Random Salary Question

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.
I hate to agree with this, but this is 100% what EM has become.

The old guard here on SDN will fight against this sentiment tooth and nail, but ultimately this is correct.

It's a dead and/or peri-arrest "specialty." To give you an idea, EM isn't even a thing in many other countries/medical systems. Our 50-60 (?) year stint was an interesting one, and effectively born out of a gap in a tragically pathetic healthcare system.

To all the young ER docs reading this, or EM residents, or even (delusional) medical students thinking about EM:

GET UP AND GET OUT, The Ship Is Sinking

This is where taking the entrepreneurial approach might have some upside, particularly as we look down the pipeline of a new administration gutting regulation.

Maybe this will be a renaissance for physicians that break free from insurance, ABMS, and medicare/medicaid.
Honestly when it's 3 am and I'm working solo in a place with no other doctors and there is a radio call for "2 month old born premature in respiratory distress" I don't always want to be the combination triage expert and resus guy anymore.

But if not me, who? When my family is sick, I want BCEM docs prepared to do what's needed, but I don't know that I want it to be me.
 
I don't, I probably wouldn't last an hour, but undertrained APPs sure think they can. AI has also been in "training" in radiology for 20 years and the field is only getting better and advancing
We need everyone in medicine. But as I have said in the past, DR better do something fast or AI is going to make you all go the way of CT surgery with interventional cardiology. It will be a mighty fall.

Cheap and easy is what insurance companies and hospital admins want; patient outcomes and standard of care can be made up over time.
 
I hate to agree with this, but this is 100% what EM has become.

The old guard here on SDN will fight against this sentiment tooth and nail, but ultimately this is correct.

It's a dead and/or peri-arrest "specialty." To give you an idea, EM isn't even a thing in many other countries/medical systems. Our 50-60 (?) year stint was an interesting one, and effectively born out of a gap in a tragically pathetic healthcare system.

To all the young ER docs reading this, or EM residents, or even (delusional) medical students thinking about EM:

GET UP AND GET OUT, The Ship Is Sinking

This is where taking the entrepreneurial approach might have some upside, particularly as we look down the pipeline of a new administration gutting regulation.

Maybe this will be a renaissance for physicians that break free from insurance, ABMS, and medicare/medicaid.

This brain rot field is so dead. Honestly I don't see any solution aside from keep shoveling money into the S&P 500 and keeping spending to a reasonable limit in order to retire early. Every "entrepreneurial" activity I've explored seems to involve shady dealings where you sell your soul even more.
 
We need everyone in medicine. But as I have said in the past, DR better do something fast or AI is going to make you all go the way of CT surgery with interventional cardiology. It will be a mighty fall.

Cheap and easy is what insurance companies and hospital admins want; patient outcomes and standard of care can be made up over time.
People have been saying this for years. Reality; avg age of a radiologist is over 55 w avg retirement age of 64, training spots have increased 8-10% since the early 2000s with image volumes continuing to rise. Rvu per scan gets cut yearly, but the technical fee is high, and $/rvu keeps increasing. If AI were to make it to where 1 rad can do the job of 5, there would still be a shortage. AI in rads is also very different (and much more advanced, years ahead. You think this voice recognition is cool? Rads has been using it since the early 2000s) than what the public is exposed to leading to this AI craze, which is why when you talk to a radiologist who is in the know they aren't concerned (even if you disagree)
 
People have been saying this for years. Reality; avg age of a radiologist is over 55 w avg retirement age of 64, training spots have increased 8-10% since the early 2000s with image volumes continuing to rise. Rvu per scan gets cut yearly, but the technical fee is high, and $/rvu keeps increasing. If AI were to make it to where 1 rad can do the job of 5, there would still be a shortage. AI in rads is also very different (and much more advanced, years ahead. You think this voice recognition is cool? Rads has been using it since the early 2000s) than what the public is exposed to leading to this AI craze, which is why when you talk to a radiologist who is in the know they aren't concerned (even if you disagree)
Personally, I don’t think AI will make the job 1 rad = 5 current rads. I worry it will be AI + 1 rad = 100 current rads. But that is just my worry because that is what HCA/United Healthcare are drooling over.

HCA is already trying to remove DR from the stroke CTs and large PE CTAs decision trees by using their AI.
 
People have been saying this for years. Reality; avg age of a radiologist is over 55 w avg retirement age of 64, training spots have increased 8-10% since the early 2000s with image volumes continuing to rise. Rvu per scan gets cut yearly, but the technical fee is high, and $/rvu keeps increasing. If AI were to make it to where 1 rad can do the job of 5, there would still be a shortage. AI in rads is also very different (and much more advanced, years ahead. You think this voice recognition is cool? Rads has been using it since the early 2000s) than what the public is exposed to leading to this AI craze, which is why when you talk to a radiologist who is in the know they aren't concerned (even if you disagree)
I've been using voice recognition software in my notes since you were probably in middle school and I've thought it's been cool the whole time. I also still have to secure chat the radiologists occasionally when they make a transcription error that entirely changes the interpretation.

But I'm glad you're hopeful on AI not tanking the specialty. I think DR is going to go the way of pathologists manually doing their own lab tests before analyzers hit the market that made the human eye obsolete except for the rare case.
 
Personally, I don’t think AI will make the job 1 rad = 5 current rads. I worry it will be AI + 1 rad = 100 current rads. But that is just my worry because that is what HCA/United Healthcare are drooling over.

HCA is already trying to remove DR from the stroke CTs and large PE CTAs decision trees by using their AI.

Underappreciated take. As long as dollars drive healthcare, and AI is cheaper than humans, you bet every damn dollar you own for-profit corporations will push boundary after boundary on this until forced to stop by congress (which they also own).

OTOH, be a wise investor. I mostly invest in space but ai is a great market as well. Invest well and get out.
 
Underappreciated take. As long as dollars drive healthcare, and AI is cheaper than humans, you bet every damn dollar you own for-profit corporations will push boundary after boundary on this until forced to stop by congress (which they also own).

OTOH, be a wise investor. I mostly invest in space but ai is a great market as well. Invest well and get out.
Moreso under-appreciated in the U.S., where it always seems as though there's infinite money and everyone can grab a pound of flesh as the stricken man rolls by.

Outside the U.S., the augmentation/replacement of clinicians is a desperate race against time and funding constraints to try and provide sufficient level of modern care to an aging population, whilst not blowing out the government health budget.
 
Personally, I don’t think AI will make the job 1 rad = 5 current rads. I worry it will be AI + 1 rad = 100 current rads. But that is just my worry because that is what HCA/United Healthcare are drooling over.

HCA is already trying to remove DR from the stroke CTs and large PE CTAs decision trees by using their AI.
If radiologists are still legally liable for the images they read, that doesn't really increase efficiency even in the setting of a decent AI. Checking images is the bulk of our work, we already have templates and voice to text that maximize efficiency for writing the report. We have pretty good versions of AI overreaders already, they're called senior residents and fellows, but attending radiologists still check every study. The only way AI becomes a threat is if a large subset of scans, or scans as a whole, are taken away from human eyes completely and the hospital/AI company takes on the liability. Too many complex studies, "edge" cases, rare pathology that doesn't have thousands of examples to train AI on, and limited/artifact-riddled studies for that to happen anytime soon.
 
If radiologists are still legally liable for the images they read, that doesn't really increase efficiency even in the setting of a decent AI. Checking images is the bulk of our work, we already have templates and voice to text that maximize efficiency for writing the report. We have pretty good versions of AI overreaders already, they're called senior residents and fellows, but attending radiologists still check every study. The only way AI becomes a threat is if a large subset of scans, or scans as a whole, are taken away from human eyes completely and the hospital/AI company takes on the liability. Too many complex studies, "edge" cases, rare pathology that doesn't have thousands of examples to train AI on, and limited/artifact-riddled studies for that to happen anytime soon.
Unlike some in this thread I don't really have a strong opinion on the subject but at least where I trained (two different institutions) it was almost exclusively residents/fellows doing reads overnight.

If AI could get to the point where it would do the entirety of all night reads / prelim reads for a system and just get double checked by Rads in the AM would that not significantly affect the Rads market in some way?
 
Unlike some in this thread I don't really have a strong opinion on the subject but at least where I trained (two different institutions) it was almost exclusively residents/fellows doing reads overnight.

If AI could get to the point where it would do the entirety of all night reads / prelim reads for a system and just get double checked by Rads in the AM would that not significantly affect the Rads market in some way?
No because you're not cutting down the total number of scans to be read. Those scans done by the residents/fellows were prelimed for the attending to finalize in the morning anyway, it didn't cut down on anyone's work. It's just a better schedule since they don't have to come in overnight.
 
No because you're not cutting down the total number of scans to be read. Those scans done by the residents/fellows were prelimed for the attending to finalize in the morning anyway, it didn't cut down on anyone's work. It's just a better schedule since they don't have to come in overnight.
But night shift usually comes at a big premium right?

I mean that recent famous Reddit dude was working 1 week on 2 week off nights - you don’t see that schedule offered as much for pure day shift jobs I wouldn’t think.
 
But night shift usually comes at a big premium right?

I mean that recent famous Reddit dude was working 1 week on 2 week off nights - you don’t see that schedule offered as much for pure day shift jobs I wouldn’t think.
Not particularly, most 1 week on 2 week off jobs make around 450-500k, that guy might have had a special deal worked out or was working in rural Wyoming or something. The schedule is good but you work 60-70 hours the weeks your on and need to be okay switching your circadian rhythms, most people don't want to work deep nights.
 
The market for night rads is relatively small. Maybe one night rad for every 6-9 day rads depending on how you count the hours. That's just an estimate off how much of total radiology volume is in overnight emergency and inpatient studies (as opposed to daytime emergency and inpatient and all outpatient studies). Given that night rads work half or third total time of day rads, if you displaced all night rads to be day rads then it would be just a minor bump in the market.

That guy is somewhere in Illinois, not quite Wyoming. I bet he's cranking at unsafe speeds. Something doesn't make sense about his story. He says he's partner track in a private practice but also counting on PSLF for loan forgiveness, which I thought is limited to nonprofit employers.
 
People have been saying this for years. Reality; avg age of a radiologist is over 55 w avg retirement age of 64, training spots have increased 8-10% since the early 2000s with image volumes continuing to rise. Rvu per scan gets cut yearly, but the technical fee is high, and $/rvu keeps increasing. If AI were to make it to where 1 rad can do the job of 5, there would still be a shortage. AI in rads is also very different (and much more advanced, years ahead. You think this voice recognition is cool? Rads has been using it since the early 2000s) than what the public is exposed to leading to this AI craze, which is why when you talk to a radiologist who is in the know they aren't concerned (even if you disagree)
I know a couple of neuro rads who are pretty concerned.
 
If radiologists are still legally liable for the images they read, that doesn't really increase efficiency even in the setting of a decent AI. Checking images is the bulk of our work, we already have templates and voice to text that maximize efficiency for writing the report. We have pretty good versions of AI overreaders already, they're called senior residents and fellows, but attending radiologists still check every study. The only way AI becomes a threat is if a large subset of scans, or scans as a whole, are taken away from human eyes completely and the hospital/AI company takes on the liability. Too many complex studies, "edge" cases, rare pathology that doesn't have thousands of examples to train AI on, and limited/artifact-riddled studies for that to happen anytime soon.
Aside from the part in which even the privately held radiology practices will use these and the norm will become humans signing off on AI reports with minimal review to maximize RVUs.
 
Unlike some in this thread I don't really have a strong opinion on the subject but at least where I trained (two different institutions) it was almost exclusively residents/fellows doing reads overnight.

If AI could get to the point where it would do the entirety of all night reads / prelim reads for a system and just get double checked by Rads in the AM would that not significantly affect the Rads market in some way?
I was spoiled in residency with 24 hour subspecialty radiology reports...and am still spoiled post residency.
 
Aside from the part in which even the privately held radiology practices will use these and the norm will become humans signing off on AI reports with minimal review to maximize RVUs.
It requires you to check everything the AI says. If you're legally liable you're not blind signing anything. Ask any breast rads if there AI over the years has sped them up, if anything it slows you down now you have to check someone's work.
 
It requires you to check everything the AI says. If you're legally liable you're not blind signing anything. Ask any breast rads if there AI over the years has sped them up, if anything it slows you down now you have to check someone's work.

I think the point is, or i suspect the point is, that AI reads will be legally binding. the risk ultimately be held by the hospitals that use AI, or the outpatient radiology companies where you get imaging done. Someone has to be held liable for AI reads. Or else we will be put malpractice attorneys out of business 🤣
 
AI can even replace CEO and executives AI will disrupt the humans working not just rads.

Also malpractice isn't hat big of an issue for AI you save money and have an insurance company pay even radiologists win
80% of malpractice cases. Since AI learns things can be changed to make more of the liability on the patient if health costs are cut drastically.
 
Med mal is a real issue but realistically.

1) let’s say it missed 2x what a human missed. 1) I doubt that’s the truth and 2) who cares since it can do the work easily of 100 rads. It works 24/7 365.

If each radiologist makes 500k. It’s $50m in savings payments. No HR headache. No benefits. If the AI costs you $10m a year which it won’t. It will be much much cheaper.

That’s $40m to pay out in lawsuits.

Realistically it probably does the work of 400 radiologists. The math is compelling, lawsuits and all. If a normal rads works 40 hours a week out of the 168 hours a week it’s crazy to this the AI couldn’t do it all incredibly cheaper and quicker. Just food for thought.

Also the amount of money you would save in trading future rads is wild. Doesn’t even go into the savings here.
 
Med mal is a real issue but realistically.

1) let’s say it missed 2x what a human missed. 1) I doubt that’s the truth and 2) who cares since it can do the work easily of 100 rads. It works 24/7 365.

If each radiologist makes 500k. It’s $50m in savings payments. No HR headache. No benefits. If the AI costs you $10m a year which it won’t. It will be much much cheaper.

That’s $40m to pay out in lawsuits.

Realistically it probably does the work of 400 radiologists. The math is compelling, lawsuits and all. If a normal rads works 40 hours a week out of the 168 hours a week it’s crazy to this the AI couldn’t do it all incredibly cheaper and quicker. Just food for thought.

Also the amount of money you would save in trading future rads is wild. Doesn’t even go into the savings here.
1. This is assuming the hospital is taking on liability as opposed to the AI company that made the product. I don’t know why any AI CEO would care about saving money on radiologists salaries when they’re not the ones paying them, and would likely just want to continue selling products to hospitals that help but don’t replace radiologists and avoid malpractice.
2. It’s not how many misses, it’s the nature of the miss that costs you. A radiology resident was sued for 120 million last year for missing a basilar stroke that left a young person incapacitated. Plus you’re leaving yourself wide open to a med mal disaster by leaving everything up to a single program. If it has a particular weak spot at missing, say, small lesions that become pancreatic cancer, now you got a lawsuit from every single missed pancreatic cancer diagnosis in the community. That’s a much larger scale miss than a single subpar radiologist could make.
 
Last edited:
The malpractice lawyers are drooling.

I have seen deposition transcripts from this year in which the plaintiff attorney routinely asks the radiologist defendant if they used AI in their interpretation.

The answer's no because it's really not common yet. But the lawyers are ready for when it happens, to impale anyone who might've taken a shortcut that can be considered a breach of the standard of care.
 
I would really love an AI program that robo-calls the referring doctor the incidental findings that require non-urgent but closed loop communication. Getting a hold of people is a pain point, especially outpatient providers, that is a source of malpractice as well.

Example - a surgical PA got a pre-op chest CT and there was an incidental pancreatic cystic lesion for which MRI was recommended. This was never attended to and the patient wound up having metastatic progression of pancreatic cancer. Patient is suing the radiologist for not communicating the result more directly than the report, and the surgeon for not reading the report obviously. Exclusive | Columbia-Presbyterian surgeon failed to notify patient of cyst that grew to stage IV cancer: suit
 
The malpractice lawyers are drooling.

I have seen deposition transcripts from this year in which the plaintiff attorney routinely asks the radiologist defendant if they used AI in their interpretation.

The answer's no because it's really not common yet. But the lawyers are ready for when it happens, to impale anyone who might've taken a shortcut that can be considered a breach of the standard of care.
I think a lot of folks here are viewing rads from an ER perspective. If an onc or GI orders a PET CT or MRCP, they are expecting a consultant-level analysis, not a "no acute cardiopulmonary abnormalities" noted on a cxr or CTA
 
I’m hoping AI leads to massive lawsuits from misses. Just like I’m hoping these ****ty new ER residencies lead to massive board failures. The 80% pass rate this year was excellent news. More job security for people who actually take their careers seriously. Everyone who failed this year already screwed themselves over into taking in person oral boards.
idk how yall old guards are, but the new ER folks are terrible. The average convo with a pgy1-3 EM resident is "hey brooooo I have this dude who is one of yalls patients, ugh he has chest pain but negative trops and no ekg findings. Can you like give your blessing before I dc him?" Go down and its clear costochondritis, now the ccu team has to waste time to do an H+P
 
I’m hoping AI leads to massive lawsuits from misses. Just like I’m hoping these ****ty new ER residencies lead to massive board failures. The 80% pass rate this year was excellent news. More job security for people who actually take their careers seriously. Everyone who failed this year already screwed themselves over into taking in person oral boards.

Wishful thinking. There will be misses but programs are correctable. Anything that saves money and time will see heavy investment.

PE, if anything, will use poor attending performance to justify AI. It's still nascent but unavoidable due to savings, the only thing our masters care about.
 
1. This is assuming the hospital is taking on liability as opposed to the AI company that made the product. I don’t know why any AI CEO would care about saving money on radiologists salaries when they’re not the ones paying them, and would likely just want to continue selling products to hospitals that help but don’t replace radiologists and avoid malpractice.
2. It’s not how many misses, it’s the nature of the miss that costs you. A radiology resident was sued for 120 million last year for missing a basilar stroke that left a young person incapacitated. Plus you’re leaving yourself wide open to a med mal disaster by leaving everything up to a single program. If it has a particular weak spot at missing, say, small lesions that become pancreatic cancer, now you got a lawsuit from every single missed pancreatic cancer diagnosis in the community. That’s a much larger scale miss than a single subpar radiologist could make.
Hospital could indemnify the AI company. Hospital or rad “provider” could soak up so many suits it wouldnt matter.

$120m.. yawn.. how much did they actually get.. hint no where close. I think you are referring to the missed stroke. Note the jury verdict was $120m.. however after you go past the glitzy headlines there was a confidential “post verdict settlement”. Also hint.. it’s less than $120m.. way less.. who knows maybe $40m.. it’s secret. The thing is they will save that in a year. It’s coming.. it will be massive..

The real question is if a human missed this why do you think AI would be worse? The more important question is if a human can make this error why would AI be worse.. never tired. I think the AI would take time. Note that while it might miss pancreatic CA you can teach it. Unlike a human it will only get better with time. It’s the crazy part. It would need a ton of QI.. again it only has to be taught once

The main issue is finding a million incidentals that dont matter. That leads to more and more imaging. But you could have reads within 5 mins of completion.

I think the financial upside is insane though.. If we have learned anything money is the prevailing motivator.

imagine this.. Rads AI Inc goes to congress and says hey you spend (I am making up the number) with Medicare and Medicaid and the VA $40B a year on rad fees. We will save you $30B a year and you pay us $10b and give us immunity.

Imagine the congressman who can say they saved the feds $300B in expenses (Since they often speak in 10 year blocks). Who isnt taking that deal.
 
idk how yall old guards are, but the new ER folks are terrible. The average convo with a pgy1-3 EM resident is "hey brooooo I have this dude who is one of yalls patients, ugh he has chest pain but negative trops and no ekg findings. Can you like give your blessing before I dc him?" Go down and its clear costochondritis, now the ccu team has to waste time to do an H+P
I see this. It’s so bad.. some new Em residents are barely fit to be docs.. it’s embarrassing.
 
I would really love an AI program that robo-calls the referring doctor the incidental findings that require non-urgent but closed loop communication. Getting a hold of people is a pain point, especially outpatient providers, that is a source of malpractice as well.

Example - a surgical PA got a pre-op chest CT and there was an incidental pancreatic cystic lesion for which MRI was recommended. This was never attended to and the patient wound up having metastatic progression of pancreatic cancer. Patient is suing the radiologist for not communicating the result more directly than the report, and the surgeon for not reading the report obviously. Exclusive | Columbia-Presbyterian surgeon failed to notify patient of cyst that grew to stage IV cancer: suit
I mean we already have coordinators who can notify providers, that’s not any big obstacle. The problem is when the Ed doc has a question about the obvious artifact that the AI is convinced is an arterial bleed, and now has no one to talk to about it.
 
Any radiologist who thinks that the complexity of their specialty is protective is fooling them self. There’s something in AI called Moravec’s paradox, basically that high cognition tasks are actually easier for AI to master. It’s harder for an AI to do glorified triage if it means interviewing illiterate people high on drugs, or being a short order cook, etc. We’re just habituated to think being ‘smart’ has intrisinic value to society because that’s what 10 years in college and a career surrounded by other nerdy high achievers teaches - it’s not actually true.
 
Big lawsuits rarely even affect physician assets unless you live in Illinois even then if your married they really can’t go after a lot and it’s very difficult to do
 
idk how yall old guards are, but the new ER folks are terrible. The average convo with a pgy1-3 EM resident is "hey brooooo I have this dude who is one of yalls patients, ugh he has chest pain but negative trops and no ekg findings. Can you like give your blessing before I dc him?" Go down and its clear costochondritis, now the ccu team has to waste time to do an H+P

PGY 15 here.
You're not wrong.
The new kids out here in the community suffer from a serious case of decision making inertia.
I hate it.
 
I think a lot of folks here are viewing rads from an ER perspective. If an onc or GI orders a PET CT or MRCP, they are expecting a consultant-level analysis, not a "no acute cardiopulmonary abnormalities" noted on a cxr or CTA
This is the Emergency Medicine (EM) forum. Of course we are going to view everything through that lens. You wandered over here. Don’t be surprised.

It’s EM. We work in the ED. We don’t work out of a single room anymore.

Don’t be rude and demeaning. Every so often someone wanders into another specialties’ forum and feels the need to trash another’s field. Unless you have dealt with the same challenges through advanced training in a field then you haven’t earned that right.

We also expect consultant-level analysis. Equally, when we catch the acute critical finding that you miss and save you from a legal landmine (even if rarely occurs - and it will happen), then you might start appreciating our field as specialists in acute care medicine.

People have time to microanalyze a tumor. The patient with a SAH or aortic dissection doesn’t have that luxury.
 
I think all this worry about getting sued as a radiologist for an AI or AI-assisted read is a very valid 2024 mindset because we're clearly not there yet with the technology.

However, given the exponential advancements in AI over just the past few years, it’s not far-fetched to envision a future where healthcare systems rely on AI exclusively, given the monumental cost savings to every stakeholder in the healthcare system.
 
I had been working at one place for a while (5-6 years) and every year they would toy with the pay structure in what was clearly an attempt to cut pay, but because we were great at gaming out the system we would always end up with a pay raise because we would just practice/bill/document to the specifications they laid out for $$ generation and never any further. So they would then tweak it a year later and we would do that even better but only to the extent that it paid us better and not to the "document every penny" level the bosses wanted.

Anyway, eventually that place got super toxic unrelated to the pay so I (and most of my coworkers) left, and its been essentially impossible to find any similar pay in the area because everyone else in my area has apparently taken pay cuts since covid and it hasnt yet recovered.
 
Re AI.. It will help us.. it will hurt us.. Same for Rads.. but if you don't think it is happening.. here is the proof..

The global artificial intelligence (AI) chip market size accounted for USD 73.27 billion in 2024, grew to USD 94.44 billion in 2025 and is predicted to surpass around USD 927.76 billion by 2034, representing a healthy CAGR of 28.90% between 2024 and 2034.

AI chip spending

People will be spending nearly 1 TRILLION a year on the chips in 2034.. It is growing at just under 30% a year.. this isn't spending without a healthy return on the investment.

I think AI will be like the iPhone.. the end result will be something we don't see today. No one thought we would use our phone to hail a driver and order food when the iPhone was introduced. Many of the businesses like uber that rely on the iPhone are themselves multibillion dollar companies. Who knows the result. We all should be both excited and worried about the future.

I read ready player 1.. seems like it was fairly predictive of what we may see.
 
People will be spending nearly 1 TRILLION a year on the chips in 2034..
King Of The Hill Eating GIF
 
I would really love an AI program that robo-calls the referring doctor the incidental findings that require non-urgent but closed loop communication. Getting a hold of people is a pain point, especially outpatient providers, that is a source of malpractice as well.

Example - a surgical PA got a pre-op chest CT and there was an incidental pancreatic cystic lesion for which MRI was recommended. This was never attended to and the patient wound up having metastatic progression of pancreatic cancer. Patient is suing the radiologist for not communicating the result more directly than the report, and the surgeon for not reading the report obviously. Exclusive | Columbia-Presbyterian surgeon failed to notify patient of cyst that grew to stage IV cancer: suit
Our local rads group now has a way to put a big pop up on the chart for incidental findings. Pretty amazing.
 
Yes, many radiologists have gotten big raises the past two years. Mainly by leveraging our great job market and doing big grinding at high $/wRVU positions. I know plenty of seven figure rads currently.

Also, yes AI is a threat for much of the routine stuff we do particularly normal CT/MR/x-rays. That being said the tech isn’t quite there yet and every day I have to ignore false positive flags from AI on LVO and ICH. Still can’t do fluoro or IR yet.

Finally, I apologize for the youngsters getting defensive about AI discussion here.
 
Lawyers on the plaintiff side will obviously argue every missed opportunity for diagnosis on the narrowest terms. Miss + patient harm = $$$.

On the defence side – if everyone is using AI-augmented rads, and everyone is having to deal with the impact of AI under-calls, then it can be reasonably argued the "standard of care" is met. It's basically the foundation of arguments against malpractice – if the same protocols and procedures were followed as per the "average" clinician acting on the same clinical information using the same tools for diagnosis, then it's not malpractice.
 
The market for night rads is relatively small. Maybe one night rad for every 6-9 day rads depending on how you count the hours. That's just an estimate off how much of total radiology volume is in overnight emergency and inpatient studies (as opposed to daytime emergency and inpatient and all outpatient studies). Given that night rads work half or third total time of day rads, if you displaced all night rads to be day rads then it would be just a minor bump in the market.

That guy is somewhere in Illinois, not quite Wyoming. I bet he's cranking at unsafe speeds. Something doesn't make sense about his story. He says he's partner track in a private practice but also counting on PSLF for loan forgiveness, which I thought is limited to nonprofit employers.
How do you know it's Illinois? I really doubt any IL group is paying partners 1.5 million. You don't even make that much up in Alaska.
 
The groups in Anchorage and Fairbanks do not pay that much. I've interviewed at both PPs. The biggest group in the state sold to RP in 2021 so there's no way they're pulling 1.5 mil post-sale. The rest of the state is either Indian Health Services, active military or covered by Washington state groups which pay significantly less (600-800ish).
 
Top