Hacker News Reader

Comment by giantg2 | Hacker News Reader

"I don't know whether this is a result of..."

The cause is pretty easy. Patients aren't treated like mysteries, they're treated like BAU Jira tickets - just get it done so you're on to the next one. The system is built to handle the 90%. If you fall into that other 10%, it won't work well for you. If you have provider companies and insurances pressing you to hit some metric, that's what you have to do. If you are concerned about malpractice, then you have to just read from the Epic system. No surprise we're in this situation.

View on HN

Up to 25% of women have endo, in some communities at least.

We need more doctors. The nation has grown, our medical professionals and courts must scale up. Automation isn't going to solve everything.

We need more of all healthcare professionals. Nurses and CNAs are getting run ragged just the same.

Have you heard of Iatrogenesis?

https://www.wikipedia.org/wiki/Iatrogenesis

> In 2013, an estimated 142,000 persons died from adverse effects of medical treatment, up from an estimated 94,000 in 1990.

Careful what you wish for!

And, imo, an awful lot of deaths are due to issues that cannot be classified as caused by doctors - eg dosage error over time, something that mitigates symptoms but allows the underlying to deteriorate, etc.

Most premature death in this country is from chronic disease. It's nearly impossible to get a PCP in many parts of the country. Gastroenterology is like a four month wait here.

I’m very ready to have a conversation about whether or not there is a point where we should stop advancing healthcare because people probably shouldn’t be immortal, but doing it by making your healthcare workers struggle under the extreme weight of all the patients they need to care for probably is not the approach.

I simply don’t think we have the medical knowledge necessary yet to know what the “right” number of health care professionals is.

We need more doctors.

Yes, unfortunately we've made doctors the gatekeepers and they don't want more doctors because that will eat into their income. This happens with every single licensed occupation where the license body is run by members of the occupation, with the possible exception of bar associations.

My theory on that one is that the more lawyers you have in society the more lawyers you need because lawyers do a great job creating work for one another (both through litigation but also through legal documentation which needs to be read and interpreted by other lawyers).

The vast majority of doctors support increasing the number of doctors. Most doctors are not members of the AMA and many disagree with their positions on many subjects. The main issue with the shortage is the lack of funding for residencies that is mostly paid for by Medicare and grants. Medical schools don't want to increase the number of students if they won't have slots to finish their training to become doctors. The limits you're talking about are comparatively minor issues as the limits imposed on MDs have led to more DOs. And what nobody wants to talk about are the thousands of slots that go to international medical students, the majority of which do not stay after training.

They currently have a senate bill to increase funding and incentivize the most needed specialties (GP and psych). We'll see how that goes. At this point, it feels like interest in being a doctor has diminished - there's too much training, many specialties don't pay well enough to justify the delayed earnings and costs, the hours can be miserable, and it's a nightmare to deal with all the regulations and legal aspects.

The main issue with the shortage is the lack of funding for residencies

No, that’s the proximate issue. The main issue is the requirement of residencies for all doctors in the first place. In particular, I’m referring to doctors from other countries who may have years or even decades of experience practicing medicine being required to compete for residency spots. These doctors infamously end up driving for Uber instead of practicing medicine.

My mom was one of those lucky few doctors that was able to redo internship and residency after moving to the US. When she first arrived she cleaned houses and worked as a waitress while awaiting entrance into a program. This was the early 80s. Of her friends, she's one of the few who was able to go through the gauntlet and become a doctor in the US. Most became other types of medical workers, including my father who became a respiratory therapist.

Contrast that to my cousin who moved to Australia and was quickly accepted as a practicing physician.

If I trained at a community hospital in Nepal, am I going to need US residency standards? We don’t know. That’s why we have standard board exams and admission pathways that all US physicians need to follow. We should not compromise on that.

We can expand US MD and US DO schools and fill our thousands of unfilled Family Medicine, Internal Medicine, and Pediatric Subspecialty residency slots first!

Board exams are fine. You can sit thousands of candidates for those. The issue is residency slots which are the main bottleneck. Why should a surgeon with 20 years of experience working in Mumbai be forced to become a resident alongside fresh-faced med school grads in the US? It’s a huge problem when it’s cheaper and safer to fly to India for surgery than to get the same surgery in the US (from a less experienced surgeon).

It’s the tyranny of the status quo. Milton Friedman was complaining about this more than 50 years ago.

[1] https://youtu.be/UmVrfbfKBIk

"Why should a surgeon with 20 years of experience working in Mumbai be forced to become a resident alongside fresh-faced med school grads in the US?"

Perhaps you could have them in some accelerated program, but i think it makes a lot of sense to require doctors to train or prove themselves in the specific country/region using the standard practices, tools, etc customary there. I wouldn't want a US doctor operating on me in India since they're likely to make assumptions and recommendations for treatment that overlook local practices, resources, etc.

I think this is a fundamental disagreement, not tyranny of the status quo. We have a functional medical training pipeline already that produces (arguably) the best physician population in the world. Let’s invest there.

The surgeon from Mumbai can come in and prove their worth through a surgical residency. That’s completely fair.

In my opinion we need to elevate doctors and invest heavily into more nurse practitioners and the like. I don't think you need to see a doctor all the time, for procedures or surgery yes, or if you need someone to actually investigate something.

But there is no point in pretending doctors are going to surge, the current administration is trying to make being a doctor harder, not easier. Doctors also avoid certain fields because it doesn't pay enough. The supply is also artificially restricted. It's a system so fucked that it's better to ignore it and pump out PRNs

At this rate, it might be easier to train ChatGPT for better diagnostics performance and give it access to health records than to untangle decades worth of legislative and institutional inertia that prevents "more doctors".

AI hallucinating an art project is one thing…but medicine…?!?

Heaven help us.

>Up to 25% of women have endo, in some communities at least.

I feel like that's a relatively novel insight. These days I hear about endometriosis once a month or so, mostly because of outspoken women advocates. I could swear I've never heard of endo before 2020.

As an example, the Australian National Action Plan for endometriosis is from 2018. Perhaps the currently practising generation of doctors did not have this kind of awareness during their training, and the next generation will fix it?

I think because it's seen as gyn health (though endometriosis can occur in men in rare cases), people often don't feel comfortable talking about it. Relatedly, pregnancy complications weren't really on my radar until I started working in the emergency department and discovered just how common they are.

The studies are wild. The reported range of affected individuals is 2-50% depending on which study you look at. While genetics are the main risk factor, there are all sorts of environmental risk factors that have been linked to it, especially in uetero. Many of those things are highly common such as BPA, PFAS, dioxins, and cosmetic use.

Well, that, and deviation from the established practices can make it difficult to get paid by the insurance company and/or open you up to legal risk (particularly if something goes wrong). Or so I understand from those in the system in the US.

I've had three doctors in the past year tell me that either I wasn't a candidate for cancer treatment because I was gonna die in the next 3 years anyway, or that my best bet was a liver and heart transplant.

All of them read the first page of my medical records upon being admitted to a hospital in January. Not a single one of them read anything after that and it shows. One of them was stricken and amazed that I was walking into his office without aid or an oxygen mask.

Gives me great confidence in their ability to pay attention to detail.

Amusingly this is why people say LLMs will beat doctors. It’s because the 90% of cases is so easy that a motivated guy with Google can get there and a smart NP can get there too.

It isn’t that it’s easy to do all a doctor does. But their training and knowledge shines in the 99th percentile case except they never exercise it there so you can usually get there with Google.

“Oh but an LLM will guess the common case and never think of the rare!”

Yeah but so will a doctor given 10 minutes on it. They’re not exactly going to House MD you. You’re gonna die.

I think a great use case for AI is to act as triage for a new case so that it can send you to the right specialist and have them evaluate you. It could potentially remove the need to see a GP for a referral to a specialist, thus freeing the GP up to spend more time treating others.

How are you going to make it reliably stop people that lie to get into contact with medical professions?

Why does it matter if they do?

I've seen a lot of doctors who insist patients must be making up things when they say "but XYZ", and my question becomes...so what?

If someone credibly lies to you and gets codeine or ritalin or something once or twice...that's not really significant, in terms of negative outcomes.

If someone lies to reach a medical professional, then you treat them like any other bad customer interaction and stop doing business with them after some point.

You're going to have to choose, denying people in need directly, or doing it indirectly because other people game the process and get in the way of those in need.

The length of the queue doesn't determine the length of the priority queue for urgent issues necessarily, and more accessible preventative care could hypothetically go a long way toward reducing the demand on first-contact touchpoints in the US's healthcare.

You don't need a GP for referral to a specialist currently. You just have to pay. It's what I do.

But I understand what you're saying. Insurance gates these but could do so with their own tech rather than relying on the third party. Could help with keeping loss ratios at the minimum.

In my experience (arthritis), specialists typically won't take direct appointments without a referral. They don't want to triage the 90%, and most offices are booked weeks-to-months in advance...

IME, it’s highly dependent on region and specialty.

I can get into local dermatologists without problem. But an endocrinologist takes a referral (because there are so few, and they’re all fully booked).

Specialists differ from GPs in that if you ask them the time, they will have their eyes open when they read the broken clock on the wall.

I have no idea what good GPs serve besides flu testing. Sounds unfair but they generally seem uninformed about pretty common medical conditions.

>Insurance gates these but could do so with their own tech rather than relying on the third party. Could help with keeping loss ratios at the minimum.

I work in insurance. In my experience, the fact that you have to go to the doctor for a referral discourages people from getting said referral.

So the tradeoff is that you would get fewer referral-specific visits (i.e. person going to their GP to get a specialist referral) at the likely expense of more specialist visits.

Strange, I have conditions that require specialists, including a very rare type of specialty.

Never had a problem just calling them up and making an appointment. Insurance never cared either.

Different insurance plans have different restrictions. A PPO typically requires a referral while an HDHP does not.

I always use PPOs. Never had a problem.

In all fairness, I’ve been seen similar specialists for a while. That might make the situation different. “Continuing care.”

I discovered a friend’s chronic medical issue that two VA specialists and a PCP couldn’t figure out using an LLM health project that had been posted here. Works when it works, n=1. Certainly, don’t trust the robot, but it doesn’t hurt to rubber duck debug with it to find blind spots. Fancy search engine sometimes is right (although it can lie too!).

(Bone tuberculosis)

Language models are really good at free association tasks, such as semantic fuzzy search. Next token prediction is among the worst possible ways to use them (although if there's no other obvious way of getting the information out of the model, it works in a pinch).

Which project did you use?

https://news.ycombinator.com/item?id=42974851

https://github.com/OpenHealthForAll/open-health

Aw, it's next-token prediction. This is not generally useful, only a "last resort" research tool.

Why "beat"? Why not "augment" or "improve"?

Indeed.

Imagine if asking a specialist the time meant they had 3 broken clocks on the wall and picked one.

That’s pretty much how initial medical diagnoses are done.

Insurance companies then limit what’s types of care/investigation can be done for various conditions.

Doctor may know that medicine X will work best but insurance demands that Y and Z be first tried before covering X. Same with tests.

I’m tired of it.

What would you propose as an alternative? Healthcare costs are already high. Imposing step therapy requirements to try cheaper treatments first is one of the few ways that insurers have to control costs. And the cheaper treatments do work well for many patients (or they recover on there own just due to time).

Look at any nation that doesn’t have these unbelievably, ridiculously, insanely high costs. This isn’t an unsolved problem, nor is it a sensible one.

What’s with the stupid “here’s the cost” bill you get, followed by the insurance company just, like, deciding things are a different price? What’s with the unreasonably stupid “out of network” medical charges? What’s with the fact that you can walk into a pharmacy and request a discount, but they can’t tell you about it once the transaction has been rung up?

There are SO MANY stupid rules in this system, I literally cannot imagine a system with more low-hanging fruit.

Many of the nations with lower healthcare costs also have step therapy requirements, or simply don't make the most expensive treatments available at all. That is literally one of the cost control mechanisms they use.

Ok, that’s fine? We still don’t have an easy time accessing the cheap and simple stuff here. I have to keep going in to get a doctor to approve a non-controlled substance I’ve been taking for a decade. It has no habit-forming effects. It has no use as a precursor. It has no use recreationally. It’s a worthless, ongoing expense.

Why are you defending this awful, awful system? Have you had even one good experience? Not “well, that wasn’t awful”, but truly one GOOD experience?

Hospitals here don’t even follow up to see if your medical care was effective.

Hell, I’d even take Australia’s approach, where basic stuff is free and you pay for additional coverage.

Public or private medicine shares the same issue. In fact there is no John Hopkins you can pay in the public system so things never get solved.

Yep this is something that only AI can solve. Same situation applies to education, sales, HR. Human powered bureaucracies and systems suck.

Any AI will most certainly reflect the biases of the bureaucracies responsible for their creation.

Nah AI can easily be programmed to be much more patient and investigate edge cases and figure out personalized solutions thoroughly and provide bespoke service. This problem would be solved, though of course there are other issues with biases of the bureaucracies.

If you can do that easily, you will have no shortage of investors. But it’s not easy - getting the data alone is a huge problem.

I think OpenAI has plenty of investors...

https://x.com/deedydas/status/1933370776264323164

OpenAI doesn't lack investment capital. What they still don't have is a good source of high quality clinical data. And this isn't just a matter of buying access to deidentified patient charts from some large health system. Most clinical data quality is kind of crap so using it directly for model training produces garbage output. You need an extensive cleansing and normalization pipeline designed by human clinicians who understand the data at a deep level.

Absolutely true, but is there a system that works perfectly that I can use now that has all that that isn't AI?

In the absence of such a thing OpenAI is already quite good, some theoretical perfect shouldn't be trotted out as a counter if it doesn't actually exist.

There is a massive difference between being able to discuss symptoms and actually being diagnosed. Diagnosing edge cases is precisely the kind of thing that a generative AI needs massive amounts of data on. Where is this magical data source that violates doctor patient confidentiality? And all the doctors just contribute to it and everything is fine? Come on…

That’s true but they have a massive shortage of clinical data. Doctor patient confidentiality is a thing.

AI would be deployed to behave like the median doctor (at best - or maybe the lowest-common-denominator) to avoid blowing up costs with 99%-likely-to-turn-up-nothing hunts for super-rare conditions.

Today you can try to cajole your human doctor into listening more, or ordering more tests, or considering things you heard online or from acquaintances. AI will be guided to take that into account less because a doctor being more sympathetic and bypassing "standard practice" is an expense caused by humanity that the machine can be trained to avoid.

Today you can go across town and try your luck with another doctor. If it's all AI, you'll just repeat your story to the same basic model and get the same basic dismissal.

The problem arose from trying to make people behave like machines in order to save money. Making a machine behave like a machine ain't gonna help.

You need to shift the goal from "saving money" to "helping people." AI doesn't do that.

The costs are so low you can easily inference a bit longer. The idea that a computer would be as lazy as a human is not even close to reality.

Not the cost of running the ai, the cost of potential tests and medical treatment if they do find something, I think?

Nah those make money not cost money (for providers).

You can do an obscene amount of inference for a fraction of the cost of an average doctor's appointment.

>Yep this is something that only AI can solve.

[citation needed]

> Yep this is something that only AI can solve.

How? I'd expect them to already have standardized lists of the most useful next thing to investigate given what's already known, and a modern "AI" would actually be worse at that than some sort of solver engine with a database of costs/risks (for tests) and conditional probabilities.

Maybe if they're still using (digitized versions of) paper flowcharts things could be improved, but the most powerful tech should be old-school stuff rather than modern "AI".

No normal person would actually be able to use a specialized solver database, the woman in this story would already be dead before the guy figures out that such a thing exists and manage to make an account. https://x.com/deedydas/status/1933370776264323164

LLMs already work fantastically with pretty good UX.

Does "here are the most useful things to find out next" really need that complex of a user experience?

Reply on news.ycombinator.com