Live data from Hacker News

Medicine's Machine Learning Problem

bostonreview.net

71–80 of 112 posts

Re: Medicine's Machine Learning Problem

#71
post #64
post #52

Earlier quoted context omitted.

I do respect your experience and take on the matter, however, let's replace this statement: "I'm an eye surgeon and self-taught machine learning practitioner, I started to learn Python in 2016 when the deep learning hype was at his highest." with: I'm a [machine learning researcher] and self-taught [ophthalmologist], I started to learn [ophthalmology] in 2016 when the [clinical medicine] hype was at his highest. In t…

Haha, you are perfectly right. I totally admit that I'm an amateur with a low level of ML expertise. One the other hand, ML researchers with a deep knowledge expertise are extremely hard to find, even among statisticians / programmers. I suppose that the people with a real expertise are working on their own startup or in FAANG. This leads to a situation where the medical research involving ML is largely without inter…

I think it's partly the incentive structure that is to be blamed. Historically, quantitative PhDs in healthcare(medical physicists, statisticians, comp. genetics) have been underpaid (in my opinion). Now with FAANG and Quant Funds willing to pay $400K+ comp packages to these PhDs, there are far more exit opportunities for these PhDs.

On a positive note, I'm so glad that clinicians are taking interest in ML! As a practicing ophthalmologist, the fact that you were able to self teach is really impressive! I do know that a lot companies are looking for people like you, who have clinical experience. If you are interested you should explore roles/potential collaborations with some of these health research teams in tech.

Re: Medicine's Machine Learning Problem

#72

Earlier quoted context omitted.

Health policy is fraught with counter-intuitive phenomenon - and screening is one of them. Seems like it should help, but in practice leads to over-diagnosis. For example - Cancer rates jumped in Korea after screening with no impact on patient outcomes [1]. There are several others. [1] Lee, J. H., & Shin, S. W. (2014). Overdiagnosis and screening for thyroid cancer in Korea. The Lancet, 384(9957), 1848.

This is a false blanket statement. Also one that could change as we start to see human+ai performance be better than just human performance. For lung cancer screening, NLST showed a 20% reduction in mortality and now NELSON has shown even stronger results in Europe. This “all screening is bad” is FUD in the medical field, frankly. Yes it has to be studied and implemented carefully, but to make blanket statements abou…

I have not stated "all screening is bad".

Broad-based population screenings as the parent comment suggests, in my opinion, are.

I'm yet to see any clinically-valid distinguishing aspects that would suggest AI would add value to screening. Curious to hear evidence that drives your optimism of human+AI.

Just to state, the NELSON study [1] focuses on high-risk segments. Their paper also recommends a "personalized risk-based approach" to screening. This seems reasonable.

[1] https://www.nejm.org/doi/full/10.1056/nejmoa1911793

Re: Medicine's Machine Learning Problem

#73
post #66

Earlier quoted context omitted.

Hi, I did not intend to be disrespectful, sorry if you read my message like this. I mainly intended to underline the fact that we (doctors) were promised a revolution in healthcare (AKA : to disappear) and we ended with diagnostic scores. However, I gladly admit that I exaggerated and that AI technologies can be helpful in some cases, of course.

Geoffrey Hinton really made things hard for folks on the AI side and even walked back that promise. I think it’s the classic thing where it’s overestimated in the short term and underestimated in the long (longggg) term. My sense is that for AI to have the full impact it will one day reach, it will take rethinking medical care entirely with online machine learning and data at the core of how decisions are made. ML wa…

I agree on the very long term possibilities. However, the first problem to solve is the data collection. Saving doctors and nurses from their horrible professional softwares and replacing them with user-friendly, well-thought, data collection friendly softwares would be a huge step forward.

Re: Medicine's Machine Learning Problem

#74

Earlier quoted context omitted.

This is a false blanket statement. Also one that could change as we start to see human+ai performance be better than just human performance. For lung cancer screening, NLST showed a 20% reduction in mortality and now NELSON has shown even stronger results in Europe. This “all screening is bad” is FUD in the medical field, frankly. Yes it has to be studied and implemented carefully, but to make blanket statements abou…

I have not stated "all screening is bad". Broad-based population screenings as the parent comment suggests, in my opinion, are. I'm yet to see any clinically-valid distinguishing aspects that would suggest AI would add value to screening. Curious to hear evidence that drives your optimism of human+AI. Just to state, the NELSON study [1] focuses on high-risk segments. Their paper also recommends a "personalized risk-b…

The general thread here is about AI helping with a more proactive approach to medicine. Screening for high risk populations certainly falls under that.

You certainly said that screening leads to over diagnosis.

I think for screening, the best results are probably the upcoming prospective study from Kheiron.

https://www.kheironmed.com/news/press-release-new-results-sh...

Re: Medicine's Machine Learning Problem

#75
post #73

Earlier quoted context omitted.

Geoffrey Hinton really made things hard for folks on the AI side and even walked back that promise. I think it’s the classic thing where it’s overestimated in the short term and underestimated in the long (longggg) term. My sense is that for AI to have the full impact it will one day reach, it will take rethinking medical care entirely with online machine learning and data at the core of how decisions are made. ML wa…

I agree on the very long term possibilities. However, the first problem to solve is the data collection. Saving doctors and nurses from their horrible professional softwares and replacing them with user-friendly, well-thought, data collection friendly softwares would be a huge step forward.

There are certainly tons of people working on this. I think that the entrenched competitors will only be displaced by other folks who are achieving things they cannot via AI. These two problems are closely linked for sure.

Re: Medicine's Machine Learning Problem

#76

Earlier quoted context omitted.

I have not stated "all screening is bad". Broad-based population screenings as the parent comment suggests, in my opinion, are. I'm yet to see any clinically-valid distinguishing aspects that would suggest AI would add value to screening. Curious to hear evidence that drives your optimism of human+AI. Just to state, the NELSON study [1] focuses on high-risk segments. Their paper also recommends a "personalized risk-b…

The general thread here is about AI helping with a more proactive approach to medicine. Screening for high risk populations certainly falls under that. You certainly said that screening leads to over diagnosis. I think for screening, the best results are probably the upcoming prospective study from Kheiron. https://www.kheironmed.com/news/press-release-new-results-sh...

I suspect, btw, that the Google model in this paper https://www.nature.com/articles/s41586-019-1799-6

will show stronger performance. But Kheiron appears to be ahead as far as proving the value of the tool since they have actually validated prospectively.

Re: Medicine's Machine Learning Problem

#77

Earlier quoted context omitted.

How do we improve on the situation? Given economic realities and racist history (consider what happened in Tuskegee as one example), in the US you would need to provide free screenings to poor people under circumstances that convinced people of color they can trust you while signing the documents to let you have their data. This is a fairly high bar to meet and one most studies are probably making zero effort to real…

Note that lung cancer screening is covered my Medicare and thus already free for anyone over 65 who smoked a pack a day for 30 years (or equivalent aka more in less time). My understanding is that there are many reasons that screening is not deployed more widely but the fact that it requires a 40 minute discussion with a physician, and those physicians in communities in need have very limited time. Then there is the…

I vaguely recall some article about bathtubs being given to poor people in Appalachia who had no running water (in like The Great Depression of the 1930s). They would put them on the front porch and use them to store coal, which got mocked by others as them being "ignorant fools" who didn't understand what a bathtub was for rather than seen as evidence that bathtubs are essentially useless for bathing if you lack running water.

If we nominally have free care available to everyone but there are systemic road blocks that make it essentially impossible for most people of color to access, this is one of those things that falls under "White-splaining."

"Oh, you just do x and it's free" only x is nigh impossible to accomplish if you aren't a fairly well off White person is one of those things that falls under "systemic racism that Whites don't really want to understand."

There's a classist forum that charges $5 for membership and claims this is merely to prevent bots from signing up and is not intended to keep out poor people and all you have to do is ask and they will give you a membership for free if it's a financial hardship. And then the mods make sure to be openly assholish to poor people so poor people won't ask.

When I went to gift a free membership for a "sock puppet" account to an existing member who had said in MeTa she couldn't afford one but needed one for privacy reasons, the mods were quite assholish to me about the entire thing every step of the way in every way possible, including telling me after I had paid "She could have a free second account now for that purpose just for asking" -- something they also hadn't volunteered to her when she said in MeTa she wanted one and couldn't afford it.

It's important that it was in MeTa because that's the only part of the site the mods are required to read all of, so you can't say they just didn't see it. They saw it and declined to inform her "Oh, that's also free for the asking if it's a financial hardships for you. That policy is not only for initial accounts. If you need a sock puppet for privacy and safety reasons, just message us." And then offered to refund me my $5 that I had paid to gift her the account while I was still homeless.

They also did not offer to hook me up with a second account for free. I had eventually paid for a second account for myself while homeless and they didn't offer to refund me $5 at that time either.

I had used the ability to leave anonymous comments a few times and they messaged me to let me know I was a bad girl and a fuck up who was misusing the system as most people only ever left one or two anonymous comments in the entire lifetime of their membership. Nowhere was there any instructions that you should only do that once or twice. It was just the social norm that most people who participated a lot and had privacy concerns had a second sock puppet account for that purpose.

Rather than going "Oh, she's extremely poor and can't afford a second account because she's homeless" they treated me like I was misbehaving. I had no idea I was doing anything "different" until then in part because I was shunned socially because of the extremely toxic classist environment that was openly hateful to me where the mods actively encouraged other members to bully me.

People of color are painfully well aware that the rules are often de facto different for them. People of color often are not notified that X can be had for free or are oblivious to the ways in which it's not really free if you don't already have access to a great deal of infrastructure that Whites have access to on a routine basis and people of color often simply do not have that infrastructure already in place, much like people in Appalachia who can't take a bath even if you give them a free tub because their shack has no running water.

Saying "It's already free...if you can check this box that requires a personal jet to check off" means it's not actually free to most people. It's only free to the current Haves.

Such policies mean that a lot of "freebies" in the US amount to perks for the mostly white Haves, not basic healthcare for all people, regardless of socioeconomic status or skin color.

Re: Medicine's Machine Learning Problem

#78
post #14

I'm an eye surgeon and self-taught machine learning practitioner, I started to learn Python in 2016 when the deep learning hype was at his highest. After 3 years of research, playing with datasets, extracting and cleaning data from EMR and from different machines, I not sure that the biggest problem with the so-called "AI" is the inequalities that it can induce ; it is rather, is it useful at all ? This is a little b…

I worked for a couple of years providing modelling services to health institutions. We did the heavy work under the hood, while med practicioners were mostly interested in getting academic papers published.

For companies with the enterprise expertise it is difficult to enter with the right mindset because sales cycle ib healthcare is way too long.

Confirm that that major botlenecks are indeed getting data in shape for modelling. Feature engineering is key and domain specific. Forget brute force approaches like DL.

Also most AI practicioners in the field seem to ignore that doctors dont need to know who is sick but who is actionable. Its a completely different game.

Re: Medicine's Machine Learning Problem

#79
post #17

Earlier quoted context omitted.

May I suggest, in response to your sentiment that applications of AI to medicine are lacking, is that you are seeing applications replace current medical practices. An AI diagnosis of a medical image seems redundant indeed, however in this situation a patient has seen a doctor out of complaints and has been sent to the radiologist for further investigation. This medical practice is reactionary, and suspicions are alr…

Health policy is fraught with counter-intuitive phenomenon - and screening is one of them. Seems like it should help, but in practice leads to over-diagnosis. For example - Cancer rates jumped in Korea after screening with no impact on patient outcomes [1]. There are several others. [1] Lee, J. H., & Shin, S. W. (2014). Overdiagnosis and screening for thyroid cancer in Korea. The Lancet, 384(9957), 1848.

You can hardly conclude that broadly screening populations are ineffective from this study. You have to consider, among other things, the treatments available for the given disease being screened and the cost of that screening program. If treatments for the disease already have a low success rate (what is low?), the timing of detection doesn't really help. Additionally, if the cost of the screening program is negligible (what is negligible?), then even successfully treating a few patients may be worth it.

Re: Medicine's Machine Learning Problem

#80

Earlier quoted context omitted.

Note that lung cancer screening is covered my Medicare and thus already free for anyone over 65 who smoked a pack a day for 30 years (or equivalent aka more in less time). My understanding is that there are many reasons that screening is not deployed more widely but the fact that it requires a 40 minute discussion with a physician, and those physicians in communities in need have very limited time. Then there is the…

I vaguely recall some article about bathtubs being given to poor people in Appalachia who had no running water (in like The Great Depression of the 1930s). They would put them on the front porch and use them to store coal, which got mocked by others as them being "ignorant fools" who didn't understand what a bathtub was for rather than seen as evidence that bathtubs are essentially useless for bathing if you lack run…

Yeah... I was just trying to explain what the challenges are in hopes that you have a better understanding of what it will take to fix it. For example, the requirement that you have a physician explain things I think should be relaxed as much as is feasible. I'm not blaming people who are poor for not having access to healthcare. Also... I'm not white.
Post reply on HN