Interested in Rads, but worried about the future.

[MedLyfee]

---

Members do not see this ad.

What do you tell a student that is interested in Rads but is worried about its future? Particularly the concern with AI.

While I don’t believe that AIs will fully replace radiologists, I worry that their use will reduce the number of Radiolgists needed in the future and one day we may find ourselves in a terrible job market.

 

[9Rads9]

Job markets and radiology, name a more iconic duo.

 

[CloudEmperor]

Path and job market

 

[howdysang]

Following this thread as I am a med student with this concern as well

 

[Radology]

We’re entering year 3 of the AI hype machine. Thus far, there hasn’t been anything close to showing that it has any real diagnostic capabilities. Not a single article of note in a peer reviewed journal

How long has the inevitable “self driving car takeover” been coming? That’s technology that we KNOW exists and performs better than humans. And it still is not being implemented.

Millions and millions of dollars have been dumped by Silicon Valley into radiology AI. If they continue to show 0 progress, those funds will start to go elsewhere.

So..I wouldn’t worry just yet. At this point it’s all hype

Nobody knows what the future holds. But there’s a much, MUCH better chance of our health care system imploding cutting the need for radiologists than AI

 

[masaraksh]

We’re entering year 3 of the AI hype machine. Thus far, there hasn’t been anything close to showing that it has any real diagnostic capabilities. Not a single article of note in a peer reviewed journal

How long has the inevitable “self driving car takeover” been coming? That’s technology that we KNOW exists and performs better than humans. And it still is not being implemented.

Millions and millions of dollars have been dumped by Silicon Valley into radiology AI. If they continue to show 0 progress, those funds will start to go elsewhere.

So..I wouldn’t worry just yet. At this point it’s all hype

Nobody knows what the future holds. But there’s a much, MUCH better chance of our health care system imploding cutting the need for radiologists than AI

Yeah if our country goes broke, radiology is pretty dispensable - compared to specialties like surgery, medicine, ortho, etc.

 

[Rads4lyfe]

Yeah if our country goes broke, radiology is pretty dispensable - compared to specialties like surgery, medicine, ortho, etc.

You really have no idea of what we do and how reliant on imaging every specialty is. If the radiologists didn't come to work, there'd be a complete hospital meltdown.

 

[masaraksh]

You really have no idea of what we do and how reliant on imaging every specialty is. If the radiologists didn't come to work, there'd be a complete hospital meltdown.

No there wouldn’t

 

[cyberchicken001]

No there wouldn’t

So are you applying to radiology or ortho? If radiology, why apply to a field you believe to be dispensable?

 

[masaraksh]

So are you applying to radiology or ortho? If radiology, why apply to a field you believe to be dispensable?

I’m gonna work for an insurance company

 

[masaraksh]

So are you applying to radiology or ortho? If radiology, why apply to a field you believe to be dispensable?

Honestly if radiologists didn’t show up to work a lot less people would die compared to not having any other specialty show up

 

[cyberchicken001]

Honestly if radiologists didn’t show up to work a lot less people would die compared to not having any other specialty show up

Maybe, but if all the other doctors in the hospital didn't show up we would also be fine because midlevel providers can do just as good a job! /s

 

[masaraksh]

Maybe, but if all the other doctors in the hospital didn't show up we would also be fine because midlevel providers can do just as good a job! /s

You win.

 

[Brorthopedic]

How long has the inevitable “self driving car takeover” been coming? That’s technology that we KNOW exists and performs better than humans. And it still is not being implemented.

There's no real incentive to perfect the technology. Nobody pays you to drive your car, why would a company take on that legal risk when there's no profit to be made?
Rads is different, a large company that is paid millions/year by the hospital to use their AI instead of Radiologists can afford to take on those malpractice claims.

I'm not saying you're wrong, but just food for thought to get this thread back on topic.

 

[808s&heartbreak]

If you fear it then don't do it. Personally I'm pretty excited about the future of radiology. Back in the day when CT was first invented, radiologists fear that they will lose their jobs because clinicians can see the pathology for themselves. The opposite happened; the new technology ushered in new research and paved the way to current radiology. I think AI will be similar. It will make radiology better, in my opinion, but as much as I love AI it will be very slow, if it ever happens. Just reading the hype going around on Twitter right now makes me realize more so how little silicon valley knows about radiology practice. Deep learning is not the answer to general medical intelligence and as long as they continue to use this supervised learning algorithm I think the best that AI can do is displace a fraction of radiologists' work, and that is the most extreme scenario.

 

[Radology]

There's no real incentive to perfect the technology. Nobody pays you to drive your car, why would a company take on that legal risk when there's no profit to be made?

Oh, give me a break.

• 2.4 million truck drivers — 1.6 million truck drivers + 800,000 delivery operators.
• 180,000 taxi drivers, 160,000 Uber drivers, 500,000 school bus drivers, and 160,000 transit bus drivers — another 1 million jobs

There are 34,000 radiologists in America.

 

[Brorthopedic]

Oh, give me a break.

• 2.4 million truck drivers — 1.6 million truck drivers + 800,000 delivery operators.
• 180,000 taxi drivers, 160,000 Uber drivers, 500,000 school bus drivers, and 160,000 transit bus drivers — another 1 million jobs

There are 34,000 radiologists in America.

Right, so imagine the volume of potential lawsuits due that many possible job errors.

Receiving $500k/year to cover 1 person's errors is different than receiving $500k/year to cover 10 person's errors.

 

[808s&heartbreak]

Also, if AI can fully automate the chest radiograph I would love it. In fact I would pay good money to not read a CXR. Get paid $10 and get sued for 13 million for missing a nodule? Thanks but no thanks. AI can have it all. Radiology has got enough new modalities on our plates and way more interesting cases to solve: molecular imaging of cancer with new tracers, dementia imaging, functional neuroimaging, radiomics. It’s a new frontier out there. I see every day the impact radiology has on the clinical team decision making, and I don’t even think radiology is in it’s golden age yet. In the future it will be even more so. I’m betting all my money on it.

 

[Radology]

Right, so imagine the volume of potential lawsuits due that many possible job errors.

Receiving $500k/year to cover 1 person's errors is different than receiving $500k/year to cover 10 person's errors.

“Oh, crap, he countered my dumb argument with facts, quick let me just pull another argument out of the grab bag and pretend that was the core issue all along”

 

[Brorthopedic]

“Oh, crap, he countered my dumb argument with facts, quick let me just pull another argument out of the grab bag and pretend that was the core issue all along”

It wasn't about changing the core issue at all, it was showing you why your argument fails kiddo

 

[ValleyRad2018]

Yeah if our country goes broke, radiology is pretty dispensable - compared to specialties like surgery, medicine, ortho, etc.

The idea that radiology is dispensable is contrary to everything I have experienced in medicine (PGY-2). Hopefully, you were only being facetious. Radiology is involved in literally 99% of inpatient hospital care (as in rarely does an inpatient stay occur without at a minimum a single read from a radiologist) and imaging is utilized with high frequency on an outpatient basis as well.

A big part of the reason I am switching into radiology is because of a realization of just how critical radiology is to patient care. The AI discussion is a different topic altogether, but suggesting specialties like surgery, medicine, ortho etc are less dispensable than radiology is simply not true in my experience. Not even close.

 

[Rads4lyfe]

The idea that radiology is dispensable is contrary to everything I have experienced in medicine (PGY-2). Hopefully, you were only being facetious. Radiology is involved in literally 99% of inpatient hospital care (as in rarely does an inpatient stay occur without at a minimum a single read from a radiologist) and imaging is utilized with high frequency on an outpatient basis as well.

A big part of the reason I am switching into radiologist is because of a realization of just how critical radiology is to patient care. The AI discussion is a different topic altogether, but suggesting specialties like surgery, medicine, ortho etc are less dispensable than radiology is simply not true in my experience. Not even close.

He's likely a med student or a premed given his apparent lack of clinical experience, or he is just trolling.

As far as AI, first of all, we aren't anywhere near having AI in clinical practice. We are likely still many years-decades away. Computers can't even reliably read the few squiggly lines of an EKG, let alone the complexity of a simple radiograph. Forget it reading CT, MRI, or ultrasound anytime soon. Even if we have the technology, the logistics of it would be a whole mess which would take years to implement.

When AI eventually is able to be used clinically, the leaders in radiology that I've spoken with think it will likely be synergistic with radiologists. For example, AI may be able to help with reading more simple things in the future like chest x-rays for pneumonia, counting pulmonary nodules on CT, bone age radiographs, etc. Basically the things that we don't get paid much for and hate to do anyways.

 

[Tiger100]

Honestly if radiologists didn’t show up to work a lot less people would die compared to not having any other specialty show up

Totally wrong.

If radiologists don't show up to work the entire ED will close. ED is the biggest place in the hospital that saves lives.

If a neurosurgeon doesn't show up to work, probably the mortality rate won't change that much, believe it or not.

A CT surgeon does probably 5-6 CABGs a week. 1-2 of them may be marginally beneficial. 1 is going to be life changing. 3 of them are going to improve the survival rate 7-8 years/each on average which is about 24 years of life. Yawwwn. I diagnosed an aortic dissection in a 24 year old on a non-con CT yesterday and probably saved equal to 60 years of life.

People barely think about what they say.

 

[Tiger100]

To OP:
Don't worry about it.

 

[masaraksh]

Totally wrong.

If radiologists don't show up to work the entire ED will close. ED is the biggest place in the hospital that saves lives.

If a neurosurgeon doesn't show up to work, probably the mortality rate won't change that much, believe it or not.

A CT surgeon does probably 5-6 CABGs a week. 1-2 of them may be marginally beneficial. 1 is going to be life changing. 3 of them are going to improve the survival rate 7-8 years/each on average which is about 24 years of life. Yawwwn. I diagnosed an aortic dissection in a 24 year old on a non-con CT yesterday and probably saved equal to 60 years of life.

People barely think about what they say.

Diagnosin’ Ain’t fixin’

 

[Rads4lyfe]

Diagnosin’ Ain’t fixin’

You aren't fixing without diagnosing...

 

[ValleyRad2018]

You aren't fixing without diagnosing...

As a former surgery resident, I can attest to the statement above; but it is clear this person is only trolling now.

 

[masaraksh]

You aren't fixing without diagnosing...

CT scan isn’t the be all end all for diagnosis. Diagnoses have been made just well without imaging - just lower sensitivity and specificity, sure.

And you can def fix without diagnosis - ever heard of Ativan, Dilaudid, Turkey Sandwich?

 

[ValleyRad2018]

CT scan isn’t the be all end all for diagnosis. Diagnoses have been made just well without imaging - just lower sensitivity and specificity, sure.

And you can def fix without diagnosis - ever heard of Ativan, Dilaudid, Turkey Sandwich?

If we are speaking seriously, I will make the counter argument that patients can definitely be hurt without an accurate diagnosis (in many, many ways). In surgery, we relied on radiology more than any other specialty; and I can imagine the same is true in other specialties as well.

If you are trolling, please stop. You've posted >1700 times - frankly, I don't have the time or desire to review your post history - but clearly a frequent contributor to sdn. Don't ruin your credibility or this thread.

Also, most of my patients who were still requiring dilaudid for pain would not be eating a turkey sandwich...

 

[thatsrad]

CT scan isn’t the be all end all for diagnosis. Diagnoses have been made just well without imaging - just lower sensitivity and specificity, sure.

Diagnosis without imaging? - ex lap for everybody! Who needs CTA to look for that dissection when you have sternotomies?

 

[thatsrad]

Right, so imagine the volume of potential lawsuits due that many possible job errors.

Receiving $500k/year to cover 1 person's errors is different than receiving $500k/year to cover 10 person's errors.

There would be less lawsuits with AI drivers because they'd be safer and more accurate than humans.

Point is, there's a lot more money to be made in automating driving than radiology. Plus it's a heck of a lot easier to do, and the technology is supposedly there already. So I wouldn't count on seeing much AI in radiology until after you see driverless cars and trucks become ubiquitous.

 

[Bbqchickenman]

Posted a comment here, but realized it deserves a new thread. New viewers might start reading this thread and think its all AI related. Posted it under “Radiology Market Consolidation.”

Would love to hear your opinions there.

Thanks!