[smfst7_929]

Dear Nostradamus,

If what you say becomes reality, why do you think proceduralists also wont be replaced by NPs and PAs? Nothing exceedingly difficult about what they do. CRNAs exist dont they? Whos to say a new speciality for NPs and PAs wont pop up. Ill call it RP-NP or RP-PA. Radiological Proceduralist NP/PA.

If we go down your rabbit hole or deeper, might as well just provide rental of equipment so that people can diagnose and treat themselves. I bet my cousin who is a plumber could throw in an IVC filter or even retrieve it. He can fish some amazing things out of pipes.

I think IR is actually more amenable to replacing now. You cant replace diag wity AI yet but NPs and PAs could do everything an IT does with one year of supervised training

[smfst7_929]

No way an NP or PA can replace diags at the current moment without some pretty advanced AI. Theyd minimum of 4 years training to do diag work.

[afazio.uk_887]

How can you believe DR cant be replaced by AI? DR images are just data, just like lab tests. In fact, the scanner doesnt even need to produce an image, it can just send all the raw data to an AI.

[tom.claikens_334]

.

[william.wang_997]

I think OP is partially correct that data analysis is coming faster and faster at us. But we would be the ones putting our signature on the final report.
 
When the EKG report comes out of the machine : Old anterior wall MI or RBBB; there is a disclaimer that Cardiology will file the formal EKG report. OB gyn or Neuroradiology or Ortho doesn’t put their final signature on the EKG report.  
 
We would be doing the same. Adding addendum to the Machine report and finalizing it. It better be us than anyone else.

[afazio.uk_887]

Sure, there are regulatory hurdles and of course privacy and medico-legal issues to be sorted out. But there is literally billions of dollars at work on the other side.

Nathaniel makes a good point that excrement tend to roll down hill and so if physicians are willing to work for far less money they will still have job. However, those who can do things that AI cant will be able to command far higher income – thus a bifurcated system may be created.

[afazio.uk_887]

Genomics has also been digitized now, with extremely low cost genomic testing and sequencing of entire genomes easily and cheaply performed. This is the nature of technology, just gets better and cheaper over time.

In the future, a patient will roll into the hospital – their vitals, all labs tests, all imaging tests, medications, historical data, genomics will all be fed into an AI and it will spit out the mostly like couple of Dx if not the exact Dx. It will also tell you the most effective treatment given the patients genetics etc.

In such a reality, which doctors are really needed? Those who can do procedures to confirm a Dx or a procedure to treat the patients. Thats it really. The rest is just pushing papers around.

Of course, robotics plus AI may someday ever change that also.

[ktmunmun1gmail.com]

What was once said is just more obvious today. For AI to replace radiologist (or any other type of M.D), it will replace most of other jobs we know.
Once that happen, it does not matter if you have a job as a radiologist or not. We will be either living in a new world where everyone gets universal income or it will be the “last of us” scenario
.

[buckeyeguy]

Quote from Mawaly

What was once said is just more obvious today. For AI to replace radiologist (or any other type of M.D), it will replace most of other jobs we know.
Once that happen, it does not matter if you have a job as a radiologist or not. We will be either living in a new world where everyone gets universal income or it will be the “last of us” scenario
.

 
Even smarter. Well done, Mawaly

[emilielegay]

It’s amazing to see you guys sweating over Chat GPT. Wanting to be a good radiologist will be the equivalent of saying I want to be good at counting popsicle sticks.

A computer used to be a person. So will a radiologist.

[afazio.uk_887]

Yeah DR has maybe 10-15 years imo.

Can extend career by doing procedures mostly for another 5 years

[afazio.uk_887]

For me, 20 years is plenty and can live off the portfolio. But my kids i will send the computer science route not medicine

[btomba_77]

I always thought that the first AI medicine replacement would be
 
4 Primary docs —> 4 RNs +1 AI algorithm + 1 MD supervising.
 
 
 
I still think that will happen well out in front of imaging.

[anabarcelo_18]

yes, it’s becoming increasingly clear that many fields – both within and outside of medicine – are in trouble.
 
just be grateful you all are in practice and were at least able to get a little (or a lot ) fat from your career in medicine/ radiology.
 
i, OTOH, am a PGY4 with 300k loans and honestly just feel defeated. so much sacrifice and commitment that will probably amount to nothing. 
 
10-15 years would be a blessing. at least i could pay off my loans. if sooner than that, well i and many of my colleagues are totally screwed. 

[ruszja]

I know chat gpt is going to be able to replace our ER PAs and NPs. They are completely reactive. Feed in the triage note, age and vitals and I can predict what kind of CT they are going to order. Based on what I tell them, I can predict the next CT they are going to order too.

[dzerangel_635]

This is why I believe we need to be proactive and decrease residency spots.  We will need radiologists for the foreseeable future but it seems likely that AI will significantly boost our productivity.  Even if it only boosts productivity by 50%, we will need 50% less radiologists. 

[afazio.uk_887]

well, the “good news” if you want to see it that way is the Radiology workforce is current very old.  So over time if each Rad becomes more productive, it will naturally pick up the slack for retiring / death of the older Rads.  
 
But, I still believe a Rad that has maintained his or her procedural skills will be needed for a long time. 
 
 

[dzerangel_635]

This is why residency funding needs to be shifted more toward procedural/surgical fields.  If anything, AI will generate more incidental findings requiring intervention. Also, we can’t rely on people retiring.  Historically Rads is a specialty where people tend not to retire early or even on time (probably because it is not physically demanding). If AI makes us more efficient and makes our jobs easier, even less people will be retiring.  We need to be proactive and decrease DR residency spots.
 

Quote from Waduh Dong

well, the “good news” if you want to see it that way is the Radiology workforce is current very old.  So over time if each Rad becomes more productive, it will naturally pick up the slack for retiring / death of the older Rads.  

But, I still believe a Rad that has maintained his or her procedural skills will be needed for a long time. 

[william.wang_997]

No.
You keep bringing this up and it is a wrong strategy.
If you don’t read the studies on time, the admin will find the people to read them ( and it won’t be radiologists). This has happened with Cardiac imaging, Ob imaging , vascular imaging, nukes imaging etc. 
 
We are way way short and imaging is growing faster than the number of rads churning out/ burning out.
 
 
 
 

Quote from rads4meplease

This is why residency funding needs to be shifted more toward procedural/surgical fields.  If anything, AI will generate more incidental findings requiring intervention. Also, we can’t rely on people retiring.  Historically Rads is a specialty where people tend not to retire early or even on time (probably because it is not physically demanding). If AI makes us more efficient and makes our jobs easier, even less people will be retiring.  We need to be proactive and decrease DR residency spots.

Quote from Waduh Dong

well, the “good news” if you want to see it that way is the Radiology workforce is current very old.  So over time if each Rad becomes more productive, it will naturally pick up the slack for retiring / death of the older Rads.  

But, I still believe a Rad that has maintained his or her procedural skills will be needed for a long time. 

[buckeyeguy]

Quote from Waduh Dong

well, the “good news” if you want to see it that way is the Radiology workforce is current very old.  So over time if each Rad becomes more productive, it will naturally pick up the slack for retiring / death of the older Rads.  

But, I still believe a Rad that has maintained his or her procedural skills will be needed for a long time. 

 
I would argue what I always have, it’s all marketing. So the fact that it gets rolled out, approved, etc. has nothing to do with quality or what “the people want.” Rather, it has to do with a broke and broken system. And that’s why as the US loses its special status in many ways (has already for decades), including the world reserve starting to slip, it will force this issue. As above, there will be bigger fish to fry for most americans at that point, and the luxury of talking about “health care is a right” will also be an afterthought.

[william.wang_997]

Radiologists who use AI will replace radiologists who don’t. 
 
 
 
 

Quote from dergon

I always thought that the first AI medicine replacement would be

4 Primary docs —> 4 RNs +1 AI algorithm + 1 MD supervising.

I still think that will happen well out in front of imaging.