Let’s start with all the things that you should look for in a radiology program except the one that I’ve alluded to in the title. Many of these features are broadly generalizable and largely not unique to Radiology, and no one needs to tell you that they’re important.
The Usual Stuff
We don’t need to spend much time discussing the usual factors much: location, academic vs community, prestige, size, blah blah. You probably have a decent idea what kind of cities you’d be willing to live in. Other than the nebulous intersection of a program’s “reputation” and your ego, everything else will generally get summed up in your overall gestalt (“feel”): program stability, subspecialty strengths, book fund, etc. For example, no one is going to pick a program because they let you use your book fund to buy an iPad, but it may play a small role in how supportive and neat the place feels overall. Academic programs are generally perceived as better in pretty much every way for training quality, but they’re also typically bigger, giving you more potential peers/friends and generally more scheduling flexibility.
You might also say that program stability, camaraderie, staff/resident interactions, and breadth of pathology are all important, and you’d be right. But no program is going to tell you that their training sucks, that their pathology is strictly bread-and-butter, and that no one is fun to work with. You’ll get a feel for the general program environment, but it’s a deeply flawed one based on the handful of people you interact with +/- how boring the conference is (note: the one conference you see is a poor proxy for educational/didactic quality). That’s one of the baked-in inadequacies of the interview process for any specialty. Radiology is generally a pretty decent field, particularly during residency: if people actually seem anything other than happy with their residency choice, consider that a huge red flag.
A “resident-run” program is generally ideal, though that term is used loosely enough that you’ll likely have to ask what that actually means if someone says it. Resident-run usually means some of combination of the following: the residents (generally the chiefs) make the annual schedule, the call schedule, handle vacation requests, and are involved in the bureaucracy of the residency and the department such that the residents have a strong say in the composition of your training and any changes that may be made to it. Resident-run programs tend to have more scheduling flexibility and a more democratic milieu. I’d argue that you’re generally unlikely to be unpleasantly surprised by program changes in a resident-run program.
Is absolutely life-changing and probably more important than you realize. Talk to the residents.
Are generally not that competitive. Most academic places will send people to “great” fellowships. And, a lot of residents will stay on at strong programs anyway to continue their lives/training at home; this speaks to the comfort and environment of the program, which I believe should be viewed as a good thing overall. If no one is staying, that means everyone is fleeing. If you’re looking at community programs, then you definitely want to see what connections they have. Everyone does a fellowship.
Note that there are academic fellowship factories out there that have a disproportionately high number of fellows. In these programs, fellows tend to take more call (often instead of or supplanting the residents). A place with a ton of fellows is often a nice name, but there may be a concomitant hit in residency training experience. Though, if you’re lazy, needing to fight fellows for cases could speak to your interests.1 As for doing a fellowship at one, jump that hurdle when you get there to see if the name is worth the likely suffering.
Since people often ask: board passage rates at big places should also be near 100%, which functions as a proxy for how a program helps its residents prepare. The pass rate for the Core exam is in the high 80s, so good programs shouldn’t have failures routinely. If the residents are taking a bunch of overnight call in May or are grinding out full days on IR the week before their June Core exam or haven’t received a bunch of dedicated/targeted Core-relevant lectures, then the program isn’t supporting its residents.
IT & Equipment
I find it embarrassing that there are still places with antiquated or clunky PACS and incomplete/terrible EMRs. It’s the 21st century. Everything should work well and be as frictionless as possible. We literally use a computer all day long, so any irritating issues will be compounded over the course of your countless hours at the workstation. Something as simple as how the PACS handles scrolling or making measurements is important for us.
Any academic place should hopefully have or is getting new-ish dual-energy or spectral CT and 3T MRI in addition to usual gazillion slice CTs and 1.5T MRI. You can actually do different things on different machines. It’s not that you can’t learn this stuff later, you absolutely can. It’s that it speaks to the buy-in of the institution to provide strong imaging services and the strength of the department to have new toys.
Pathology & Volume
No one has a monopoly on sick people (or scanners). That said, high clinical volume is a plus. If a single resident can cover an entire service without any independently reading faculty, that’s not a high-volume place. Increased volume is good because it allows attendings to share interesting cases with you in addition to the cases you read on your own. It also means that you probably won’t have a responsibility to “clean the list” during the day (because you literally cannot), which tends to mean you have the opportunity to read at your own pace, look things up, etc.
If your academic center is the only one in a large city/metropolitan area, you’re going to be drawing from a lot more people and seeing a lot more variety than if your hospital is one of five smaller academic centers within a 5-mile area. You’ll still get trained for the high yield stuff regardless; the question will be how much weird stuff did you see on call vs. only in a book. Being a big “referral” center can help account for a relative paucity of volume, but typically when there are multiple “referral centers” in one place, they tend to have their own turfs and aren’t the refer-ee for everything.
Additionally, MRI should be a big part of your training. Practically speaking, you probably won’t be doing cardiac MRI out in practice without a fellowship, but that doesn’t mean your program shouldn’t teach you how to do it anyway. Your foundation in MRI physics should be rock solid. You should be exposed to everything a radiologist can do. And not just in your senior year or some BS; your radiology skills should develop as a longitudinal experience with routine reinforcement. The semi-classic way of doing plain films/fluoro +/- ultrasound for a year, then CT, then MRI is antiquated, silly, and completely out of touch. Ditto goes for taking “plain film call” before taking “cross-sectional” call. Ew.
Prestige & Research
Research strength may be important for the passionate few, but know that the name of your potential residency and its prestige factor are most relevant if you are seriously considering a career in academia. Also, note that your ability to get research done and further your personal growth as a physician-scientist does not necessarily correlate perfectly with how research-y your institution may seem on paper. The amount of research getting done at an institution generally has more to do with PhDs doing things you don’t care about than the opportunities afforded to you as a resident to do work that you are passionate about.
In some cases, for actually doing science, it’s better to have underutilized resources in the department and receive unbridled joy at your aspirations than to be in a department operating at full thrust with no one available to mentor you. On the other hand, doing strong research and writing impactful publications is never a guarantee; getting your card punched at a big name is (once you’ve matched).
Most academic programs will be happy to support you and your projects unless they’re expensive and you can’t get a grant. A dedicated research track and plenty of conference/travel support is a definite plus but is likely taken into account in your gut reaction.
Additionally, when considering different programs, the aspiring scientist needs to consider the call density and seriousness of the daytime work. While a heavy call burden and having lots of independent work responsibilities will undoubtedly lead you to be a stronger clinician, it may cut into your time for research if that’s what you’re really passionate about. If it’s easy to routinely disappear from clinical service for “research meetings,” you may not learn as much, but you can certainly get a lot more research done. I would argue that residency is your time to learn diagnostic radiology and that you should push yourself clinically as much as possible, but if you really don’t care about clinical medicine, then make sure to take this dynamic into consideration.2
How does call work here?
Radiology Call Patterns
There are several ways to do “call” (i.e. cover the ED/hospital at night).
Like the day work: There are dedicated night radiologists or even a whole “emergency radiology” division who sit in the room with you. You probably send the radiographs and ultrasounds straight to them to be over-read immediately. You likely read out CT (and MRI if performed after hours) with them 1-on-1 like a daytime rotation and then submit the report after discussing the case. Your prelim report, if anyone can even seen it, will reflect an attending’s vision. As you get more advanced or they learn to trust you, they may let you sign your stuff prior to discussion, but anything weird is still gonna get their eyeballs early on. Sure, you’re there to see more fat stranding for your education, but your real value is mostly to answer the phone and deal with the technologists.
Pseudo-independent call: You don’t talk about most cases with anyone, but they’re still totally there, if not in the room then on the list from somewhere. You read the cases and submit prelims independently, but you’ll be overread in short order. If you have questions or a tough case, the expectation is likely that you’ll call them. After all, they’re going to change it before you can do much damage anyway. If you make a strange call or the patient seems incongruously sick/healthy, the ED will wait for the final read to do anything. They may even call over to get it instead of calling you. In some cases, you read alone but reports aren’t released to clinicians until final signed by an attending. In others, residents are overread by fellows who are in turn overread by attendings. In short, you only kinda matter.
Independent call: The residents are responsible for imaging overnight. Generally, this means they cover the ER and at least the stat inpatient studies or all cross-sectional studies. You read everything that needs reading. There are no staff in-house for a significant chunk of the night. As in, a study read at 10 pm probably won’t be over-read until 8 am. While obviously staff are always available by phone, the expectation is that the residents practice radiology. Cultural milieu at the institution here is also important. If it’s independent call but there is no buy-in and the ED demands that you page your attending for a “staff read” for anything remotely unusual or complicated, then you’ll still feel pretty tethered.
Night float is increasingly the most common way to handle call. Long-short-off, overnights after a normal workday, random 24-hour shifts, or other styles are frankly problematic in radiology. Willpower and attention are limited resources that are fragile in the face of sleep deprivation.
Shift length is also important. 14-hour shifts from 5pm-7am are a lot more painful than 10-hour shifts from 9pm-7am, but obviously, there’s no free lunch. If evenings are covered separately, you’ll be working a bunch of evening shifts at some point unless you have late staff coverage instead. Having worked pretty much every combination of shift over my training, including solo 24-hour calls, I think splitting evenings and deep nights and having a relatively short night float shift is preferable for both education, sleep, and—most importantly—patient care.
Real-time readout: The attending is either periodically in the room or calls you. You read out some fraction of your cases (e.g. the CTs) in real-time or shortly thereafter. If they give you a longer leash or decide to take a
nap walk, they might come back in, overread you, and then only talk to you about things they disagree with or thought were interesting.3
The morning after readout: You sit down with an attending in the morning when you’re exhausted and go over your cases. The good? You get a detail-rich readout and “learn a lot.” The real? Watching someone scroll through images after you’ve been up all night is a form of medieval torture. You may not be able to learn effectively because you’re too tired to care.
Asynchronous feedback: Staff grade you and give you written feedback through some software package. You look over your overreads in PACS or the EMR to see what the final reports say, probably in the afternoon when you wake up or when you get back to work. You get to see what you missed on your own time as well as how the staff massage the phraseology, and then fold the stuff you like into your practice patterns.
I think virtual feedback is a better, more humane system. Especially as you increase in seniority, your desire and patience to watch someone scroll will inevitably and rapidly wane.
Preliminary vs Full Reads
A common practice at institutions with pseudo-independent and independent call is for the resident to but in a brief “prelim” read instead of a full dictated report. So the negative appendicitis CT might be “Normal appendix. No findings to account for RLQ pain” and the positive appendicitis CT might be “Acute uncomplicated appendicitis.” Drop the mic. Move on. This makes call much, much easier. It’s also less realistic and sets you up for laziness/satisfaction of search.4 The classic resident prelim is a way for you to handle volume without drowning, particularly given the complexity of cross-sectional imaging, but it partially shields you from your biggest growing experience.
Full reports—while undoubtedly more painful at 3am—will help make you a better radiologist. Saying “no fracture” is way easier on a whole-spine CT for a 75-year-old after an MVC than describing multilevel degenerative disease and grading neural foraminal stenoses, but the more you do of everything the better you get. Speed and quality are intrinsically inversely correlated, so it takes an overall significant increase in skill to increase both at the same time. You’re not braving the full volume of cases you read if you’re not reporting on the findings. Dictating the real volume will allow you to get faster, more thorough, more confident, and more capable of handling a big list—skills you’ll need to develop at some point.
Which brings us fully to independent call, which is, I believe, the single most important differentiating factor for a diagnostic radiology resident’s actual training. Unfortunately, independent call has become increasingly uncommon due to demands for 24-hour staff coverage from the ED and a general nationwide push for greater resident supervision. I find this trend problematic in radiology.
Radiology is by its nature a consultative role, and the general diagnostic work doesn’t involve repeated follow-up or clinical assessments. You look at the pictures, dictate what you think they mean, and then send the report out into the ether, maybe accompanied by a phone call when there’s blood or air somewhere they’re not supposed to be. That report is in the EMR, so if you’re wrong, it gets copied and pasted into a whole bunch of notes and you look stupid. Unlike the OR, where the surgeon is composing a narrative generally removed from scrutiny, the images don’t lie. Everyone can see them. And yes, you can also hurt people. It’s not hard to see why a department like the ED—where the attendings are there 24/7 and turnaround times are the most important metric—would want the same for the imaging that makes up such a huge part of what they do.
But think about clinical residencies: the more experience you have, the more autonomy you get. Individual experiences vary wildly, but these scenarios are common: A third-year EM resident sees patients semi-autonomously, orders labs and imaging tests, and makes clinical determinations, usually checking in with their attending verbally at some point, often at the end before discharge. The attending may eyeball the patient.5 A medicine resident on call overnight admits a patient and does a bunch of stuff, and frequently the attending doesn’t see them until rounds the next morning.6 Senior surgery residents operate and teach junior residents often without an attending scrubbed in. OB residents perform deliveries with cursory staff involvement. Hell, residents in other services are often interpreting images on their own overnight while awaiting (or in place of) radiology reads!
While there are an increasing number of programs with very hands-on tight supervision, I think it’s obvious which scenarios allow for more resident growth and decision making. The bottom line is that being autonomous for a few hours as a resident may seem scary, but graduation is coming and then the supervision is gone. Turfing out any semblance of autonomy until after residency isn’t going to help anyone, least of all the patients. Graded responsibility is a good thing.
Depending on where you do your internship, you may find yourself with a surprising amount of autonomy (that you’ll likely grow into). Then you slide back and start from scratch in radiology. Not only are you new and newly incompetent, but outside of procedural/fluoro months, you probably have no responsibility. You theoretically need to learn everything but have few external pressures to force you to improve (except for the fear of looking stupid or the fear of call). It’s a difficult transition for a lot of residents. Then as you advance, your friends and colleagues from medical school start to finish residency and start their fellowships or enter practice. And this is where the lack of independence on call would begin to get really galling. Your PGY4 equivalent former friend/new attending in the ER can defensively order a boatload of inappropriate CTA chests for PE that are going to be negative, but you still can’t speak your mind autonomously and have anyone listen. How infantilizing.
In radiology—and perhaps especially in radiology—I believe independent call is important clinically and completely changes the tone of your training. Radiology residents who are never alone or functioning independently might go through their entire training potentially never having truly made a clinically-meaningful decision themselves or directly impacted patient care. They might never have had to “make the call.” And then they’re going to graduate and then finally hit sign on a report that someone will take seriously for the first time as a fellow or attending? It’s completely different to consider making a challenging diagnosis than it is to put it out there. Better to start when you have someone who can overread you in a reasonable timeframe than to never have done it until you’re really alone. I’ll never forget my first solo 24-hour call as a beginning PGY3. It changed the trajectory of my studies and pushed me more in one day than I’d been pushed in the entire year that proceeded it. I can’t imagine going back and deferring that to my PGY6 or PGY7 year.
Psychologically, I doubt most people can ignore the reality of their training environment. If you’re barely on call and then when you are you just check the “negative” box or type in a one-sentence prelim prior to a fellow who immediately overreads you followed by an attending who overreads them (yes, this happens), then the pressure on you to do a good job, learn, and help patients is undoubtedly weaker than for someone who thinks the work they do matters.
Let’s take a step back
So, that’s all a bit dramatic. I’m overstating this a bit for effect. This isn’t to say that you can’t be a good radiologist without independent call. Of course you can, and many do. And people at the many programs without independent call are I’m sure very happy with their training and likely for good reason. Because no one is able to do residency for the first time twice, it’s impossible to know how much it matters on an individual basis or how an individual would respond to different training environments. It helps to know what kind of person you are. One of independent call’s benefits is that it forces you to learn and grow throughout your training instead of displacing an outsize portion of that growth into a rough transition to autonomy as a fellow or attending. But if you’re self-motivated, you’ll be fine. If you can emotionally treat a 1-hour overread turnaround like a 14-hour turnaround, then you’ll be in good shape. A strong call experience challenges you to be faster and more confident, particularly if you’re doing full reports, allowing you to hit the ground running out in practice. Again, after some period, it all evens out. How helpful (or not) experiences like that would also depend on your learning style and how you handle pressure. It may be that deferring that sort of anxiety until you have years of experience suits you better, or that all you really want to do is academic nuclear medicine and that a bunch of brutal general call will be a largely irrelevant skill.
I didn’t know how important the call experience would be to me at the time when I was interviewing for residency. I can definitely say now that I’m glad I had that head start, and I’m especially glad it was there to help minimize the unavoidable sensation of feeling like a student transcriptionist that occurs when first joining the field. To me, call is the core of training.
The trend across medicine of demeaning resident skills and requiring granular scrutiny is reactionary and ultimately self-defeating. People grow when they’re permitted to grow. If new residents are lacking in certain skills, then we need to reevaluate the medical school curricula. And if medical school is too full to put more clinical training in, then we need to look at the BS premed requirements. But lengthening training or turning a young attending into the equivalent of a better-paid senior resident is not a desirable or viable option for many reasons, not least because the midlevels successfully advocating for greater autonomy have managed to nearly sidestep this debate entirely. But that’s something for a different post.