Hacker News Reader

Comment by codingslave | Hacker News Reader

I've long believed that once startups and tech companies get their hands on medical data (patient outcomes, medication treatment outcomes, mal practice rates, surgery outcomes) that a reckoning will take place on the medical industry. The level of legitimacy and respect that doctors and hospitals currently have is far too high, and hinges on branding and the exclusivity of becoming a doctor. Once all of this data becomes available, it will be obvious how worthless and expensive most medical treatments really are. Most prescription drugs do nothing, and the ones that do something are more so just covering the symptoms.

HIPAA laws are hiding patient outcomes and holding this whole sham afloat.

View on HN

> Once all of this data becomes available, it will be obvious how worthless and expensive most medical treatments really are. Most prescription drugs do nothing, and the ones that do something are more so just covering the symptoms.

I've read Bad Pharma (and will recommend it to anyone interested in this subject). And one way to summarize (perhaps oversimplify) it is: pharmaceutical corporations manipulate data to justify expensive drugs of questionable effectivness.

But to think "startups and tech companies" are going to help solve this problem... man, that's optimistic. And as Bad Pharma demonstrates, the data is already available. There's just not a lot of interest (economically or politically, especially where those two are strongly correlated) to seek it out or listen to it.

Also I don't think it was HIPAA's intention to hide data from well-meaning startups. That's just an unintended consequence perhaps. I'd like to think its primary motivation was to protect patient data from predatory insurance companies and unscrupulous marketing forces. (Does it succeed in doing that? I don't know. Are startups and tech companies going to help with that if it doesn't? Hmmm.)

I briefly worked for a startup back in 2005 that was doing data analysis on prescriptions to reverse analyze a patient's conditions.

The customers were the insurance companies.

How does that work? Insurance companies have access to all the prescription purchase and tests ordered,but not the diagnoses?

> I've long believed that once startups and tech companies get their hands on medical data

Honestly I think the last people that are going to improve anything in medicine are startups and tech companies.

Quite the contrary. I know that quite a few people in evidence-based medicine are worried that Tech people who want to "disrupt" things come with their own poor scientific standards to undermine the minimum standards this field has (and has fought hard for to get).

We've started with ubiome/23andme providing customers data previously not available.

Another area is improving existing processes. We see techniques where photos are being read more accurately and quickly compared to manual processes.

Medicine now resolves around tests. Those can be improved and automated.

We will see tools that will empower personal medicine and shift the power balance away from the family clinics.

To be fair, ubiome is a fraud, and (anecdote) 23andme made a mistake in reading my genome and their data showed that I had an incredibly deadly cancer predisposition (Li-Fraumeni), which a reputable medical lab showed to be false.

ubiome is a ycombinator company.

Why would you call them a fraud. I've done a number of tests over a period after antiobiotics and I definately saw the change into diversity correlate with the foods I introduced.

They've been raided by the FBI for insurance fraud, they reportedly conducted "tests" using samples that hadn't been stored properly, their CEOs/founders have resigned, they've let half their staff go recently and currently are not capable of actually doing testing ...

That's not saying that they never produced useful data, but ...

You'd be surprised how many scams Ycombinator accepts. Quantstamp cough cough.

23andme has been found to be highly inaccurate.

https://www.livescience.com/63997-dna-ancestry-test-results-...

There were and are a lot of articles about the fact that many prescription drugs and many treatments are overrated and I totally belive that. But saying that "most" medical treatments & prescription drugs are worthless it's quite a radical view based on... pure speculation.

I think your statement sounds perfectly reasonable, but it actually holds a pretty interesting corollary! If he is wrong that "most" medical treatments and prescriptions are worthless then that would necessarily mean that "most" are valuable. The reason this is interesting is because the exponentially increasing rate of both the amount of medication out there and the amount people are using. For instance from 1997 to 2016 the number of prescriptions filled grew from 2.4 billion to 4.5 billion [1]. And an exponentially increasing number of people, across all ages and demographics, are also being medicated with all sorts of drugs for various reasons.

So, if we assume that most medical treatments and prescriptions are not worthless we should see increases in healthfulness that are at least somewhat comparable with our increases in pharmaceutical usage. In reality we saw sharply diminishing returns on life expectancy that have now gradually shifted towards an ongoing decline in life expectancy over the past few years. By any standard of effectiveness, this is not looking good for pharmaceuticals.

So I suppose I'd turn it around. I think the view that medicines (and especially their recent exponential growth) are providing a substantial social benefit could be true, but it would seem to be the statement that needs to be explained. E.g. perhaps our own increasingly unhealthy lifestyles are contravening the gains from the medical technology? No idea there, but I do think occam's razor would tend to simply suggest that the reason we're not seeing improvements is because we're not improving.

[1] - https://www.webmd.com/drug-medication/news/20170803/american...

I agree with you but in a different way :) (what follows its a long text written in bad English but I hope I'll manage to make my idea clear)

Regarding your idea that "By any standard of effectiveness, this is not looking good for pharmaceuticals": This is a very complex topic and I don't think anyone has a clear and objective view/explanation about this.

For example, we have a big problem with the antibiotics. We eat a lot of meat from animals that where treated with antibiotics. We "overuse" antibiotics. Etc. Because of this, new antibiotics-resistant bacteria appeared and classical antibiotics became less and less effective in fighting against them. How should we categorize this issue? Should we see it as a proof of the fact that medicines are not as effective as advertised? It's not a problem with antibiotics per se. Or as a proof that modern medicine is overrated? Medicine correctly identified the problem(bacteria) and the treatment(antibiotics). In reality it is a problem generated by mostly by the rising popularity of automedication, lack of basic medical/pharmaceutical education and urban myths (antibiotics are good for anything), medicines used on animals (which is not a problem of human medicine), and, I would say, only a tiny part of the problem is the treatments that unnecessarily include antibiotics and are prescribed by medicine doctors. But, let's say that we don't know anything about the above mentioned facts and we study some statistics. What would we think when seeing a clear correlation between the increase in the number of antibiotics and decrease of the effectiveness of the treatments? We could see it as a proof that "the quality of medicine drugs has decreased resulting in an increase of ineffective treatment".

The point I'm trying to make is that the reason why the effectiveness of many treatments decreased and now we have a decline in life expectancy it's a very complex issue and it's not fair to see it as simple as a fact failure of pharmaceutics or/and medicine.

Regarding your last statement, I think that we are not seeing improvements because our modern environment and life style has changed and is changing in such a rapid and unpredictable way that medicine can keep the pace. Medicine it's improving, but it's improving too slow. It is as out environment and lifestyle is traveling by plane while medicine and pharmaceutics are traveling by bus...

The health supplement industry is still alive and kicking. And that's an industry that's so sleazy many companies can't be bothered to even put their claimed active ingredients in their products. Yet it's a very healthy industry that has very cushy anti-regulation protections.

So even if you were right on the lack of efficacy of modern medical treatment (and I strongly suspect you are not), I highly doubt that its exposure would lead to a collapse in the industry.

>I've long believed that once startups and tech companies get their hands on medical data (patient outcomes, medication treatment outcomes, mal practice rates, surgery outcomes) that a reckoning will take place on the medical industry.

Startups have been going at this problem for years and years already and have had the data. I've worked at them. You know what the goal of most of those startups was? To make money. Patients were a means to an end. If they helped people, it was a thing that happened but not the actual goal.

See, actually trying to help patients and do so responsibly and not just cause a bigger mess is hard. Medicine is complicated, people are complicated, the health system is complicated. Easier to just ignore all that, make some money and sell out before people realize all you did was funnel some money around.

>Once all of this data becomes available, it will be obvious how worthless and expensive most medical treatments really are.

There's plenty of existing literature that notes various medical treatments are worthless. Yet they still happen and are still prescribed and are still paid for. Knowledge doesn't change behavior.

>HIPAA laws are hiding patient outcomes and holding this whole sham afloat.

You can easily get medical billing data for a massive percentage of the US population which includes doctors, dates, procedures, diagnoses and aggregated costs. De-identified but that doesn't matter for these sorts of things.

edit: Also, what is probably the biggest hurdle for health care startups isn't data. It's the fact that AB testing, the cornerstone of how every tech company iterates, is considered unethical except in very constrained forms. Something about there being a difference between crashing the web page of 1% of your customers versus literally killing 1% of your customers.

Thank you for articulating these points so well. Possibly preaching to the choir here, but you should check out "Geek Heresy" by Kentaro Toyama.

How would a startup make money by showing that treatments are worthless? Startups are going to have the same incentives as other pharmaceutical companies.

Well, the hope I have (and OP probably as well) is that machine learning techniques properly applied to the raw data could uncover a lot of statistical errors in the historical literature, which could be both profitable and disruptive.

... Unfortunately, this may happen first in countries with poor human rights records (i.e. China) because there is likely a greater availability of patient data under those circumstances.

ML doesn't fix bad statistics. ML is really just a sexier way of saying real-time statistics with lots of data.

I would agree that ML is just statistics, but I also kind of think that a median approach to statistics in medicine are currently at an even poorer level of quality than the low bar set by typical ML, so there could be an improvement nonetheless.

ML is not in a state to be of much use in this context. Source: me, I work as a data scientist to apply ML and other techniques, with claims, EHR, and a bunch of other data, specifically for the purpose of identification of conditions, effectiveness of treatments and non-medical intervention. The tech world is very much in a condition of being unjustifiably arrogant about its ability to solve problems in this space. The problems are very hard, and the answers aren't gonna be found by some egotistical software eng who tries to shift his toy kaggle crap into this space, or the kinds of startups that are built up around that kind of thinking.

I kind of agree- A rigorous solution to this problem would probably involve running large numbers of (human?) cell cultures with deep analysis fed through ML systems. i.e. we first need to have a full understanding of cell chemistry/physiology and then scale it up to larger organs and the whole human organism.

ML will simply be used to find all sorts of new spurious correlations that will spawn a whole new flood of unnecessary medications and treatments. After all, statistics puts a whole lot more care into avoiding those sorts of errors than ML and, well, here we are anyway. Technology doesn't solve societal problems.

Yeah you see that's the problem. The raw data is not there, because we don't have the necessary information retrieval capability yet.

Data is really sparse and of the worst quality. It is interesting that the hype is exactly the opposite: big bad hospitals keep essential info for themselves.

It's a deep, wide and sensitive subject. White coats give people extreme social powers. But they are far from perfect.

At a high level, I agree. My take is that for things like broken bones, and other accute trauma, doctors are great. For everything else, especially anything chronic, and they are completely worthless.

Any good doctor will advise you to get plenty of exercise and eat a health diet. That's easily the best thing you can do for your health.

Only the vast majority of people ignore that advise.

Also the combination of vaccinations and nutrient-fortified food have worked wonders for public health.

You can get the same advice from a crossfit website. Doctors aren’t the ones fortifying food and nurses seem perfectly capable of administering a vaccine.

Oh yeah, of course! Startups and tech are so much more honest than big pharma and clinicians...

The reckoning will not go the way you think it will.

Anti-virus vendors have been doing great, what leads you to believe that medical tech startups won't just look to "disrupt" the market by funneling the money into a different (likely: their) direction?

Maybe you just described why this data will not become available...

"The level of legitimacy and respect that doctors and hospitals currently have is far too high, and hinges on branding and the exclusivity of becoming a doctor. Once all of this data becomes available, it will be obvious how worthless and expensive most medical treatments really are."

I'm the spouse/partner of a recently-trained doctor, and I'm a programmer and startup founder with 2 decades of experience. With all respect, you have no friggen clue what you're talking about.

Would it be possible for you to be more specific about the ways in which the commenter is wrong?

Have tech solved the issues of a client changing their minds about what they want?

Well, all the interesting bits of medicine are built on social interactions. You ask the patient where it hurts and they lie about where.

I don't think this is relevant to my question, or am I missing something?

are those the same social interactions that explain why surgeons resisted using things like checklist because it was "beneath them", "they never make mistakes", etc.?

because if so i agree, there's some real low hanging fruit there.

Would you kindly elaborate on that?

I have enormous respect for the difficulty of the training that doctors like your spouse have gone through, and yet I share the quoted sentiment. We assume that difficulty is essential to the work, and I think that's an assumption worth questioning, as the dental profession has been recently.

Yes, doctors are very smart and well trained. But medicine as a whole has a lot of completely insane practices.

For example, have you ever tried to sleep in a hospital? They generally run very cool, daylight temperature lights 24/7, with tons of noisy machines, and the nurses wake patients every hour.

And it's not just the patients who are getting their circadian rhythms destroyed, the residents themselves are doing 30+ hour shifts. It dates back to the first medical residency, started by William Stewart Halsted who was a cocaine fiend and expected everyone else to keep the same hours.

We know that sleep deprivation impairs your driving ability just as much as being drunk. So why do we accept doctors and nurses being sleep deprived?

My understanding is that they believe they are exception to the rule and that such schedule is more effective for them.

Because money. You may become sick and have to deal with a sleep-deprived doc someday. But you have to pay said doc every month as sure as the sun rises. Plus, being sick only happens to others, right?

> I'm the spouse/partner of a recently-trained doctor, and I'm a programmer and startup founder with 2 decades of experience. With all respect, you have no friggen clue what you're talking about.

Then you should know very well that the entire MD education pipeline optimizes for rote memorization and regurgitation of facts (from undergrad pre-med all the way through post-residency fellowships). Any patient with even the slightest deviation from what the physician searched on WebMD is given generic advice similar to how fortune tellers recite generic phrases. If you have a non-standard question or health problem, go to 3 doctors (or even specialists) and you'll get 3 different answers.

I work closely with clinicians at Tier 1 urban medical institutions daily, both at the in-patient and outpatient settings. Most doctors are dumb as a rock.

Note: Physicians with a MD/PhD are much better at critical thinking and make more of an effort to solve your problem.

By definition, most patient issues don’t deviate from the normal. Optimizing for the most common case makes sense.

This lines up extremely well with my personal experiences over the years.

It makes total sense if you think about it: medicine is an infinitely complex field, and at the end of the day even though doctors have (in the past) undertaken significant training, they are still human beings, subject to the same limitations and errors as the rest of us.

New discoveries and nuanced understandings are made in medicine constantly, the idea that it is safe to assume that any individual doctor is going to know and understand all discoveries seems to me like extremely wishful thinking.

I can second this! I studied medicine for two years in a SE Asian country. My late dad was a doctor. Both of my host family parents are doctors in the US. My gf is a doctor.

The main reason I left med school and decided to study math and comp sci. is because of the need of rote memorization necessary in medical training. This is especially worse in the US because the system here requires 4 yrs (undergrad) + 4 yrs (med school) + 3 yrs (residency) = 11 years of training to become a doctor. In contrast, it takes 5 yrs (med school) + 2-3 yrs (residency) of training to become a doctor in my home country.

The journey to become a doctor in the US is unnecessarily lengthy and it makes doctors feel more entitled (because they have spent a lot of time and money as investment) to charge exorbitant fees despite what most of them are doing is simply prescribing antibiotics; analgesic; insulin; and garden-variety stuff (or read up "Up-to-date" [https://www.uptodate.com/home] to refresh on what they learned or what the current consensus treatment is).

Most of them do the same thing day-in-day-out. Sure, there may be some specialists that are top notch, but they are like the minority and even their work is mostly about being able to keep steady hand and cool head during emergency situations.

And there's no Dunning-Kruger effect at play here, at all. You're entirely equipped to evaluate highly-trained technical specialists in a field not your own.

What you don't perceive is why medical education has to be that way in the current state of affairs. Yes, has to.

You may be brilliant at other things, but your analysis of this particular problem is no better than the doctors you describe.

> What you don't perceive is why medical education has to be that way in the current state of affairs. Yes, has to.

No, medical education does not have to be this way. It is purely gatekeeping.

It's similar to how "You need a PhD" to land top AI/ML jobs. That's gatekeeping as well, though it is slowly changing.

Not many doctors or AI researchers will openly state this gatekeeping because they benefit, at the expense of those not in their field.

Nope, not gatekeeping. I know I won't convince you, but just think about that:

Most parts of the country are approximately medical deserts. So once you're educated, you are dispatched to said desert. Your hospital has standard pathways where nursing teams operate. Problem is, everything that escapes the standard pathway comes your way with no one to help you.

"Yeah, but there's UpToDate, MEDLINE, etc.". Well, no. NO! You simply don't have the time. Your job is to operate as a well-oiled machine to handle all the impromptu shit 70+ hours/week. That's why you did all this rote memorization.

A doctor is not an engineer. He is trained for a hostile environment where you have no time to think or search. As long as automation can't help (and it currently can't), docs are bound to memorization training.

Yes, but doctors often refer to us as "the IT guys", and also have no idea what they're talking about.

> you have no friggen clue what you're talking about

Please do better than this, regardless of how wrong or ignorant another comment is. Your post gives us no information other than about yourself.

If you know more, please share some of what you know so the rest of us can learn. Please don't just name-call. This is in the site guidelines: https://news.ycombinator.com/newsguidelines.html.

I winced at this comment too, because of that line, but have to point out that the "other side" of this argument could stand to be a lot more respectful as well. The "most doctors are dumb" sentiment on this thread fits the caricature lots of people have in their heads about how "HN people" think.

I don't really think it's a caricature.

"Doctors are idiots following algorithms and don't actually have any knowledge or information and just scam everyone for money all the time until their eventual replacement by the first person who thinks of feeding electronic medical records to machine learning algos" would cover about 95% of HN comments in these sorts of threads. You might cover another 2% with variants of "there's no point in going to a doctor for your problem, any decent engineer can just spend a few dozen hours reading up on their problem and figure it out just as well."

It's gotten to the point where I only chime in once in a while when the thread has gone deeply, deeply, off the deep end. But then, HN seems to have variants of that attitude towards pretty much every profession that doesn't have an engineering degree.

Sorry. You're right, dang, I could have done better.

I was not intending to name-call; to me, the entire comment was wrong and I posted a curt reaction.

I'll provide more info here:

"The level of legitimacy and respect that doctors and hospitals currently have is far too high..."

Ask any young doctor in the US whether their field has a lot of "respect". Medical residents work 80-100 hours per week in windowless, flourescently lit rooms, of which 40-60 hours is often spent doing clerical work at sub-minimum-wage to service private insurance and EMR overlords. Patients regularly tell these same doctors they don't trust them, don't respect them, think they're wrong, want to see the "real" doctor, can prove how wrong they are with WebMD, etc. My spouse is now doing her 13th and 14th year of post-secondary training toward the goal of being a specialized surgeon. She has done hundreds of successful surgeries/procedures already and saved countless lives. Yet, she is burnt out, and often ends her workdays completely destroyed by the humiliation of being a cog in the hospital machine. She is not alone; physician depression and suicide is at all-time highs.

"... and hinges on branding and the exclusivity of becoming a doctor."

Like I said, she is on her 13th and 14th year of post secondary training. Aside from this being a massive time investment, it also cost her dollars she didn't have (no support from parents) on order of $300k in post-secondary education, just in tuition. If becoming a doctor is "exclusive" and confers a "brand", I'd say the exclusivity is "extreme passion for helping humans at great personal sacrifice and cost", and the conferred "brand" is that you are allowed to practice medicine, at all.

"Once all of this data becomes available, it will be obvious how worthless and expensive most medical treatments really are."

Medicine -- and the tireless and ethical work of thousands of well-trained and professional doctors -- saves countless, countless lives every single day. Modern medical treatments are not "worthless" and your comment is honestly insulting to how effective and life-saving many treatments are. I think this part of your comment is what got the rise out of me: spreading this kind of misinformation is the closest thing HN commenting gets to "malpractice".

Society should respect and pay doctors much more for the sacrifice they put toward this profession. Patients should respect doctors a lot more for the benefit they provide society. Hospitals should give doctors a better work environment. Doctors are not perfect, nor is medicine. But to think the root cause of problems in the present medical system is an undeserving respect for doctors and modern medical treatments expresses a profound ignorance of reality.

The original comment is so wrong that it truly stunned me into an uncharacteristically curt ad hominem response, for which I apologize.

Much better. Thanks!

What this really makes me believe even more is that medicine as an institution is completely bankrupt. There are for sure selfless doctors who sacrifice for their patients. And medical care is fantastic for acute injuries and conditions.

But the system as a whole is very ineffectual for chronic diseases, and I would argue, outright harmed the public with its dietary recommendations.

Didn't a big Silicon Valley startup already get their hands on a bunch of money and data and try to disrupt the medical industry? Can't quite remember... Themanos? Therablos? Wonder whatever happened to them...

Reply on news.ycombinator.com