Google DeepMind's AI Beats Doctors at Spotting Eye Disease in Scan (cnet.com) 40
DeepMind, Google's artificial intelligence business, is planning clinical trials of technology that can help diagnose eye disease by analyzing medical images after early tests showed its results were more accurate than human doctors. From a report: Published in the scientific journal Nature, the study claims that DeepMind, in partnership with Moorfields Eye Hospital in London, has trained its algorithms to detect over 50 sight-threatening conditions to the same accuracy as expert clinicians. It is also capable of correctly recommending the most appropriate course of action for patients and prioritise those in most urgent need of care. In a project that began two years ago, DeepMind trained its machine learning algorithms using thousands of historic and fully anonymized eye scans to identify diseases that could lead to sight loss. According to the study, they can now do so with 94 percent accuracy, and the hope is that they could eventually be used to transform how eye exams are conducted around the world. You might be wondering why we need AI to do this job that has up until now been carried out by medical staff. But diagnosing eye diseases from ocular scans is incredibly time-consuming for doctors due to their complexity. Due to the aging global population, eye disease is also becoming more prevalent not less, increasing the burden on healthcare systems.
Re: (Score:2)
Well it is good, but more so that our level of healthcare can improve without ever more skyrocketing costs. With an aging population and a nearly never-ending supply of new drugs and treatments, we could probably automate 50% of the medical industry and still have a shortage of doctors. Every time we can move a diagnostic test from requiring 30 minutes of a doctor's time to 30 seconds of a computer's time, that is huge savings.
These stories are often spun as computers taking over a doctor's job, when they r
Re: (Score:3)
Re: (Score:3)
The main reason health care costs more now than in 1950 (age-adjusted) is all the new technology. MRIs still aren't anything like cheap. What people with a political agenda tend to gloss over is that you get much better care now, thanks to that tech.
More accurate diagnostics are a net win, but don't expect costs to come down.
Re: (Score:2)
Oh, administrative costs (and insurance company profits, in the US) may be as much as 1/3rd of overall health care costs, and that sucks. But health care cost has gone up vastly more than 30%, inflation adjusted, in the past 60 years.
Re: (Score:2)
The main reason health care costs more now than in 1950 (age-adjusted) is all the new technology. MRIs still aren't anything like cheap. What people with a political agenda tend to gloss over is that you get much better care now, thanks to that tech.
More accurate diagnostics are a net win, but don't expect costs to come down.
MRIs increase overall healthcare costs because it provides a service which couldn't even be done without the technology. It creates a net new cost (but also improves care). Using software to replace humans providing an existing service is very different, and has a high likelihood of reducing costs. If rules are created were this software can only be used to confirm existing diagnosis then it would be considered a net new service, and would increase costs.
Re: (Score:2)
Doctors spend about 30 seconds looking at such things. Not much cost there to reduce.
Re: (Score:2)
Doctors spend about 30 seconds looking at such things. Not much cost there to reduce.
I am not sure that is true, since TFA says "diagnosing eye diseases from ocular scans is complex and time-consuming for doctors". Perhaps ocular scans are far more complex than most MRI/CT scans, but in my experience it can take days to get results from some MRI scans so I doubt it is a 30 second process.
Then again, there are over 100 million MRI/CT scans performed per year in the US alone, so even 30 seconds saved would be 50 million man-hours of doctors' time saved. That's probably worth around $100 milli
Re: (Score:2)
This is the downside of making docs the gatekeepers for the things patients really need: A LOT of docs make a significant portion of their income from referrals for labs, imaging, prescriptions, etc. Straight kickbacks, of course, are illegal, but there are all kinds of investment groups, etc. that are used to get around these, and it's not uncommon at all for docs in the US to get connected with one of these groups and make tens of thousands of dollars a month (and I've seen reports listing almost a mill
Re: (Score:2)
Unless one company has a monopoly through patents or private data.
Re: (Score:2)
You think adding expensive technology on top of the doctors that are there will decrease costs?
Yes, I think software capable of detecting sight threatening conditions would reduce the cost of preventative eye care. I believe the cost per scan will be less than the cost of having an optometrist do it. Doctors are very expensive.
Automation isn't the problem (Score:3)
we could probably automate 50% of the medical industry and still have a shortage of doctors.
That's a nice little made up statistic you have there. First off a lot of potential automation is refused by patients. They WANT a person to come in and talk to them about what they are experiencing and there is no way to automate this. Second, automation is only cheaper if you can do it in volume. Small medical practices don't have the money for expensive test equipment. There is a reason hospitals have the MRI machine and not your family doctor. This is neither good nor bad but just a reality of aut
Re: (Score:2)
we could probably automate 50% of the medical industry and still have a shortage of doctors.
That's a nice little made up statistic you have there. First off a lot of potential automation is refused by patients.
You can hide the automation from the patient. In cases like the one we're discussing here, the patient wouldn't even be aware of it, because the doctor doesn't study the material in front of them. They can spend the time they would normally use to do so doing something else productive if the computer does it.
Not solving the big problem (Score:2)
You can hide the automation from the patient.
Sometimes though not always. And you can't hide the fact that the doctor isn't there talking to them.
They can spend the time they would normally use to do so doing something else productive if the computer does it.
I guess I'm not making my point clear. The opportunity for time savings isn't generally in diagnosis. That's rather efficient in quite a lot of cases. It's on the administrative side of things that is where the real time burden is and where the opportunity for automation really stands out. Digital medical records, more efficient billing, reducing the need for office staff, etc. If we can improve treatm
Re: (Score:2)
we could probably automate 50% of the medical industry and still have a shortage of doctors
That's a nice little made up statistic you have there.
I should have preceded that statement with an IF. I meant to imply that we would still have plenty of work for human doctors no matter how much we automate (at least for the foreseeable future). I did not mean to imply that we could actually automate 50% of the medical industry, I only used that as a hyperbolic level of automation to prove a point.
While it is true that any time savings is likely a cost savings, I think you may not appreciate how short a time doctors usually spend on a single case. My wife is a pathologist so I see some of this up close. I've seen pathologist go through 100 to as many as 300 cases in a single 8-12 hour shift.
If they are going through that many cases per day, I would say that increases the benefit automation can do. If they were only doing 10 cases per day, then an aut
Re: (Score:2)
And how much will be charged for this service?
A lot, but much less than a doctor or else why use the service in the first place?
The US performs 70 million CT scans per year, and about 40 million MRI exams. Add other exams I don't know about and there is likely a market of around 150 million scans per year in the US alone. Probably closer to 1 billion scans worldwide. Even at $1 per scan, which is worth less than 30 seconds of a doctor's time, that is a billion dollar industry all by itself. My guess is economy of scale will make this service quite chea
I did not see that coming (Score:4, Funny)
But it's great.
London so they have NHS unlike us where (Score:1)
London so they have NHS unlike us where under the GOP system this can be used to quickly black list people.
Re: (Score:3)
London so they have NHS unlike us where under the GOP system this can be used to quickly black list people.
Sure - the NHS just blacklists you for being overweight or a smoker. It's a cheaper blacklisting system that avoids the costs of fancy US tests.
Umm... there just might be a flaw (Score:2)
If it's hard for a human to see whether the scan shows signs of a disease waiting to happen, what was the AI trained by? And by whom? Do we know that the eye scans are actually relevant to the diseases? This is the part that always strikes me odd in those "humans have a hard time to notice X, so we train an AI to do it" stories. If humans have a hard time telling whether something is or is not relevant to a certain disease, and if the AI can only be trained by humans because there is no other source of info
Re: (Score:1)
You feed the computer a bunch of scans and say "these patients ended up having XYZ disease". The diagnoses are determined later, sometimes much later. The doctor may not have initially diagnosed XYZ, but lo and beyond, 3 years later we know for sure the patient indeed has XYZ because by then the disease has progressed further and the diagnosis is clear. (Unfortunately for the patient, it may be too late to fully treat by then...)
You feed the early scans, any intervening scans, and the final diagnoses to the
Re: (Score:2)
Doctors are blind, but not because they can't see. They are blinded by their conflict of interest between patient care and controlling costs. This conflict of interest is never disclosed to patients, but is very real. A doctor will never tell you "hey I think you might have pre-cancerous cells, but I'm not going to order the test because it costs $2000 and it may very well not be pre-cancer. We'll just wait and see if an actual tumor develops," but 99 times out of 100, if a doctor doesn't order a test for something, this is exactly the reasoning.
My wife of 37 years died from cervical cancer that started as precancerous cells in her cervix that were detected but never acted upon by her physician. The doctor wanted to wait another 6 months before running any tests on it because, she said, the cells were not conclusively abnormal.
In the trial, the emails I had subpoenaed showed that the cells were in fact a high grade pre-cancerous legion, but an email from the insurance company instructed the doctor to put off further testing or a LEEP procedure because they were way over that quarters expenditure numbers and that it could probably wait 6 months without issue. I will never forget that email, "these things can always wait 6 more months."
Well, those six months cost her her life, cost me my wife, and cost our children their mother.
I am all for replacing doctors with machines and AI, so long as insurance companies are never allowed to tinker with the algorithms. Let's face it though, they will figure out a way.
Don't wait for a greedy insurance company to murder a family member to demand change. We need a medical system that puts health first.
That is messed up. I hope you publicize this more and make more people aware of what happened.
I've had this opposite experience though. I went to the doctor for a routine checkup and ended up doing 3 tests for no reason. Chest X-ray, EEG and some blood work that I was charged for. They were not expensive tests but there was no reason for any of the tests. I even overheard one of the nurses look at my insurance details and say something giddily about deductibles.
94% accuracy means nothing (Score:1)
By itself, the statement "94% accuracy" means nothing - without an understanding of the rate of false positives and false negatives in the diagnoses. Because of the generally low incidence of the specific diseases in the general population, better than 94% accuracy could be easily achieved by a black box that says "healthy" with respect to all diseases to be diagnosed. Of course, such a black box is totally useless, though it certainly could "transform medical care"!
Re: (Score:1)
I agree that it would be nice to see the actual numbers, but they said it performs better than human doctors. So I doubt it is a black box you described it could be.
Google beats doctor... (Score:2)
Google beats doctor... but your doctor won't share your data with companies all over the web and start advertising to you based on what condition it discovers and trying to take advantage of you financially based on what ails you. (they will just milk you a bunch for your visit instead).
Re: (Score:2)
This is one reason I won't have any Obamacare-compliant "health insurance" plan - I want control over my own health information, and expressly do not want it digitized or shared with anyone electronically, ever. (I've spent way too many years building and working with large-scale electronic medical records systems and healthcare networks to ever want my data in there...)
Re: (Score:2)
Google beats doctor... but your doctor won't share your data with companies all over the web and start advertising to you based on what condition it discovers and trying to take advantage of you financially based on what ails you. (they will just milk you a bunch for your visit instead).
That concern was touched on in the BBC report [bbc.com]. As it stands, this analysis requires the hospital to share patient data with Google. Many people will be understandably unhappy about that.
Why does Google succeed when IBM fails? (Score:1)