DeepMind readies first commercial product • Hacker News Board

Re: DeepMind readies first commercial product

#101

Post by zaroth » Mon, Apr 01, 2019, 2:40 PM UTC

cure wrote: post #94

bufferoverflow wrote: post #83

Earlier quoted context omitted.

Another high cost of healthcare source is the lack of competition on price. You have zero idea how much each procedure costs. Imagine going to a supermarket where there are no price tags, and you must buy the items once you pick them off the shelf. It's madness.

It's worse. If you have insurance, the price tag will be X, and they will pay for some part (most?) of it. If you don't have insurance, the price tag can easily be 100X and it's your responsibility. If you have insurance, but you happened to pick up that can of pepsi from the left aisle (out of network) instead of the right aisle (in network), they won't cover it, and you're stuck with the 100X price. You won't know…

It should be illegal to charge two different people two different prices for the same procedure at the same facility.

The whole model of negotiated rates, rebates, etc. needs to go.

I also think the concept of networks is bizarre on the face of it. A certified medical provider should be covered to perform procedures in their area of expertise. Period.

Personally I like the idea of anyone who wants to be able to buy into Medicare A & B. And if you don’t have Medicare then you can always pay the Medicare rate of the procedure at 100% (versus having Medicare where your copay is 20%).

If insurance companies can’t compete with that then great.

Re: DeepMind readies first commercial product

#102

Post by tntn » Mon, Apr 01, 2019, 2:41 PM UTC

nitrogen wrote: post #60

bartimus wrote: post #49

Earlier quoted context omitted.

It's my understanding it's sometimes exactly those obvious abnormalities that are missed by radiologists.

Adding another lossy step will only make even more anomalies get missed, or worse, misdiagnosed as something else (I'm thinking about the photocopiers that sometimes change numbers in documents). The argument that "our system is already bad so we should just merge this new bad component because it doesn't make it worse" is bad in software, and unacceptable in medicine.

I don't think that is a fair characterization of what I wrote.

If the scanning process has a sensitivity of 99.999%, and the next stage in the signal chain has a sensitivity of 50%, and we consider what happens if the MRI sensitivity drops to 99.9%, that's well in the noise in terms of diagnostic value. Use some of the extra money freed up by doing the MRI 10x faster to pay a radiologist to look at the scan for 10% longer, and perhaps the net accuracy has improved.

What is unacceptable about this kind of holistic reasoning about the system? High performing systems are frequently not composed of perfect components.

Re: DeepMind readies first commercial product

#103

Post by tntn » Mon, Apr 01, 2019, 2:50 PM UTC

dontreact wrote: post #44

tntn wrote: post #34

Earlier quoted context omitted.

I don't have strong feelings about this, but a few thoughts come to mind regarding exploring accelerated MRIs via neural networks. 1. If an MRI can be done 10x faster with the same results except in exceptionally rare cases, might that not still be a win? Order of magnitude reduction in time may translate to substantial reduction in cost and increase opportunity for applications. It seems like it is worth considering…

1. Maybe, but that should be a tradeoff that is made consciously with some analysis and care, instead of just jumping over a low FDA bar, which is what everyone in this space seems to be doing. I think there are enough abnormalities that can be fatal which a net would just fill in (e.g. aortic dissection?) 2. There are thousands of different abnormalities. From what I understand about the FDA validation process for t…

> And once again, this would be a question that should be studied carefully when people's lives are at risk

I agree 100% with this, but how can this question be studied of this research to develop these systems doesn't go forward? You earlier characterized the FastMRI research as, in your view, "actually very dangerous." As I see it, the research here is potentially extremely valuable, and the danger comes not from the research, but from somebody deciding to deploy it without considering these questions. Typically the road from research program to wide deployment is quite long, and I disagree that we should discourage the research because of potential flaws in the productization process. (Unless there is ill will on the part of the researchers, which I'm assuming there is not).

Re: DeepMind readies first commercial product

#104

Post by johnnycab » Mon, Apr 01, 2019, 2:57 PM UTC

leesec wrote: post #76

easytiger wrote: post #58

Earlier quoted context omitted.

That has been available since the 1970s. Healthcare is not a very complex compsci issue for 80% of cases. The problem is presentation of symptoms is highly subjective and needs highly complex interpretation that no AI will ever be able to achieve and deal with the liability conundrum

>that no AI will ever be able to achieve The Luddites are back! We'll see how this prediction holds up in 30 years.

The jury is out on whether the Luddites were indeed troglodytes[1]. If you were to apply the 30 year nostalgia rule in 2049, then we might be hankering for the present, where 'AI' and the surrounding issues are still in their infancy.

However, a pragmatic approach would suggest that any form of AI and it's derivatives, would be assistive in the medical field and play a hybrid role, rather than being a panacea[2].

[1] https://news.ycombinator.com/item?id=17667375

[2] https://towardsdatascience.com/why-ai-will-not-replace-radio...

Re: DeepMind readies first commercial product

#105

Post by xiphias2 » Mon, Apr 01, 2019, 3:08 PM UTC

eismcc wrote: post #17

xiphias2 wrote: post #13

Earlier quoted context omitted.

Which medical company doesn't want to move costs down? You don't need high costs to charge a lot.

If your business is making money on treating issues, preventing them isn’t aligned. Worst case, you neglect issues until they get to a point where you start making money. Legacy dialysis business is a good exsmple - they need ass-in-seats to justify clinics, so prevention isn’t as big a priority. Also, doctors often have shares in clinic revenue.

This makes sense, but Facebook / Google have enough resources to push their solutions onto the hospitals (give few years for free, maybe give the CEO some other fun stuff).

Re: DeepMind readies first commercial product

#106

Post by dmix » Mon, Apr 01, 2019, 3:13 PM UTC

zaroth wrote: post #101

cure wrote: post #94

Earlier quoted context omitted.

It's worse. If you have insurance, the price tag will be X, and they will pay for some part (most?) of it. If you don't have insurance, the price tag can easily be 100X and it's your responsibility. If you have insurance, but you happened to pick up that can of pepsi from the left aisle (out of network) instead of the right aisle (in network), they won't cover it, and you're stuck with the 100X price. You won't know…

It should be illegal to charge two different people two different prices for the same procedure at the same facility. The whole model of negotiated rates, rebates, etc. needs to go. I also think the concept of networks is bizarre on the face of it. A certified medical provider should be covered to perform procedures in their area of expertise. Period. Personally I like the idea of anyone who wants to be able to buy i…

So the solution is more layers of administration over the prices of the pseudo marketplace?

Re: DeepMind readies first commercial product

#107

Post by est31 » Mon, Apr 01, 2019, 3:13 PM UTC

easytiger wrote: post #58

fluffycat wrote: post #54

Earlier quoted context omitted.

They could be actually better and much faster than humans for initial diagnosis. This has far greater implications, e.g. early detection actually reduces big burden from system.

That has been available since the 1970s. Healthcare is not a very complex compsci issue for 80% of cases. The problem is presentation of symptoms is highly subjective and needs highly complex interpretation that no AI will ever be able to achieve and deal with the liability conundrum

> liability conundrum

So many programs and devices are used where someone would be liable if they malfunctioned. In a production line for example, if something goes wrong and it has to be turned off, every hour costs $$$ to the production plant owner. Similarly for robots: there have been cases where industrial robots have killed people. Accidents with machines can happen in so many industries. If the machine is wrong in 0.2% of cases, that's a risk that can be calculated. If its rate of misdiagnoses is equal to the rate of a human expert, then replacing non-experts with it will improve patient experience. Of course, there might be super experts whose patients would be worse off if they were treated by an AI.

Re: DeepMind readies first commercial product

#108

Post by midnightclubbed » Mon, Apr 01, 2019, 3:28 PM UTC

nightfly wrote: post #57

dr1337 wrote: post #52

The cost blowout in healthcare isn't due to diagnosis or care but administrative costs. All these diagnostic AIs won't significantly reduce the cost of healthcare nor improve quality. What we really need is automated administration and billings to really move the needle and that's more human bottlenecked than technology.

I really don't believe we want more automation in this area. Think of Youtube or Google play store moderation, do you want that for medical billing? And I doubt even 10% of medical costs could be contributed to regular administrative overhead.

Depending on the type of procedure/practice a quick google search indicates somewhere north of 20% is administration. https://www.nytimes.com/2018/07/16/upshot/costs-health-care-...

Every time I visit the doctors in the US there is a trail of paperwork produced. Outside of basic checkups (where you don't dare mention any niggling health issues because your instantly going to get billed consultation fees) there is always a back and forth between the patient, doctors office and insurance over what the doctors office is asking to be paid and what the insurance says it is prepared to pay. This is an enormous time wasting game - the doctors (or their medical group) ask for much more than the procedure cost to perform knowing full well that the insurance company will push back with a lower accepted payment and the patient will sat in the middle having to negotiate between the two.

Removing the pricing game would lower costs substantially (and vastly reduce the time wasted by regular Americans calling their insurance/doctors trying to sort out their individual billing messes).

Having lived in both the USA and the UK there is no question in my mind that the US health care system is a dysfunctional mess. In my experience the only place the US system has an advantage is that the healthcare providers are very (too) willing to send a patient for tests/scans to check for every eventuality and are happy to prescribe whatever medicine is 'best' for an ailment.

In the UK a fit person with a common cold would get sent home and told to not waste the doctors time, in the US you'll get a consultation and a dose of Z-Pack (and hey if I'm paying for my insurance I should use it when I get sick, shouldn't I).

Re: DeepMind readies first commercial product

#109

Post by nitrogen » Mon, Apr 01, 2019, 3:46 PM UTC

tntn wrote: post #102

nitrogen wrote: post #60

Earlier quoted context omitted.

Adding another lossy step will only make even more anomalies get missed, or worse, misdiagnosed as something else (I'm thinking about the photocopiers that sometimes change numbers in documents). The argument that "our system is already bad so we should just merge this new bad component because it doesn't make it worse" is bad in software, and unacceptable in medicine.

I don't think that is a fair characterization of what I wrote. If the scanning process has a sensitivity of 99.999%, and the next stage in the signal chain has a sensitivity of 50%, and we consider what happens if the MRI sensitivity drops to 99.9%, that's well in the noise in terms of diagnostic value. Use some of the extra money freed up by doing the MRI 10x faster to pay a radiologist to look at the scan for 10% l…

It's fine to talk about statistical accuracy when the distribution is somewhat normal. ML systems we've seen so far don't typically degrade in nice, statistical, predictable ways.

Omitting 90% of samples (based on 10x speed claim) leaves an awful lot of room for bizarre errors.

Re: DeepMind readies first commercial product

#110

Post by hyperbovine » Mon, Apr 01, 2019, 4:08 PM UTC

kaolti wrote: post #84

hyperbovine wrote: post #65

Earlier quoted context omitted.

I’ll bet you say the same thing when you are forced to pay for groceries before eating them.

I'm sure you're joking. Information is not a consumable, it doesn't get used up in the process.

And yet it still costs money to produce...