What specialties can be least easily done by computers?

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

SchroedingrsCat

Full Member
10+ Year Member
Joined
Jan 6, 2012
Messages
69
Reaction score
0
Computers are starting to take over tons of previously high-skill jobs (ex. pharmacy, law cases) and I see much diagnosis work in medical fields is becoming automized as well with IM physicians basically becoming machine technicians. What fields will be good to prevent your job getting taken by a computer?

Members don't see this ad.
 
Members don't see this ad :)
easy answer: surgery of any type...
 
Computers are starting to take over tons of previously high-skill jobs (ex. pharmacy, law cases) and I see much diagnosis work in medical fields is becoming automized as well with IM physicians basically becoming machine technicians. What fields will be good to prevent your job getting taken by a computer?

Gawande addresses this in his book "Complications." He uses a cardiologist vs. a machine for reading EKGs to diagnose MIs as an example.
 
easy answer: surgery of any type...

Have you seen/read about the DaVinci surgical machine...blows my mind. Still requires a surgeon to operate, but blows my mind all the same.
 
Haha so many people are terrified of their jobs being taken over by new technology/outsourcing/nurses/machines/cures for diseases. Just relax.
 
I'm more worried about the Indian robot nurses manning the teleradiology/telesurgery centers.

Until patients are allowed to sue people in India for medical malpractice, it ain't gonna happen
 
Computers are starting to take over tons of previously high-skill jobs (ex. pharmacy, law cases) and I see much diagnosis work in medical fields is becoming automized as well with IM physicians basically becoming machine technicians. What fields will be good to prevent your job getting taken by a computer?

Something as simple as filing one's income taxes, which is taught to all high school seniors in economics, is still done by professionals.

Relax, nothing of this nonsense will happen. However, I agree with some of the above posters that the real threat comes from people replacing people.

Excel at what you do, and you'll always be fine.
 
Members don't see this ad :)
Until patients are allowed to sue people in India for medical malpractice, it ain't gonna happen

You really think that?

Money talks. 10 years from now, if the US were to fall in to dire financial trouble, I doubt they would be so desperate about making sure patients could sue their radiologist, especially if any randomized studies came back to show equivalent effectiveness.

Just realize, anything can happen.

Yet, I don't think any of us are in trouble as far as job security.
 
One of my primary reasons for becoming a physician is a surefire demand for my skills in the upcoming zombie apocalypse.

That being said, while there might be a grind towards computer diagnostics, look at unmanned aircraft. There are still fighter pilots, (or pilots supervising the drones). Physicians will always be needed.
 
One of my primary reasons for becoming a physician is a surefire demand for my skills in the upcoming zombie apocalypse.

You can't use your physician/surgical skills during the zombie apocalypse if you don't make it out of the hospital alive during the first day of the outbreak (assuming it initially presents near you). Just keep this thread in mind for when the first dozen people come into the ED after suffering severe bites, and more and more people keep coming. Get out of the hospital while you can! :smuggrin:





jk
 
Psych
Ob/Gyn (because having your baby delivered by robots would just be creepy)

But really, I don't think being replaced by computers is a huge risk:
1. A computer may be able to diagnose and come up with a treatment algorithm but until we get real artificial intelligence, I don't think it could handle the complexity of actually applying that algorithm to a real, live patient because there are too many variables.
2. Patient comfort. A computer just isn't the same as an actual person...I'm incredibly antisocial and even I acknowledge that. Being left alone in the room with a machine can be quite intimidating. To be fair, doctors can, too, but in my experience as a patient, the number of doctors I've encountered that I've found reassuring far outweighs the number of machines.

But the medical profession will need to continuously assert its relevance. If doctors take an increasingly hands-off role in patient care then, yes, there could be a risk of someone concluding "software could do this."
 
The specialty most in danger of losing to computers is anesthesiology. Computers may not be able to administer the gases correctly, but they damn sure can complete a crossword puzzle.

they took er jerbssss
 
Until patients are allowed to sue people in India for medical malpractice, it ain't gonna happen

Agreed, we will have robots practicing medicine long before we have significant outsourcing, based on our legal system. It will be the US robots worrying about losing their jobs to Indian robots. :laugh:
 
Robots taking over won't happen for a looooong time. Not bc they can't, but b/c our society just isn't mentally prepared for that. Docs make mistakes, and people accept this (though often suing the pants off them in the process), but the tolerance for mistakes from machines is MUCH lower.

For example: Human drivers cause tens of thousands of fatalities annually. But just 1 or 2 caused by failed brake override tech in toyotas triggered a national uproar. All it takes is a few robot surgeons to cut a major artery, a few automated radiology programs to miss a tumor, a few misdiagnosis by robot Dx programs and the backlash will be tremendous, even if the robots are statistically "better" then a human doc.

What is far more likely (and what we are starting to see already) is computer assistance rather than computer replacement. The computer will perform a prelim-read of the radiographic study and highlight areas for the rads doc to focus, but the final read will be done by a radiologist. The computer will spit out a probability weighted Ddx based on a set of symptoms and management recommendations, but actual decisions will be made by the internist. Also remember that someone has to properly identify said symptoms + perform a good/accurate physical exam.
 
...

What is far more likely (and what we are starting to see already) is computer assistance rather than computer replacement. The computer will perform a prelim-read of the radiographic study and highlight areas for the rads doc to focus, but the final read will be done by a radiologist. The computer will spit out a probability weighted Ddx based on a set of symptoms and management recommendations, but actual decisions will be made by the internist. Also remember that someone has to properly identify said symptoms + perform a good/accurate physical exam.

they've already been trying this in radiology (CAD), particularly in mammography, for a number of years, and there are mixed reports as to whether it adds any value whatsoever -- to do it effectively I think the programmers need to better understand how humans recognize patterns and nuances a lot better, given large variations of normal, and from what I've seen, it may never come to fruition. I did some research tangentially related to this as a college student many many years ago and the advances since then are remarkably minimal. Computers such as Watson have been far more effective in matching signs and symptoms to differential diagnoses, so expect useful inroads to come to IM and primary care first.
 
they've already been trying this in radiology (CAD), particularly in mammography, for a number of years, and there are mixed reports as to whether it adds any value whatsoever -- to do it effectively I think the programmers need to better understand how humans recognize patterns and nuances a lot better,

Not necessarily. With massive databases, the computer may be able to "learn" that some lab value or test has predictive value and suggest it to the physician.

given large variations of normal, and from what I've seen, it may never come to fruition. I did some research tangentially related to this as a college student many many years ago and the advances since then are remarkably minimal. Computers such as Watson have been far more effective in matching signs and symptoms to differential diagnoses, so expect useful inroads to come to IM and primary care first.

Best case scenario for Watson: decision support - make sure the guidelines for management are being followed; very fast literature search and association building in instances of rare disease; optimize care.

Worst case scenario for Watson: replace medical school/residency - mid-levels with Watson might be able to do much more at a lower cost than a physician with "acceptable" morbidity/mortality.

One thing that a thing like Watson can't do is move too far beyond the existing data. Patients won't always fit the inclusion criteria for a study, so the management becomes a very gray area.
 
You really think that?

Money talks. 10 years from now, if the US were to fall in to dire financial trouble, I doubt they would be so desperate about making sure patients could sue their radiologist, especially if any randomized studies came back to show equivalent effectiveness.

Just realize, anything can happen.

Yet, I don't think any of us are in trouble as far as job security.

and then someone would need to read all those indian studies. Besides, radiologists do a lot more than that and do have important skills that necessitate being in a hospital.
 
With better EMR, I think the current way we do medical research may be outdated. You could have an intelligent system constantly testing hypotheses based upon all the relevant patient data in a large institution. This could generate much more complicated and accurate diagnostic algorithms than what has typically been developed by large scale studies.

Keeping algorithms simple enough for mental arithmetic is a serious limitation. If we lose that requirement and rely more on tools built into better EMR we can get much better results.
 
Surgery... for every robot you still need a surgeon to operate it. That + the fact that most people wouldnt want a robot doing anything on them even with a surgeon in control.

what about procedure heavy specialties? GI, interventionalists (cards,rads,pulms).
Or things like path? psych?
or even something like derm which requires visual checks and specific skills sets to distinguish.
 
Surgery... for every robot you still need a surgeon to operate it. That + the fact that most people wouldnt want a robot doing anything on them even with a surgeon in control.
Same could have been said about cars - within 10 years I fully intend to purchase a Google self-driving car.

I think the Da Vinci is the Trojan horse here. It reduces all of a surgeons actions to electronic data. It's a long way off, but eventually someone will be able to use that data to create surgical "macros" where the surgeon chooses an action and the machine automates it (eg suturing). From there it's just a matter of time.

what about procedure heavy specialties? GI, interventionalists (cards,rads,pulms).
Interventional/fiber optic procedures should be even simpler to automate. I haven't heard of anyone actively working on this, but when the only movements are torque or advancing a cable/catheter, that isn't outside the realm of possibility.
Or things like path?
Improved image recognition algorithms + labeling for slides. Genomics and proteomics will play a key role are too.
An SSRI pez dispenser and one of those fake boyfriend phone numbers ("How was your day?", "I really care about your feelings...")
or even something like derm which requires visual checks and specific skills sets to distinguish.
Again, image recognition algorithms.

Image recognition is a tough problem, but top minds in private enterprise (Google) and the government (NSA) are actively working on those problems for other reasons. Their advances will eventually trickle down to medicine.
 
... Computers such as Watson have been far more effective in matching signs and symptoms to differential diagnoses, so expect useful inroads to come to IM and primary care first.

Bingo.

Moreover, I'd expect a watson to be used heavily for simple and routine work-ups, therefore reducing the need for mid-levels first and foremost. They'll replace PAs/NPs before MDs/DOs.
 
Same could have been said about cars - within 10 years I fully intend to purchase a Google self-driving car.

I think the Da Vinci is the Trojan horse here. It reduces all of a surgeons actions to electronic data. It's a long way off, but eventually someone will be able to use that data to create surgical "macros" where the surgeon chooses an action and the machine automates it (eg suturing). From there it's just a matter of time.


Interventional/fiber optic procedures should be even simpler to automate. I haven't heard of anyone actively working on this, but when the only movements are torque or advancing a cable/catheter, that isn't outside the realm of possibility.

Improved image recognition algorithms + labeling for slides. Genomics and proteomics will play a key role are too.

An SSRI pez dispenser and one of those fake boyfriend phone numbers ("How was your day?", "I really care about your feelings...")

Again, image recognition algorithms.

Image recognition is a tough problem, but top minds in private enterprise (Google) and the government (NSA) are actively working on those problems for other reasons. Their advances will eventually trickle down to medicine.

I don't know about you all. But these technological advances are amazing and something to look forward to. The future is here!
 
Bingo.

Moreover, I'd expect a watson to be used heavily for simple and routine work-ups, therefore reducing the need for mid-levels first and foremost. They'll replace PAs/NPs before MDs/DOs.

They'll eventually replace a lot of intern scut too. Most basic patient labs and orders could be better accomplished by a machine (assuming you've got a well-integrated EMR).
 
Same could have been said about cars - within 10 years I fully intend to purchase a Google self-driving car.

I think the Da Vinci is the Trojan horse here. It reduces all of a surgeons actions to electronic data. It's a long way off, but eventually someone will be able to use that data to create surgical "macros" where the surgeon chooses an action and the machine automates it (eg suturing). From there it's just a matter of time.


Interventional/fiber optic procedures should be even simpler to automate. I haven't heard of anyone actively working on this, but when the only movements are torque or advancing a cable/catheter, that isn't outside the realm of possibility.

Improved image recognition algorithms + labeling for slides. Genomics and proteomics will play a key role are too.

An SSRI pez dispenser and one of those fake boyfriend phone numbers ("How was your day?", "I really care about your feelings...")

Again, image recognition algorithms.

Image recognition is a tough problem, but top minds in private enterprise (Google) and the government (NSA) are actively working on those problems for other reasons. Their advances will eventually trickle down to medicine.

:laugh: @ fake boyfriend numbers

But none of this will happen in our working lives... because by the time it does, >95% of the population will be unemployed. That number will probably be very close to 99.9% actually.
Medicine will be among the final field hit by extremely high levels of technology anyway.

But your surgical thingy still makes no sense :confused: the amount of possible error that exists would mean only the biggest idiots would accept to be operated on by a robot like that.

But anyway, dont count on a self driving car in 10 years lol. Realistic predictions have shown cars in 10 years wont be much different than cars that exist now.

If you want to think with an out-of-world mentality, keep in mind we should have been on Mars since 10 years ago.
 
:laugh: @ fake boyfriend numbers

But none of this will happen in our working lives... because by the time it does, >95% of the population will be unemployed. That number will probably be very close to 99.9% actually.
Medicine will be among the final field hit by extremely high levels of technology anyway.

But your surgical thingy still makes no sense :confused: the amount of possible error that exists would mean only the biggest idiots would accept to be operated on by a robot like that.

But anyway, dont count on a self driving car in 10 years lol. Realistic predictions have shown cars in 10 years wont be much different than cars that exist now.

If you want to think with an out-of-world mentality, keep in mind we should have been on Mars since 10 years ago.

There are already self-driving cars now, just not commercially available.

http://www.youtube.com/watch?v=-nYhKD8leAg

For legal reasons there will need to be a driver in the driver's seat, but the technology's already here.

And I believe some versions of the Da Vinci robot already allow for superhuman precision and some correction for tremors etc.

Computational power has reached the point where automation threatens white collar as well as blue collar jobs.

We won't be outright replaced, but automation will allow one person to do the work of multiple people, decreasing the demand for human labor.

I, for one, welcome our new robot overlords.
 
what was the conclusion?

It is part of a chapter long discussion (in reference to Atul Gawandes book 'Complications'. In short he makes a couple of main points:

-The computer outperformed the expert. With the same random 2240 EKGs the computer identified about 100 more MIs (750ish vs. 650ish) The expert rarely made obvious mistakes, but the computer picked up on the gray areas.

-Would you take your groceries, put them in a pile, and trust the clerk to say, "Yea, that looks like $17 worth."? No. If the clerk was very well trained they MAY be able to eventually become very accurate, even to the levels of the computer, but why bother?

-Technology often seems to oppose compassion, but the two are not incompatible. He thinks technology may be medicine's best friend. Trust between doctor and patient can only increase when there aren't errors and computers help reduce those.

-Allowing technology to play a bigger role will potentially allow physicians to focus on the care that existed before technology even came about, such as just plain talking to the patients.

-Though a computer can do a lot of the work, it can never add that human touch and take the time to explain to a patient, parent, spouse, etc. what is wrong with a patient and why it is happening.


That was just my interpretation. It is a very appropriate chapter for this topic, easy read and very thought-provoking.
 
But none of this will happen in our working lives... Medicine will be among the final field hit by extremely high levels of technology anyway.

Agreed. Just look at some of the archaic crap that we still use now... Take the stethoscope, for example. Are we really not able to innovate beyond the stethoscope? I have a 32GB touch-screen PC/phone/video game console in my pocket, yet we still we haven't adopted a way to quantify auscultation with cheap hand-held devices....
 
One of my primary reasons for becoming a physician is a surefire demand for my skills in the upcoming zombie apocalypse.

hahha. damn I choose the wrong specialty. Derm will be in high demand when zombies come to power with all their skin lesions. :laugh:
 
There are already self-driving cars now, just not commercially available.

http://www.youtube.com/watch?v=-nYhKD8leAg

For legal reasons there will need to be a driver in the driver's seat, but the technology's already here.

And I believe some versions of the Da Vinci robot already allow for superhuman precision and some correction for tremors etc.

Computational power has reached the point where automation threatens white collar as well as blue collar jobs.

We won't be outright replaced, but automation will allow one person to do the work of multiple people, decreasing the demand for human labor.

I, for one, welcome our new robot overlords.

Do you really know enough about surgery and surgical subspecialties to say that a robot could perform those tasks? cmon man... using suturing as an example...lol.

Just cause technology is available doesn't mean it will be utiliziled to the fullest or..even at all.

And this level of technology at the fullest will = 99.9% unemployment rate. Only jobs will be those of high ranking government officials and some people to manage the robots, but im sure we could make robots to manage other robots too.
 
Agreed. Just look at some of the archaic crap that we still use now... Take the stethoscope, for example. Are we really not able to innovate beyond the stethoscope? I have a 32GB touch-screen PC/phone/video game console in my pocket, yet we still we haven't adopted a way to quantify auscultation with cheap hand-held devices....

You can visualize auscultation with an electronic stethoscope + a smartphone, but you could also just go straight to ultrasound + smartphone?
 
hahha. damn I choose the wrong specialty. Derm will be in high demand when zombies come to power with all their skin lesions. :laugh:

Meanwhile, the ID people will be the first to be zombified...

I have a bad feeling neuro and path, on the other hand, will become obsolete. Either that, or their field will shift from medicine to culinary arts.
 
Do you really know enough about surgery and surgical subspecialties to say that a robot could perform those tasks? cmon man... using suturing as an example...lol.

You're not even a medical student.

Have you ever sutured anything? (I hope not a patient, but you could have tried pigs' feet or orange peels.)

I'm not going into a surgical specialty, but I've sutured plenty of wounds/incisions and honored my surgery rotation.

Suturing would be difficult to automate, but far from impossible. The hardest part would be wound approximation. (And obviously the needles would be different.)
 
You're not even a medical student.

Have you ever sutured anything? (I hope not a patient, but you could have tried pigs' feet or orange peels.)

I'm not going into a surgical specialty, but I've sutured plenty of wounds/incisions and honored my surgery rotation.

Suturing would be difficult to automate, but far from impossible. The hardest part would be wound approximation. (And obviously the needles would be different.)

oh come on... A computer still isn't allowed to give final reads on EKGs because even that is too complex for them most of the time. And you think a computer will be given the power to suture someone's wound? lol not in your lifetime that's for sure.

And buddy, you're radiology... I've done several weeks of radiology and even I know that field is still too vastly complex for computers. They aren't even close to being able to do any of it really (besides mammo but no radiologist or patient is trusting a diagnosis of breast cancer to a computer any time soon). If a computer cannot read a cxr then a computer will not be suturing a surgical wound.
 
oh come on... A computer still isn't allowed to give final reads on EKGs because even that is too complex for them most of the time. And you think a computer will be given the power to suture someone's wound? lol not in your lifetime that's for sure.

And buddy, you're radiology... I've done several weeks of radiology and even I know that field is still too vastly complex for computers. They aren't even close to being able to do any of it really (besides mammo but no radiologist or patient is trusting a diagnosis of breast cancer to a computer any time soon). If a computer cannot read a cxr then a computer will not be suturing a surgical wound.

I've spent more time than you in radiology, and more importantly I've spent more time working with machine learning and artificial intelligence.

Underestimate what is possible at your peril.

If I thought this would happen in the next 10 years, I would not be going into radiology. But in the next 30? Particularly with the transition from plain films to cross-sectional imaging, computer detection has become a much easier problem.
 
The worries about technology replacing doctors is overblown. Technology can only replace things that are simple and highly repetitive--exactly the things doctors no longer do. Those things are already done by techs. example: drug dispensing machines didn't put pharmacists out of work, it put pharmacy techs out of work. If a machine takes simple history from patients like tobacco, EtOH use, LMP, then it leaves more time for me to go deeper into current medical problem.
 
The worries about technology replacing doctors is overblown. Technology can only replace things that are simple and highly repetitive--exactly the things doctors no longer do. Those things are already done by techs. example: drug dispensing machines didn't put pharmacists out of work, it put pharmacy techs out of work. If a machine takes simple history from patients like tobacco, EtOH use, LMP, then it leaves more time for me to go deeper into current medical problem.

Who's worried? I want to help develop it. :D
 
I've spent more time than you in radiology, and more importantly I've spent more time working with machine learning and artificial intelligence.

Underestimate what is possible at your peril.

If I thought this would happen in the next 10 years, I would not be going into radiology. But in the next 30? Particularly with the transition from plain films to cross-sectional imaging, computer detection has become a much easier problem.

Those things aren't going to happen. Not because its not possible from a tech perspective (it is/will be), but b/c our tolerance for error from machines is significantly lower then from humans. We accept that humans make mistakes (though we sue them), but computer/robot error drives us into a frenzy (see the few toyota brake over-ride failure crashes). All it will take is a few missed tumors by the image recognition software or a few screwups by a surgical robot and the backlash will be unbelievable and slow/prevent such change from occuring. What we'll get instead is robotic/computer assistance of human docs.
 
You're not even a medical student.

Have you ever sutured anything? (I hope not a patient, but you could have tried pigs' feet or orange peels.)

I'm not going into a surgical specialty, but I've sutured plenty of wounds/incisions and honored my surgery rotation.

Suturing would be difficult to automate, but far from impossible. The hardest part would be wound approximation. (And obviously the needles would be different.)

lol that's my exact point. Something like suturing requires precision and is a long time away from being done by a robot. Yet you're saying we can have robots removing deep brain tumours and doing spinal fusion on their own.. :luck:

oh come on... A computer still isn't allowed to give final reads on EKGs because even that is too complex for them most of the time. And you think a computer will be given the power to suture someone's wound? lol not in your lifetime that's for sure.

And buddy, you're radiology... I've done several weeks of radiology and even I know that field is still too vastly complex for computers. They aren't even close to being able to do any of it really (besides mammo but no radiologist or patient is trusting a diagnosis of breast cancer to a computer any time soon). If a computer cannot read a cxr then a computer will not be suturing a surgical wound.
This. Key point is, even most of the dumb patients wouldnt trust a machine.
 
Those things aren't going to happen. Not because its not possible from a tech perspective (it is/will be), but b/c our tolerance for error from machines is significantly lower then from humans. We accept that humans make mistakes (though we sue them), but computer/robot error drives us into a frenzy (see the few toyota brake over-ride failure crashes). All it will take is a few missed tumors by the image recognition software or a few screwups by a surgical robot and the backlash will be unbelievable and slow/prevent such change from occuring. What we'll get instead is robotic/computer assistance of human docs.

Oh, I don't disagree, but that doesn't mean the technology won't get to the point where it is as good or better than us.

But instead of direct replacement, it will increase productivity, which will decrease the total number of physicians required for the same work.
 
Oh, I don't disagree, but that doesn't mean the technology won't get to the point where it is as good or better than us.

But instead of direct replacement, it will increase productivity, which will decrease the total number of physicians required for the same work.
Again, by that point we'll have all time records for unemployment rates. You dont think there will be a massive uproar?
 
Top