Hacker News Reader

Comment by dharmapure | Hacker News Reader

In the book Antifragile Nassim Taleb cautions against medical care unless it's truly necessary. He argues that since so many things can go wrong in a hospital (you could get sick from someone else, mistakes could be made, etc), only go if the benefits outweigh the risks.

And so if you have cancer, it's probably worth the risk. But if you have something that could just go away naturally, that's got a higher expected value of positive outcome.

He is in favor of subtractive medicine - many harms are caused by adding such things as smoking, sugar, preservatives, etc, and instead of adding medicine or surgery, remove the elements that the body isn't evolved to tolerate and see if it heals itself.

He gives the example of how he hurt his back weight lifting. The doctor proposed surgery. He just waited and his back healed itself, and in a way that is now more robust than if he had had surgery.

As with all things, it's not a black and white kind of decision, but this mindset has worked well for me as a heuristic for health.

View on HN

Nassim Taleb is great at providing backward looking advice. My back got better, therefore I was right to ignore the expert advice. He has predicted about ten of the last two financial crises. I wouldn't take medical advice from a financial pundit who can't seem to go more than a day without picking a fight on twitter.

My wife professionally interacts with a lot of top level surgeons -- the ones that have institutes named after or created for. They universally say it is all about risk management and risk mitigation. Trauma and neuro and cardiac surgeons of today are likely to be able to fix amazing number of issues just to have patients become gravely ill from secondary infections/non-top-notch care after the surgeries.

That's the frustrating thing though. Doctors tend not to talk in probabilities of outcomes. Fortunately I haven't had a ton of experience trying to get them to do so.

That's because most Americans have no idea how probability or risk mitigation work. It's their body, dammit, and they want you to fix it!

I dislike Taleb, but in this case he's right. Conventional, standard, advice in anywhere but the US is that over-testing, over-diagnosis, and over-treatment cause harm and should be avoided.

The problem is that Taleb is packaging up this standard advice and pushing out as some wisdom he has that he wants to impart to the masses.

My blog making fun of Taleb used to have higher page rank than his personal website. However, Taleb is totally correct about this, and his Antifragile book is very good (though he still needs an editor).

Most of modern medical science is quackery and pharma companies trying to milk the populace for profit. Trauma surgery is quite good, and antibiotics and vaccines are important. Other than that, I'm pretty sure going to the gym and the salad section of the grocery is much more important than going to the doctor.

He may be right in this case, I am not a doctor and don't know the details of his anecdote. My point is that if I'm looking for medical advice from a pop science author I'd much prefer Dr. Gawande's evidence and experience based account to some half baked story about convexity.

I'm not even sure whether Taleb's style of writing is meant to communicate much at all, what's the point of all this phony formalism? http://www.fooledbyrandomness.com/medconvex

I'd like to see Taleb acknowledge at least that medical and financial estimation risks have a different character. Financial mispricing is adversarial, whereas you'd have to be more cynical than me to think doctors are always trying to get you to take the maximum care they can sell you.

> can't seem to go more than a day without picking a fight on twitter

Thanks for that. I was blocked for correcting him (I can't remember about what, it was years ago).

Also, this "only go if necessary" is bad advice unless you are a doctor and can tell, from your symptoms, whether it is or not necessary to go. Make an appointment or show up at the ER and let the doctor decide whether you need help or not.

The doctor proposed surgery. He just waited and his back healed itself

Isn't that when you should actually go talk to the surgeon?

I've had two occasions when a doctor proposed surgery, I went to the surgeon, and the surgeon said "Let's give it 6 month and see if the gets better", which it did in both cases. And one case where I ignored the doctor, lived in growing discomfort for a couple of years, ended up in the emergency room for emergency surgery and the surgeon going "couldn't you tell something was obviously wrong with you? Why didn't you come a year ago when this would have been trivial fix?"

I have a couple of dislodged discs, which cause the right arm to go numb sometimes.

My GP referred me to one of the top surgeons for spin surgery, especially at the top of the spine, where cutting it up is the most risky.

After prescribing an MRI and given that the numbness is occasional and that the pain is minimal his suggestion for cutting me up was a clear NO!

This gives me a lot of respect and instills trust in an expert who provides me with advise, which runs counter towards his business model.

I hope that the situation clears itself with occasional physical therapy and exercise. But if it gets worse it's fantastic to have one of the foremost experts in the field, who can perform the surgery at your disposal (and adequate insurance to pay for him).

To make a long story short: It seems that my doc very much agrees with Nassim Taleb.

> if you have terminal cancer

Maybe you meant "if you have treatable cancer"

If it's truly terminal, not much a doctor can do for you other than palliative care, which generally shouldn't require a hospital.

Good catch. I've made an edit.

That's not really true. Having just been through this with my mother there is a huge amount of hospital care required during terminal cancer.

For starters radiotherapy is often used to control symptoms. It may not do anything for length of life but targeting key tumours can keep you mobile for longer, reduce the risk of seizures etc. It can be a boon to quality of life.

There will also almost certainly be other things that occur along the way that require visits to hospital. That is complications that could kill you but can be treated without too much discomfort where you can get back onto your feet. Infections are incredibly common for example. Particularly if you're on steroids.

Obviously you'll rule some things out, e.g. chemo as not worth the cost during palliative care.

>For starters radiotherapy is often used to control symptoms.

I believe this falls under palliative care, if it is only to control symptoms.

But it definitely requires a hospital.

He's just regurgitating conventional medical wisdom that's routine in countries that aren't the US.

We've been talking about the risks of over-testing, over-diagnosis, and over-treatment for many years. Mostly in the context of full body MRI scans for healthy people, or early testing for prostate cancer, but also many other things.

What's the problem with full body MRI scans for healthy people? They don't cause any harm, just cost money.

Apart from the fact that it costs a lot of money. What do you want to achieve with a full body MRI scan?

A friend of mine who wasn't even 30 years old had a serious brain tumour where they said "ouh, if we'd had done a MRI scan some years ago, it would have been much easier to fix."

So that is why it might be a okay'ish idea to do a MRI scan of the brain every 5 years or so? Depending where you live they are actually cheap and can be paid out of pocket.

Fine and fair enough.

But that scan would be limited to the head and not cover the entire body.

I'm also not convinced that MRIs are a great, general purpose diagnostic tool, if it's not used to diagnose concrete issues. The experience of your friend, who I hope recovered from the tumor not withstanding.

With those caveats in place, absolutely.

In the US, studies have been shown that physicians who own shares in CT (not so much any more, as it becomes almost 'commodity') and MRI (which is a lot, manufacturers approach physicians to form imaging cooperatives) are ordering imaging up at more than 2 standard deviations from expectations.

As noted, MRIs in themselves aren't harmful. But with prices in the four digits in the US... you can easily pay off your $1.5M MRI machine in 18 months and be printing money with your over-imaging.

MRIs should be safe. It's the unnecessary CT scans you want to avoid. Those blast you with x-rays and increase your risk of cancer by a non-negligible amount.

MRIs on healthy, asymptomatic individuals are not "safe," they cause harm due to overdiagnosis and overtreatment. I know TFA was a long read but it explains it, CTRL-F "turtle" to get to the part about tumors specifically. Here's a similar explanation: https://sciencebasedmedicine.org/a-skeptical-look-at-screeni...

Agreed - I was simply stating that the act of getting an MRI does not cause harm, whereas some other tests (CT scans, X-rays, etc.) actually cause harm during the course of the test. I certainly would not recommend healthy people to get MRIs. But if something is really messed up in your body, an MRI is one of the safer imaging techniques (albeit expensive and not always the best for the job).

Not really. The absolute worst estimates for CT scans increasing your risk of fatal cancer is around 1 in every 2,000 scans. That statistic almost certainly isn't true, since we'd see skyrocketing cancer rates since the development of CT scans, which we haven't. But even assuming it was true, that's pretty close to a "negligible" risk. 1 in 5 people will die from cancer. That's 400 out of 2000. Getting a CT scan makes your risk 401 out of 2000 in the worst possible scenario. If anything, that's an argument for more CT scans (as long as they're necessary or preventative), especially if you're at risk for certain types of cancer and could detect it early with preventative scans. MRIs would be better in that case, but you aren't going to die from a CT.

They cause harm due to overdiagnosis and over treatment.

https://sciencebasedmedicine.org/a-skeptical-look-at-screeni...

Scans don’t cause overdiagnosis. Eager checklist-driven doctors cause overdiagnosis. You will see incidentalomas, how you use that information is not preordained.

Scans directly cause overdiagnosis.

It's very non-intuitive but scans and tests have the potential to cause more harm than good.

From TFA:

>Overtesting has also created a new, unanticipated problem: overdiagnosis. This isn’t misdiagnosis—the erroneous diagnosis of a disease. This is the correct diagnosis of a disease that is never going to bother you in your lifetime. We’ve long assumed that if we screen a healthy population for diseases like cancer or coronary-artery disease, and catch those diseases early, we’ll be able to treat them before they get dangerously advanced, and save lives in large numbers. But it hasn’t turned out that way. For instance, cancer screening with mammography, ultrasound, and blood testing has dramatically increased the detection of breast, thyroid, and prostate cancer during the past quarter century. We’re treating hundreds of thousands more people each year for these diseases than we ever have. Yet only a tiny reduction in death, if any, has resulted.

>H. Gilbert Welch, a Dartmouth Medical School professor, is an expert on overdiagnosis, and in his excellent new book, “Less Medicine, More Health,” he explains the phenomenon this way: we’ve assumed, he says, that cancers are all like rabbits that you want to catch before they escape the barnyard pen. But some are more like birds—the most aggressive cancers have already taken flight before you can discover them, which is why some people still die from cancer, despite early detection. And lots are more like turtles. They aren’t going anywhere. Removing them won’t make any difference.

>Over the past two decades, we’ve tripled the number of thyroid cancers we detect and remove in the United States, but we haven’t reduced the death rate at all. In South Korea, widespread ultrasound screening has led to a fifteen-fold increase in detection of small thyroid cancers. Thyroid cancer is now the No. 1 cancer diagnosed and treated in that country. But, as Welch points out, the death rate hasn’t dropped one iota there, either. (Meanwhile, the number of people with permanent complications from thyroid surgery has skyrocketed.) It’s all over-diagnosis. We’re just catching turtles.

Solution -> never show a scan to a doctor who will over-react to an incidentaloma

It's not "overreacting."

If you do a scan in someone asymptomatic and see a tumor it's oftentimes completely impossible to know if it's a "turtle" or a "bird." They can look exactly the same and we don't have a crystal ball.

Just wait and see how it develops. My doctor agrees, you could check with yours.

By your argument, if there's no need to scan, there's no need to treat it right away either. Of course, if you are the sort of person who is prone to panic, you probably don't want to get a scan

That doctor needs to be somehow shielded from lawsuits, or they'll be incentivised to give undue weight to the risk of leaving something real untreated.

Good point. Of course, the overdiagnosis is then caused by the legal environment (not the scan). As an individual of course you can always sign a release for your doctor. If he seems uneasy about nontreatment, just write a letter to release him from liability for nontreatment so he knows he can relax and actually consider your wellbeing rather than liability.

Interesting. But you also don't want to bias them toward nontreatment (do you?). I wonder if one would get better outcomes in lawsuit-heavy countries by just starting all interactions with health professionals by handing out unsolicited full liability releases like candy. "I want you to do what you think can work, not what is easy to justify if it fails." Maybe only if one has a particularly difficult health condition.

Shouldn't a distinction be made between asserting this for a population vs. for an individual? Yes, for a population it results in overdiagnosis and over treatment. As a personal decision, though, I understand this and will choose how to proceed with that in mind.

>Shouldn't a distinction be made between asserting this for a population vs. for an individual? Yes, for a population it results in overdiagnosis and over treatment.

We get overdiagnosis and overtreatment of the population precisely because it is done at an individual level.

Right, micromotives. But what's best for me, personally, may be different than what is best for a whole population.

Medicine is individualized. If you have certain risk factors you may benefit from a particular screening tests or even prophylactic treatment and procedures. These should be evidence based too, however. Same when you're having symptoms.

The topic at hand is, as stated up-thread, "full body MRI scans for healthy people."

It seems possible to do better than average here? Just from reading this article, we learn about some "cancer" that is less scary than it sounds and unlikely to cause a problem.

I would be more skeptical if there were a study showing overtreatment still happens in a population informed and wary of overtreatment.

It's bad for the individual too. Many checks have non-trivial rates of false positives, many treatments have non-trivial rates of severe complication.

Checking occasionally for something specific that's reasonably suspected due to symptoms is good. If you're always checking for everything, you have a high risk of finding something you don't actually have. It turns out that that risk overrides the beneficial chance of finding something real before it has any symptoms.

They do cause harm. People find all kinds of lumps and shadows when they get an MRI done. So now they have this thing, no-one knows what it is, and the patient has to decide whether to just live with it or have a biopsy to find out what it is. The biopsy carries a risk of harm. And then you get the results of the biopsy. But that's not always clear. Maybe they find cancer. So, is it a rapid cancer that will kill you if left untreated? Or is it a slow growing cancer that's just going to sit there not doing much until you're 93? The treatment can be harsh and cause harm.

Here's some articles. (Note how old these are. This isn't new advice).

http://www.bbc.co.uk/programmes/articles/15g8LGyBzYvMHpq6qNh...

> The first problem she uncovers is that detailed but general scans, such as a full body MRI or CT scan, can frequently bring to light small abnormalities which have no implications for health at all. They are the equivalent of tiny moles, freckles or scars on the outside of our body. Yet if they are picked up on a scan, then they can cause worry and further, much more invasive, testing – usually for no reason.

> The second problem that she uncovers is that even for some very specific tests, such as ultrasound of the arteries to assess the risk of stroke, the treatment available to rectify any small issues found carries very high risks to the patient. In the case of the ultrasound of the Carotid arteries in the neck, surgery to help clear any blockages would not be advised unless the patient had already had a stroke, or mini-stroke, making a test on an otherwise healthy person pointless.

http://www.independent.co.uk/life-style/health-and-families/...

> An American patient who had a whole body scan described how it threw up an early warning of lung cancer. He had a lung biopsy after a nodule was detected but it turned out to be a healed scar, posing no threat. However, the investigation involved four days in hospital, a painful procedure, several weeks of recuperation - and a $47,000 (£25,000) bill at the end.

http://news.bbc.co.uk/1/hi/health/7150583.stm

> "Full body MOTs" may often find benign abnormalities - so-called false positives - while missing real problems, said Professor Nicholas Wald, a specialist in preventive medicine.

The rate of false-positives is almost always non-zero. So having an MRI scan may show a problem that doesn't actually exist.

I think I remember seeing something about not doing mammogram and/or prostate screenings too often, lest they find a false-positive and do a whole bunch of followup tests and procedures that in turn keep widening the risk for complications.

  Mostly in the context of full body MRI scans for healthy people

Full body MRI scan? That seems completely weird.

I had a number of MRIs and believe it can be a god send. When I see the results it always feels like borderline magic and depending on the diagnosis it can be very, very useful.

It's very much my impression that MRIs are prescribed for very local issues. A bum knee, a broken toe which doesn't heal and where the doc needs a 3d view (to rule out splinters, which are not visible on x-rays) or for parts of the spine.

A full body scan doesn't only sound extremely invasive (how long do you have to lie absolutely still while the scan progesses?) it also sounds totally useless.

Edit: spello

People are _really_ bad at self diagnosing. So who's to say that the risks of going to the doctor are greater than the risks of self diagnosing?

Off the top of my head I know three family members who now suffer some disability because they didn't go to the doctor for symptoms they thought were non-threatening.

Are they though? In my lifetime the vast majority of my acute health issues went away on their own without any particular medical intervention. I mean well over 99.9%. I know what's "minor ache" feels like and I know what "this seems serious" feels like. I try to let my immune system take care illnesses when possible.

On the other hand I know someone who died from a complication of anesthesia during an (almost certainly) unnecessary surgery.

It's a multifaceted problem.

Patients (especially the kind that present to the doctors office for a cough) tend to want doctors to do something, they don't like hearing "rest up, you'll get better on your own." Doctors provide a service, patients want to feel like their time was worth it.

At the same time doctors don't want to put themselves in a position of missing something. Everyone remembers the time the doctor missed a serious disease but since we can't have alternative timeline where the treatment and/or tests weren't done then its very difficult to say, on an individual level, what was over treatment.

Sprinkle in a healthy dose of the placebo effect stir.

Judgement gets clouded by biases on all sides.

Not only all that, but sometimes the data just isn't there on what the best course of action is. The studies haven't been done so we just do what feels "right." And we've learned that what feels right might not lead to the best outcomes.

> In my lifetime the vast majority of my acute health issues went away on their own without any particular medical intervention. I mean well over 99.9%.

To be fair, if they didn't, you might not be here to talk about it, so there's definitely some survivorship bias there. The people who waited and died obviously can't say anything because they're dead.

In my experience, your average doctor is really bad at diagnosis too, they just have way more confidence in their diagnosis than the average person has in their own.

Doctors are _really_ bad at aligning the interests of their patient and their employer. So who's to say that the risks of going to the doctor are less than the risks of self diagnosing?

Off the top of my head I know two family members who were murdered because they did go to the doctor for symptoms that were non-threatening, and instead received an opioid addiction and premature death.

> But if you have something that could just go away naturally,

isn't that the case for most diseases?

After having a glance at the submission title, I directly CTRL-Fed, and wasn't disappointed to see the mention of stressor avoidance from the Lebanese philosopher and twitter sophist.

Great advice but how do you know ahead of time whether the medical care is necessary?

"Blind" second opinions. Go to another doctor - perhaps even at a different pay-grade - and see if she recommends the same treatment without you indicating to her anything about what the other doctor(s) said.

This is slightly easier for surgeries and anything non-acute where you have at least a month or two to make a decision.

>He is in favor of subtractive medicine - many harms are caused by adding such things as smoking, sugar, preservatives, etc, and instead of adding medicine or surgery, remove the elements that the body isn't evolved to tolerate and see if it heals itself.

Yes yes yes I agree! All those poor people living in poor and crime-ridden neighbourhoods should just shop at Whole Foods! Yes indeed!

Reply on news.ycombinator.com