I’ve been fielding job questions via email over the past months and after my talks at RSNA, ACR, and ASNR, so it seemed like the easy solution was to collect some responses here for dissemination (instead of trying to turn them into a bunch of separate posts).
As always, no one has to agree with me.
Private Practice
Is private practice still viable?
and
If private practice is so great, then why are so many groups struggling and failing?
See my answer here.
Teleradiology & Locums
So many people seem to want to do teleradiology, is the stigma/taboo gone?
The taboo is absolutely gone in the sense that you won’t be shunned, but I think many groups still find tele to be a necessary evil brought on by market demands. Off-hours swing/overnight work aside—where I think there is absolutely a growing realization that there is a very important and meaningful unique value in providing that service just like there is for driving around and doing fluoro—I think people generally find it frustrating and would be happier if they didn’t have to resort to hiring for it.
To be clear, everyone loves a hybrid work schedule, and almost everyone wants to work from home sometimes. But hiring 100% remote rads to absorb remote work often means a greater fraction of on-site work for local rads and an altogether different kind of easily comparable job.
This often results in functional dual-class structures in order to create fairness between on-site and remote rads etc. That might be no remote partners. That might be higher RVU requirements. That might be less pay. It varies.
I suspect that, over the long term, cultural dilution may be a real problem, and I’m not sure it isn’t intrinsically destabilizing for the field. I suspect we will see multiple larger tele practices opening up and grabbing business. Things are so hot that they can get great rates from desperate hospitals. I don’t know how sustainable any of this is.
Is the teleradiology trend temporary?
Overall, I think the teleradiology trend is real and not going anywhere. And at least for those willing to do off-hours work, like evening swing shifts and overnight, the demand for those services will probably never go away so long as humans are involved in interpreting diagnostic imaging. I think for the foreseeable future, there is really no reason why anybody needs to be on-site for most of those jobs.
For better or worse, there is a growing contingent of people, including fresh graduates, who want to have that same lifestyle experience during the daytime. Part of that cohort are those who are unhappy in their current positions—fairly or unfairly—and looking for greener pastures but don’t want to uproot their families and move. I do think some of that trend may go away if we do see significant efficiency gains with the coming radiology products over the next few years. Some of the AI stuff is hype or won’t change productivity, but some of it is absolutely real. You don’t need to replace a radiologist to completely change supply and demand in the workforce.
If an organization becomes fully staffed or—hard to imagine—overstaffed, the first person to go is the remote, employed or (especially) contractor, daytime-only-no-weekend teleradiologist.
I think one question becomes, if that happens, how much of a chilling factor does that lead to for the tele market and does it create an opportunity to create meaningful on-the-ground jobs again in radiology? Right now, nobody wants to do the work of practice building, procedures, and other personal touches. I think the COVID era broke a lot of residents’ brains and has made many completely discount how important meaning and connection are to long-term satisfaction. I think those residents have taken the physical component of the job for granted and don’t realize how isolated they will be working from home all the time.
Yes, work-life balance etc is better as a remote worker. The lack of a commute is a huge deal. I absolutely do like working from home for a fraction of my job. And I absolutely think that in 2025, even many happy local rads are essentially demanding a hybrid schedule. But there is a difference between regularly or frequently working from home and always working at home.
I think at least some young radiologists are chasing a good job but perhaps not building a meaningful career. I’m not sure that is a great long-term plan for life satisfaction for a 30+ year career for everyone making that choice, even if it does unquestionably make your life easier day-to-day. At the same time, we shouldn’t paint with too broad a brush either; there are obviously individuals who are much, much happier working 100% remote for a wide variety of personal reasons, and certainly not all 100% remote jobs are the same. If nothing else, simplified school logistics alone are huge.
Anyway: different strokes, and all that.
It seems like locums folks are making great money, is that what I should be doing?
Many hospitals are fine paying high rates temporarily out of need but are terrified of agreeing to a healthy long-term. Part of this is organizational dysfunction in the sense that ultimately the managers’ skin in the game is limited: overpaying out of necessity isn’t their fault, but signing a costly contract makes them look bad (and they worry about ripples across the org). This is compounded by, I think, a willful ignorance of the state of the market (which more are coming around to now). Needless to say, I think this particular strain of management is shortsighted.
No doubt, you can make good money as a mercenary right now and for the foreseeable short-term future. The question is, when things shift, can you find yourself good employment or are you left out to dry with slimmer, less-desirable pickings? Nailing the landing could be hard or it could be a total non-issue. But the nature of locums is that you can usually make more money—especially if you’re geographically or temporally flexible—because you’re trading safety and stability in order to fill those urgent, well-paying coverage gaps. Nothing wrong with that as long as you acknowledge the tradeoff.
Training & Mobility
Do I Need to Do a Fellowship?
To get a job? Absolutely not. There are jobs right now for fresh trainees without a fellowship.
To get a specific job? In many cases, yes.
There are plenty of places—both groups and hospitals—that still want fellowship training for radiologists, whether that’s because the paper has actual value, because the marketing of said paper has value, or because they want specific skills that are hard to get on the job without some pain or investment and are easier to build on with that additional foundation.
It’s not unreasonable for residents to ask themselves the fellowship question—especially since, while many radiologists don’t necessarily do a lot of deliberate learning on the job outside of gradual improvement with experience, we do know that we can learn new things if we want to, even outside of the confines of a residency training program.
If you want to forgo a fellowship and there’s a place you want to live—especially if you’re willing to be in smaller or more rural areas—it makes sense to ask those groups.
It makes sense to go look at job listings (on Independent Radiology, for example) and see if there are options in the right spot.
If it’s not clear from the internet, then start the real job search early, put feelers out, and find out if people are willing to give you a job that you actually want without additional training.
The follow-up question is always: “But what happens if I don’t like that first job?”
And to that, I have no idea.
There are certain situations where fellowship is permanently important. For example, there are some hospitals and places that truly do want to have CAQ-holding neuroradiologists on staff, and that is something that is challenging to achieve without playing the game.
But in many other situations, multiple years of experience should matter more than how you spent a single year a decade in the past.
So the reality is: if you stay in your first job for five years and then move to another one, is anyone going to care that you didn’t do a body fellowship?
I would think not—but I couldn’t say for sure. Probably depends on how subspecialized you’ve been in practice and what your skillset actually is. Fellowship is certainly a terrible proxy for skill, but it is still widely used.
I think if you do take a job and you hate it and you want to quit within a year—that might be more awkward. But as long as the radiology shortage persists, I think you’ll have some flexibility on that front.
What’s the downside if I hate my job and leave it?
Well, for one, it’s sad—and it means that you probably didn’t enjoy yourself while you were working there.
Logistically, it means you’ll probably have to deal with a non-compete, which could mean either working in a different geographical area—possibly needing to move—or entering the remote teleradiology workforce.
Now, you may not mind working 100% remotely, but for those who were hoping to have support, connection, and community (particularly in their early career), that may be less fun.
If you’re interested in becoming a partner in a private practice, it means you’re going to have to restart that path when you get your next job—which again may not even be possible if you’re working as a teleradiologist. Given the relatively short workups these days, this is less of an issue today than even just a few years ago.
There may be some mostly minor benefits issues: delayed 401(k) access, potentially lost 401(k) matching, giving back signing bonuses, paying for tail insurance, and other such things.
In reality, all of the financial considerations can be handled, mitigated, and dealt with—and none are the end of the world.
If you are a remote employee leaving your current gig to be a remote employee somewhere else, the downside obviously is mostly just that breed of transient hassle, maybe learning a new system, etc. There’s no reason for us to pretend that in our increasingly mobile workforce that swapping tele gigs necessarily carries a huge cost. The musical chairs that teleradiology enables is one of the reasons the market is so crazy right now.
Mammo
What on earth is going on in breast imaging??
High demand and low supply, coupled with typically high reimbursement and productivity of breast imaging.
An increasing share of trainees have been drawn to breast by the market. Of course, many are intrinsically passionate or want a patient-facing job without the IR lifestyle and stress, but we should also just acknowledge that some are chasing the perceived chill of well-paying jobs with no evenings, weekends, or call. For this latter group, I am somewhat more concerned, if only because picking anything based on the current state of idiosyncratic and unpredictable market forces seems like a pretty poor decision metric.
(But, as long as they won’t be disappointed when things change, obviously it can be a very fulfilling career even if it wasn’t some deep-seated passion. I think passion is highly overrated. Breast imaging isn’t going anywhere, and the personalized patient counseling component and procedures remain critical healthcare tasks that are obviously deeply meaningful to both doctors and patients and not as prone to outsourcing or AI disruption.)
We currently live in a world where “telemammo” is definitely a thing, of course in part to remote diagnostic coverage at far flung centers but also capitalizing on screener volume and specifically tomo to allow for things like really well-paying 4-day 100% remote breast jobs without call. Some of that is often remote diagnostics with actually talking to patients variable/optional and all procedures driven to either centralized on-site staffed locations or to specific days when coverage is present. (I haven’t done any breast imaging since training, but I will admit I honestly don’t get the idea of breast imaging with essentially zero direct patient contact.)
As a semi-outside observer of women’s imaging, I think the opportunity for disruption there is extremely high, and the odds of those jobs staying the way they are–often “higher” pay and “better” lifestyle than other rads–is perhaps the least likely status quo to remain completely unchanged. For one, we aren’t done seeing screening reimbursement cuts, especially for tomosynthesis, which I’m told is currently the biggest financial driver for enabling all these shenanigans.
To top it off, the corporate footprint in breast imaging is large (Solis, SimonMed, RP, etc) and the economic case for AI for these companies is huge. If you ask around about AI in radiology, it’s AI for breast screening that really gets these folks salivating. And that’s because AI-enabled improvements to productivity and efficiency for screening mammograms could be more easily economically profitable compared to essentially any other part of diagnostic radiology.
Give me an AI tool that provides a frequently signable report like an upper-level resident, and that will make me more efficient, yes, especially for certain things such as radiographs, but the reality is that me reading twice as many radiographs doesn’t yield massive RVU gains or windfall profits. Thanks to the BIRADs reporting structure, breast imaging is already the most standardized written report of anything in all of radiology. If AI-generated reports were ready to click and go for the majority of screeners with a well-calibrated callback rate—and there are no breast-specific legislative rules preventing their use as an autopilot, which of course there could be—then I think that would move the needle. I don’t do any breast imaging and only vaguely know the volumes in play with the current shortage, but I would suspect that it wouldn’t take a massive efficiency gain (20%?) to fundamentally change the supply and demand that has led to these lifestyle positions that have essentially derailed recruitment and retention throughout radiology, especially when it comes to young rads.
I am not exactly sure why no one is talking about this, but I’ll admit I was hesitant to include this part in the Q&A. Because to be clear, I’m not saying breast is “easy” or that breast imagers are going to be replaced first or any nonsense like that. Obviously nothing on the market right now is anything other than an improved CAD, so nothing is happening currently. I certainly don’t put stock in my predictive capabilities, but I do think this recent post-tomo status quo with breast over the past few years is the least sustainable part of the radiology workflow, if only because with the money at stake and the narrow task, the opportunity for disruption is too high (but yes, on an unknowably multiyear timescale). Either a change to reimbursement or productivity or both could easily substantially adjust the sweetheart offers of these employed and private-equity positions that are driving the market.
I don’t know when, and I don’t know if breast will actually be affected more than other subspecialties with AI, because it’s possible that radiology capabilities of the coming vision-language models will be satisfactorily broad across the whole field. I do know companies are looking at mammo as the exciting thing they’re willing to actually pay money for in AI. Anyone can see that the current tools are overhyped and brittle, but it’s also a mistake to assume that everything is and will remain vaporware, especially for narrow diagnostic tasks.
Of course, how things get deployed and what guardrails are put in place are total unknowns, and the timeline and scale of reimbursement changes are a constant battle.
But to answer the question: in addition to body/general radiology skills, breast is the most in-demand subspeciality. For now.