Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

A.I. vs. M.D.

newyorker.com

Post reply on HN
Print view View original

21–30 of 44 posts

Re: A.I. vs. M.D.

#21

Post by na85 »

notfour wrote: post #9
shiven wrote: post #4

A few decades from now, 30 - 50 years at most, doctors and patients, will look back at us and wonder, why we were not using algorithms to assist with diagnosis already.

If you take surgeons, they are trained for one thing. I know a bariatric surgeon who is an expert in medical weight loss but he could not define the word ketogenic. He told me I was foolish for going on a high fat diet (despite the fact that he's overweight). Also my oncologist didn't know if some vitamins could actually help my cancer and he was skeptical that my diet could effect my IGF1 levels. He couldn't tell me…

It's almost like our understanding of cancer is incomplete.

Re: A.I. vs. M.D.

#22

Post by matt4077 »

amelius wrote: post #19

Is there any data openly available for developing this kind of system?

Kaggle has a few datasets, including I believe mammographs and proteomes if you want to do breast cancer specifically.

If you're not in it for the money, or if just fame is enough, I'd suggest to go upriver: individual diagnostic data is rare because of medical privacy and market forces, but there is boundless open data in, for example genomics and proteomics. If you search for [bioinformatics competition], you should find a nice selection of opportunities with good data availability and clearly defined objectives. ML is slowly revolutionising this field, although it's a good idea to pay attention to what happened before–there were some seriously smart people working on these problems for a long time, and they found some rather ways to extract the most value from data with the tools available at the time.

Re: A.I. vs. M.D.

#23

Post by bike4beer »

Nadya wrote: post #5
shiven wrote: post #4

A few decades from now, 30 - 50 years at most, doctors and patients, will look back at us and wonder, why we were not using algorithms to assist with diagnosis already.

As with many things - fear that the AI will be wrong (no matter how less frequently this happens than with doctors) and a desire for a human to blame if there is a mistake. If a doctor makes a misdiagnosis and kills the patient as a result it is less bad than if an AI makes a misdiagnosis and kills the patient as a result. It's the one thing I don't understand about most of society. Same with self driving cars. Self…

A major failure of USA medicine is the legal system, where a DOC if a fails is subject to a mal-practice lawsuit.

Frequently if a doctor say makes $300k/year, its not unheard of him having to pay $150k/year premium for mal-practice.

Ok, now here's the problem, he fails regularly, whoops I killed a patient, and the surviving family get's paid, lawyers get paid, the system moves on.

Where I live, and I can speak for most of ASIA, if a doctor kills a patient, most certainly the family will kill that doctor, now surprisingly mal-practice is RARE, imagine that????

IMHO one BIG reason why USA medicine is such a mess is that the lawyers ( insurance ) have created this fault-free mess where tired doctors do frequently make mistakes and infect, kill, or mis-diagnose their patients. The system largely gives them a big mea-culpa, its all part of doing business.

Given we're now on robots and diagnostic sw, certainly the issue in the USA will be WHO pay's?? IMHO once again our system revolves around going after the 'deepest pockets';

There is no simple solution to the USA way of doing things, thankfully bankruptcy is coming, and out of the ashes the USA will once again learn about self-treatment, and cheap-drugs, just like was done pre 1910, before Rockefeller (big oil ) created the AMA and decided to take over health-care, you know make opium and pot illegal, which had been used by granny forever for stomach-aches, ....

Re: A.I. vs. M.D.

#24

Post by nl »

amelius wrote: post #19

Is there any data openly available for developing this kind of system?

I recently found http://www.cancerimagingarchive.net/ which actually has some reasonably large collections. There's a public 1400 person breast cancer dataset, and instructions for how to apply for access to the 26254 person National Lung Screening Trial dataset.

Re: A.I. vs. M.D.

#25

Post by JusticeJuice »

digitalzombie wrote: post #10

Hopefully more companies are doing this. My research thesis is in this area, hoping for a nice career doing modelling to predict illness. I'd like to have a nice paying job doing something that actually help people instead of just only making somebody more richer. Also it doesn't matter if the AI gets wrong, the algorithm gives say 80% accuracy and it tells you base on your genetic make up if you should take the surg…

I recently talk to somebody from https://www.lumiata.com/

For my thesis I'm trying to design better medical software. Good luck with yours!

Re: A.I. vs. M.D.

#27

Post by tragomaskhalos »

ubernostrum wrote: post #20
adrianN wrote: post #18

Earlier quoted context omitted.

And yet there are cases where humans interfering with the automation caused a crash.

Yes, humans refusing to trust or work with the automation can cause a disaster. That's not an argument for going fully automated or fully human, it's an argument for developing better cooperation between the humans and the automation.

And in many cases it has been poor UI design preventing the humans from recognising or understanding what the machine is telling them, particularly when they are under the extreme stress of an undiagnosed emergency situation.

Re: A.I. vs. M.D.

#28

Post by ppod »

rrherr wrote: post #16

I recommend reading "Do machines actually beat doctors?" by Dr. Luke Oakden-Rayner, who is both a radiology doctor and deep learning researcher. https://lukeoakdenrayner.wordpress.com/2016/11/27/do-compute...

>Can’t call me a cynical, turf-protecting doctor now.

I think the overall tone of the article is a bit harsh on the machines, especially considering the coda.

Re: A.I. vs. M.D.

#29

Post by drchiu »

TuringNYC wrote: post #11
digitalzombie wrote: post #10

Hopefully more companies are doing this. My research thesis is in this area, hoping for a nice career doing modelling to predict illness. I'd like to have a nice paying job doing something that actually help people instead of just only making somebody more richer. Also it doesn't matter if the AI gets wrong, the algorithm gives say 80% accuracy and it tells you base on your genetic make up if you should take the surg…

I am doing this, full-time for almost three years soon. The company is at http://dochuddle.com/ So here is the warning: Medical startups are hard, really hard. - It is difficult to apply research on 32x32 cat icons on 15megapixel xrays. - It is difficult to get data without year-long contract efforts - It is very difficult/expensive, if not impossible, to use cloud resources due to HIPAA and localization rules so you…

This is probably one of the most accurate portrayals of medical systems i countries with highly developed and sophisticated legal system.

It's sad but one of those things that is only apparent once you're neck deep in it.

The people who want to truly help patients are swept aside by those who are in positions to benefit themselves or their friends at the expense of others, including patients. Think administrators, bureaucrats, regulators.

When this thread was raised the other day, I jokingly said how machines could never replace doctors.

The reason is that you can't sue a machine that makes a mistake. Poor medical outcomes like how random flies get hit by cars on a highway.

The powers that be will always insist a doctor be there to sign off. There needs to be a fall guy.

Re: A.I. vs. M.D.

#30

Post by josefx »

adrianN wrote: post #18
ubernostrum wrote: post #15

Earlier quoted context omitted.

I think aviation is a good analogy. Most of the automation in a modern airliner, for example, isn't there to make the plane "fly itself" from point A to point B -- it's there to do things machines are good at in order to reduce the workload of the human crew, so they have more time/energy to focus on the things humans are good at. Flight plans, for example, are still decided on by humans, even if a programmed compute…

And yet there are cases where humans interfering with the automation caused a crash.

There are also cases where humans had to completely disable automation to avoid a crash. The systems are not always redundant enough to detect bad information when just two or three inputs are affected.
Post reply on HN