times they’re the ones that are doing this stuff day in and day out, and that you want to trust to do your procedures.
JENA: Yeah, I couldn’t agree more. And the other thing is I would also say is I remember when I was a senior resident and I’d go into a room with a junior resident and we’d be talking to a patient about doing a procedure, how welcoming and open they were to be allowing someone to be supervised, to learn. I mean we’re fortunate and we’re lucky that patients are often willing to allow others to learn on them, which is good. To be honest, I don’t know how I would feel if, uh, if I was getting even something as simple as an arterial line placed in my radial artery, because it’s going to take multiple sticks by a trainee, multiple points of pain. Would I do it? I hope I do it, but after the first failed stick, I’d say, “Okay, all right. Let’s move on.”
ARORA: No. And I think that’s the thing is that a lot of the hospitals to maximize patient experience, there’s a threshold, it’s like three sticks, and then you move on
So, what about things where a threshold isn’t really possible — where you have to get it right the first time, and mistakes could cost much more than a bit of unnecessary pain?
ARORA: I do think the question about July is always a tricky one because somebody always has to be the first and that’s where the supervision is critical. And so, is there somebody in the room who is the experienced person who can say, “You try once and then if you can’t get it, I’m going to try”. That’s the kind of thing I’d be looking for.
JENA: Now, let me take a sort of 30,000-foot view here and just talk about the trade-offs here. So, on one hand, clearly having an intern who has good supervision is going to be important. Um, but historically we have seen a world in medicine where there was a lot more autonomy among interns and I’m curious to get your sense of, how do we balance the need for supervision with the need for autonomy.
ARORA: when we think about this pendulum of too much supervision and too much autonomy, you want to get the right balance, right. Uh, because when residents feel too supervised and too micromanaged, if you will, a lot of our research shows they unplug and they don’t learn anything. They just cede control to the micromanaging attending. And similarly, when they feel like they have too much autonomy and the attending is too absent and they’re kind of making decisions on their own, they feel very anxious and nervous. And that’s when you worry that a mistake could arise.
VINNY: And so, I think we do need to invest to train supervisors to be effective in their roles so that they can strike the right balance between autonomy and supervision.
That balance is important — and it’s one of the things that has changed dramatically since that day Bob Wachter described, when a third-year resident handed him a beeper and said, good luck. Hospitals have also adjusted the number of hours interns work each week, from 100 down to around 80. Studies have shown that this reduction doesn’t impact competency. These are just a couple of ways things are improving — and I bet if we studied their impact, we’d see these changes are helping to decrease the overall July Effect. Which, as Vinny agrees, is the ultimate goal.
ARORA: As medical training culture advances, so do the innovations and practice. And so, we have to learn how to use those to get better. And so that’s really the holy grail, right? Is that the system is making us better.
JENA: If you were to redesign the system, how might you do it differently?
ARORA: could we consider potentially a mid-year onboarding and maybe there’s two types of onboarding, so there’s a December and a July. So, there’s a stagger. I could see that happening.
It’s a big question, one that economists and health system administrators and educators have been thinking about for a while. It might be easier in theory than in practice…
WACHTER: It’s difficult for me to imagine how that would work logistically from a regulatory standpoint.
Bob Wachter from U.C.S.F. again.
WAHCTER: You’d have to transform the way you think about medical schools, residencies fellowships, because each stage is very tightly coupled with the next stage. And they all are pretty much dependent on a group finishing on one day and starting the next gig the next day or the next week.
Harvard Business School Professor Rob Huckman’s work has offered some practical advice, specifically around the role of nurses during this precarious July period.
HUCKMAN: When you think about who those residents are interacting with when they’re providing care to patients, they’re often doing their primary interaction with nurses. So, one question would be — how could an organization leverage the skillset of those nurses to try to ease the on–ramp for new clinicians who are learning the ropes of not just clinical practice, but are also learning how to operate within a given set of institutional norms and policies, where those nurses have been there for, in many cases, many years.
Up until now most of the research on the July effect has focused on whether it exists, rather than what to do about it. Bob Wachter is ready to see that change.
WACHTER: So it would be really nice to have a toolkit from rigorously done studies that say the right number of patients that a new intern should take care of in the ambulatory environment is X. And within a month, you can dial that up to three quarters of what they will get to. And over six weeks, they can get to their full cohort. Be nice to design a study that said for an inpatient team, this is how many patients they can take care of. This is the level of intervention and oversight that the attending needs to apply and try to figure out sort of, how can we quantify that? What does that look like? Or is that just a general statement of being more careful?
At U.C.S.F., the administrators make a point of assigning the best attendings to work in July, the best medical educators really. They’re reminded to be extra vigilant, and to communicate — maybe even over-communicate — during that initial learning curve. Bob also notes that U.C.S.F. maintains the same staffing levels in July as in the busier winter months, so that new residents can be assigned fewer patients, at least at first.
Training day one doctors is a learning process — one that Bob’s happy to be a part of — both as a doctor.
WACHTER: I tell people that you can teach trainees facts in January and February. But in July, you can teach them philosophy.
And as an architect of the system.
WACHTER: We need to figure out and continuously sort of recalibrate and rethink: how do you gain the experience? How do you allow them to traverse that learning curve safely? And one of the reasons I enjoy working in the training space is I think that’s an infinitely interesting and important problem. How do we train people in a way that gives them graded responsibility? Because at some point they will be fully responsible, but they also have to do their first case. It’s really on all of us to come up with a system that ethically shepherds people through that stage.
I don’t think I could put it any better than Bob did, but I do want to end with a little reflection from my own time in training. I remember my very first day on call as an intern. I was obviously a little scared, no surprise there. But what did surprise me was that every patient I cared for that night seemed happy that I was there. They might’ve been scared but they didn’t show it. And that was true for the next night I was on call and actually ever since. It’s something I’m grateful for. And it’s a gift from patients to everyone in medicine who’s learning to do their job. So, thank you.
Alright, that was a fun episode. I want to thank Bob Wachter, Rob Huckman, and Vinny Arora for joining me and thanks to you all, for supporting the show. As always, leave us a review, if you haven’t, on Apple podcasts and wherever you get your podcasts. And tell your friends. It really helps.
And that brings me to next week on Freakonomics, M.D.
KOMAROFF: Most of the people I’ve seen say that they were perfectly fine, perfectly healthy. And then one day they came down with a virus, but this time they never got better.
One in five adults who’ve been infected with Covid in the U.S. experience lingering symptoms from the illness. Long Covid is a relatively new syndrome, but the experience of lingering symptoms after an illness isn’t new to some people.
WINSTON: I woke up one day and I felt like my legs were lead. I remember getting into the shower and just feeling slightly off balance. And I had this strange sense of fatigue come over me.
We’re going to talk about a chronic, debilitating disease that affects millions of people around the globe. There isn’t a totally reliable way to diagnose it, and there aren’t any approved treatments. So, why not?
WHITTEMORE: There was really an outcry from the community that N.I.H. was not doing enough.
But the N.I.H. is taking action when it comes to Long Covid. They’ve designated more than a billion dollars toward studying the disease. Could this research and funding also help people living with chronic fatigue syndrome, better known as M.E./C.F.S.?
That’s all coming up next week, on Freakonomics, M.D. As always, I’d love to hear your thoughts, and your questions. My email is email@example.com.
* * *
Freakonomics, M.D. is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, and People I (Mostly) Admire. 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 Danielle Elliot and mixed by Eleanor Osborne. Our senior producer is Julie Kanfer. Our staff also includes Neal Carruth, Gabriel Roth, Greg Rippin, Rebecca Lee Douglas, Morgan Levey, Zack Lapinski, Ryan Kelley, Jasmin Klinger, Emma Tyrrell, Lyric Bowditch, Jacob Clemente, Alina Kulman, and Stephen Dubner. Original music composed by Luis Guerra. To find a transcript, links to research, and a newsletter sign-up, go to Freakonomics dot com. 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.
* * *
ARORA: You’re reminding me of a time when I was a chief resident and the intern said, “Oh, I’m so excited to work with you next month.” I’m like, “Oh, that’s great. Eho’s our attending?” And then there was like this weird pause and everyone’s like, “You’re the attending.” And I’m like, “Really? Is that happening?”
- “Mortality Among High-Risk Patients With Acute Myocardial Infarction Admitted to US Teaching-Intensive Hospitals in July,” by Anupam B. Jena, Eric C. Sun, and John A. Romley (Circulation, 2013).
- “‘July Effect’: Impact of the Academic Year-End Changeover on Patient Outcomes: A Systematic Review,” by John Q. Young, Sumant R. Ranji, Robert M. Wachter, Connie M. Lee, Brian Niehaus, and Andrew D. Auerbach (Annals of Internal Medicine, 2011).
- “On-Call Supervision and Resident Autonomy: From Micromanager to Absentee Attending,” by Jeanne M. Farnan, Julie K. Johnson, David O. Meltzer, Holly J. Humphrey, and Vineet M. Arora (The American Journal of Medicine, 2009).
- “What Happens to Patients When Thousands of Cardiologists Leave Town?” by Freakonomics, M.D. (2021).