Ai (artifical Intelligence) In Medical Diagnosis, Would You Trust?

I think its grand that we are utilizing modern Tech in order to better understand what we are up against!

But I still think that human intervention is completely necessary, and should never be underestimated.

 

I think AI in medicine is a great idea and it can be utilised to perform mundane tasks and analysis to free up the specialist's time enabling him/her to spend more time on cases that are much more complicated or have other factors at play requiring consideration beyond the scope of a machine program.

Of course if it was me I would be happy for the machines to do their thing so long as at some stage, if a problem is discovered/encountered, there is a review by a qualified specialist. Diagnostic tests are pretty easy as most are a no problem or problem at some level and analysing test data, the machines will generally be far more consistent than a human but humans excel when there are a few shades of gray because humans can quickly see the bigger picture.

Here is, I think, a very important point even if the AI/machines could be demonstrated to be 99.9% effective and reliable, if you were one of the affected 0.1% failures then nothing would make you happy about the failure because you may have to live with or die because of this failure. So if machines are doing surgical procedures then I think there should be some oversight by humans........off on a tangent.........we have a national health system here in Australia and you can imagine how happy the government would be to reduce the number of surgeons/specialists on the payroll so you need to temper this adoption of automation as a clinical improvement against adopting it as an economic cost cutting measure. The real problem would occur when politicians driving economic savings will say they will cut healthcare funding unless AI is adopted regardless and even though some patients would have a worse outcome the politicians would still be happy because of the savings realised.

As a diagnostic tool AI is already invaluable and is getting better at an astonishing rate and I think soon there will be the equivalent of robot doctors at some level within my lifetime. Many of the tests that were done on me in the 1990s by doctors are now done by machines although I don't get to speak with them...............hmmm some of them have a similar bedside manner to a couple of my doctors.

 

How could it go wrong?
https://i.postimg.cc/ZRDM2jbj/ezgif-2-c3e9566aaa3c.gif

 

I can testify to that.

Last year, 11 days after a triple CABG, waking up, the top third of my vision from the right eye was black. A few days later the cardiologist booked me for appointments at an eye clinic, neuro clinic, as well as a CT scan. At a second visit to the eye clinic the ophthalmologist diagnosed optic neuritis.
And then, the results from the CT scan ruled out optic neuritis. The neurologist and head of the neurology department decided to rely on the diagnoses of the relatively young ophthalmologist even though I'm highly unlikely to contract optic neuritis. I was treated with steroids every day for 5 days. Within 2 days my field of vision improved.
Thankfully there was no shrinkage...

While the above isn't AI it's an example of how indispensable real doctors are. More so considering an ophthalmoscope is a relatively simple machine compared to a CT scanner and it's related equipment.

For those interested...
Optic Neuritis
https://www.webmd.com/multiple-sclerosis/optic-neuritis-ms-vision#1

 

Much as machines can be very helpful, the end answer for me is- can I see a proper doctor, or peace of mind.

 

Something just in........my 88yo mum has a friend that has started to have problems with her eyes and she was telling mum that she doesn't know what to do because she keeps getting her glasses changed but it wasn't really helping. Mum suggested she get a referral to the same specialist eye clinic that successfully repaired my detached retina last year.
Long story short she didn't think it would do much good because she had been seeing her eye specialist in country Victoria for over 30 years (turns out he was only an optometrist).
She ended up getting a referral and they did a whole range of tests using their high tech scanners and machines. As I mentioned previously to DavidGP in a PM when I had my surgery, this clinic has some of the highest tech equipment in Australia so my mum's friend was really surprised that most of the tests were done by machines with a young operator and was thinking that the machines might not be up to the task. She was therefore surprised that after a battery of tests and then a couple of repeat tests to double check the results, she was ushered in to speak with two ophthalmologists. They told her she had glaucoma and there was something they could do to help and suggested surgery as soon as possible. Three days later she had the surgery and has had a very good outcome.
She thanked my mum for giving her the info on the clinic and she now thinks the machines do a great job so is a convert for technology.

 

Thanks for the extra info Dave
The hardest part was mum convincing her to come into town to have it checked out. Mum's friend lives over on the east coast of Victoria (near Lakes Entrance) and it is a 4 hour drive to the big smoke...........she was thinking if there was a problem her optometrist would surely know but as it turns out the optometrist was just in the business of selling new glasses.
Lucky my mum is fairly "au fait" with eye problems and when her friend said that the venetian blinds were no longer straight and the doors on her cupboards looked warped........that's when my mum became insistent on a proper eye specialist. Luckily her friend listened because the specialists said another few months and the surgery wouldn't have been as successful and she would have had ongoing problems. After the surgery she was given the all clear and she just has to take some pills until her next review.

Yes now she is a convert to high tech, she is telling all her friends to get checked out where ever they have modern test equipment.
Out in rural areas where the doctors have a smaller patient base investment in expensive test equipment is not a priority and is seen as a luxury investment they can manage without. Of course this will slowly change as the equipment gets more affordable.

 

It's not either or, they need to work together. The problem is usually the human even when they are right. People don't like to be questioned. That will be the ongoing challenge, can the human doctor accept the help an AI doctor can provide?
On a side note, while not AI, our latest doctor not taking good advice problem. Wifey needed pills, doctor gave her a prescription, nurse mentioned that dose was rare in a single pill and recommended two pills a day of half the dose. Doctor gave her a stern look and she left. A month later we needed to reduce the dose and had to toss $350 in pills as the manufacturer does not allow cutting them. We bought the same pill the nurse recommended originally to cut the dose in half. The doctor maintains he did the right thing.

 

i would hope that its primary objective would be crash avoidance. if that is not possible, then a series of decisions would need to be made. people inside of cars have built in protection, so the priority should be saving ( not crashing into ) the pedestrian/cyclist. the car should be programmed to either crash into another car or other stationary solid object. cars and objects can be repaired or replaced, humans can not be just switched out/easily repaired/completely replaced.

back in october, one week after my 60th birthday, i was on the freeway in stop and go traffic. i was stopped, looked up in the rear view mirror, only to see the car behind me, going really fast, and rear ending me with enough force to push me into the car in front of me. my gut feeling is that in a moment of panic, he stomped on the gas pedal instead of the brake. we were both injury free, but his car was totaled and my car was totaled.

anyway, autonomous cars should protect the pedestrian/cyclist first and foremost. and maybe it is just me, but until they can program autonomous cars to deal with rain, sleet, hail, high wind gusts, snow, flurries, freezing rain, landslides, potholes, etc etc [ every single weather condition imaginable ] i don't ever want to own one. but then, i am old and cranky

 

from my "humanitarian" viewpoint, protecting the pedestrian/cyclist, should be primary, first and foremost. however, the large corporations, who are writing the software, would probably prefer to protect their property ( car/software ) over those annoying humans, since all we do is cause problems and keep them from making money.

it will probably come down to something similar to how worker's compensation insurance works, in that.... i used to do bookkeeping for a company that employed iron workers and built buildings. worker's comp insurance for iron workers erecting buildings 3 stories and under was incredibly expensive compared to buildings over 3 stories. basically, someone falling off a building over 3 stories would die, and all the insurance had to do is pay out death benefits. if the same worker fell off a building under 3 stories, he would most likely survive, but then the insurance company would be on the hook for his medical care for the rest of his life, and that could get really expensive.

so, unless compelled by law, i see those autonomous car corporations taking a similar viewpoint. basically : humans are costly, property is not.

 

Thats the problem.........cyclists and pedestrians need to take heed about their own safety. If we continue to place them in the 'Hard done by' bin, then they will continue to not take care at major congestion points, because they can.

Yes, we should be aware of all road users be it on wheels or afoot, but placing the responsibility upon those who occupy a larger surface will never work.

Cars are a luxury we know, they are designed to please those willing to pay for the privelage. So Yes, they should absolutely be designed to protect those that fork out for that privilege, not the third-party.
Why would a multi-million dollar business do otherwise??