Hacker News Reader

Comment by ThomPete | Hacker News Reader

No matter how this story ends (whether they eventually ends up with having a fully functioning product or not) this should be a cautionary tale to anyone wanting to actually change the world in healthcare.

Science speaks for itself and does not care how much you raised at what valuation. It requires a rigorous approach to testings and facts and data or it will be merciless if you aren't living up to your own claims.

So hopefully next time someone has a great disruptive idea in healthcare, they will spend a little more time getting the science and tech right before they start worrying about becoming a unicorn.

View on HN

"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled" - Richard Feynman

how about reality distortion fields?

It is a common, romantic view of how science works - but it is only compatible with reality on a 30-50 year timescale, and probably only because of survivor bias.

The science of Dan Shchetman, J. Robin Warren, Barbara McClintock and many others spoke for itself, and for a long time "science" essentially didn't listen. I put "science" in quotes because "science" doesn't speak or listen - scientists do - and the scientists shunned those luminaries (eventually to recognize them with one of the highest honours - the Nobel Prize).

Warren had to use himself as a test subject for others to notice. Shechtman was a cast out for years (and then it turned out that he wasn't the first to notice 5-fold symmetry - he was just the only one for a long time to insist it's real; the other "scientists" who noticed it cared too much about their status, rather than the science). McClintock was shunned by her peers and essentially lost her job -- though evidence for how badly she was treated by "science" for her radical ideas (now mainstream and assumed true) is being whitewashed.

I keep wondering how much progress we have lost because "scientists" (unlike "science") does care how much you've raised, how much you've published, and who your friends are.

> anyone wanting to actually change the world in healthcare ... Science speaks for itself

The belief that modern healthcare is nothing but the application of science is also romantic and misguided. In the US, healthcare is a business and not even a free-market one. It is informed by science, most definitely. But e.g. no doctor will recommend the single most effective treatment for type 2 diabetes early in the disease[0] (calorie restriction whether directly, by direct fasting or intermittent fasting). In fact, it turns out that the American Diabetic Association recommendations are the everything but science, and possibly the worst you could do.

[0] - http://www.theguardian.com/lifeandstyle/2013/may/12/type-2-d..., see http://www.ncl.ac.uk/magres/research/diabetes/reversal.htm for references

I agree that good science often doesn't get recognized because it happens to rub some established scientists the wrong way.

However, the issue with Theranos is the opposite: it is getting recognized, even though we don't know whether it's good science or not, because there isn't any visibility into what they do and how they do it. The scientists you mention all published their work for the world to see; they didn't hide behind "trade secrets".

> The belief that modern healthcare is nothing but the application of science is also romantic and misguided.

I agree with this too; but again I don't see how it applies to Theranos, because they are explicitly claming that what they are doing is science. So they should be judged by scientific standards, even if the rest of the health care industry can't live up to those standards.

I was not remarking about Theranos, but rather to ThomPete's statement.

Rigorous science is not sufficient (e.g. the Shechtmans/Warrens that we don't know about -- it is pure hubris to believe that all those were eventually vindicated), nor is it required (e.g. Vioxx, Thalidomide, ADA recommendations, baby peanut allergy recommendation, dietary cholesterol recommendation, dietary sodium recommendation).

If Theranos is a lesson to entrepreneurial scientists it is probably: "Don't waste your time failing quietly and cheaply, when you can fail extravagantly and retire rich"

They might have also learned that lesson with Solyndra in 2011.

Not sure what you mean.

Their product is scientific. It's not just a health app or some nutriment recommendation which can be claimed successful without any scientific claims to back it.

Theranos has a very a scientific product with a binary output and so it's claims will stand up to that (i.e. being approved by passing a number of tests)

My lesson was for those who wanted to change the world via healthcare, not those who wanted to get rich.

> Their product is scientific. It's not just a health app or some nutriment recommendation which can be claimed successful without any scientific claims to back it.

So was VioXX[0]. Read the Withdrawal section, especially the last paragraph. As someone who was following the case closely while it was happening, I can tell you that everyone involved looked guilty as hell, even though the court found only the marketing people "overzealous".

You see, "science" doesn't lie. But scientists do. Some liars are better than others, though.

Also, the standard US medical advice about things like nuts that might cause allergy are "delay introduction as much as possible" (e.g., do not introduce peanuts before age 3 or so, for fear of allergic reaction). You will hear this from doctors and find this in official manuals, although there is no science to support this (never was), and in fact, there's data indicating that early exposure to nuts reduces allergy.

> Theranos has a very a scientific product with a binary output and so it's claims will stand up to that (i.e. being approved by passing a number of tests)

Output is not binary. Most (often all) data used to evaluate medical products is provided exclusively by manufacturer, who is able to massage data to yield specific desired results, see e.g. VioXX.

> My lesson was for those who wanted to change the world via healthcare, not those who wanted to get rich.

Well then, they shouldn't look to Theranos - those guys obviously care more about getting rich than about changing the world (well, changing the world for the global good, anyway).

[0] https://en.wikipedia.org/wiki/Rofecoxib

I recognize your point in showing scientists whose data were initially discounted. The Theranos case is different, however, as they have not attempted to publish their methods. If they had tried and been rejected, it might begin to be fair to compare them to the distinguished researchers you mention.

> no doctor will recommend the single most effective treatment for type 2 diabetes early in the disease

How do you know this is true? Dietary counselling is recommended as part of the treatment of newly-diagnosed T2DM, per UpToDate[1], a reference for this kind of thing.

> American Diabetic Association recommendations are the everything but science, and possibly the worst you could do.

I'm not sure what you mean here. What are the guidelines and how are they deleterious?

1. http://www.uptodate.com/contents/nutritional-considerations-...

> If they had tried and been rejected, it might begin to be fair to compare them to the distinguished researchers you mention.

I might have been misunderstood. I am not defending Theranos or claiming they deserve any accolade or even sympathy. I am just pointing out that the romantic view that "science triumphs" is not, historically, supported by evidence on time scales shorter than a few decades. As most people's careers -- scientists included -- are on that time scale, it is a bad idea to assume that "science triumph" for any specific person or company.

> How do you know this is true?

I've been following these issues for a while. To borrow Gary Taub's analogy - the guides ("eat this much or that much, mostly carbs, low fat, moderate protein") are similar to an advice about how to make a nightclub popular: "get more people in". Yes, if you manage to do that consistently and with sustainable effort, then everything will be fine - yet the compliance rates are ridiculously low (as your link points out). But now we have the people, rather than doctors, to blame for non-compliant behavior, so the ADA and medical institution is not to blame.

However, there are studies (no time to look for refs right now, sorry), showing that people on low-carb, high-fat, medium protein (i.e. "keto" style) have a much, much higher compliance rate, and much better outcomes (in term of controlling blood sugar and reversing type 2) -- yet the recommended ADA (IIRC, can't find it right now) is 60% carb, 30% protein, 10% fat. That's high fat, mind you - not high protein.

Science and results don't care about secrecy and PR either. As far as I can tell, the only purpose of this company is to provide the media with a steady stream of glamour shots of the CEO so they can fawn over her and promote the never ending "pretty Stanford dropout founds a company and is like Steve Jobs" narrative.

Well that's all very moral, but what Theranos is doing is mostly engineering. So while "hype the fuck out of it, raise big money, and fake it 'till you make it" might feel all icky, it's not a unique plan.

Will we be saying the same things about Palantir? Maybe, maybe not. There's not that much difference between Palantir and Theranos. What has Palantir got that others don't? Answer: The ability to raise money at incredible valuations. Both are living off a mythologized ("PayPal has great fraud detection" and "you don't need an incompetent phlebotomist poking holes in your arm") advantage. We'll see if they can turn all that money into a real advantage.

But you can't count on science to spank these people.

That's what I thought at first, but $400M and 12 years should have been enough to put together a pretty solid uTAS seeing as microfluidic integration wasn't exactly bleeding edge back then either. But they seem to have drunk their own kool-aid and created a working environment so toxic that they sabotaged their own chances to fake it till they made it. They might still get there, but even if they do there's a lesson to be had about expectation management.

EDIT: to clarify, I'm not the source of downvotes. Before the recent media explosion I believed exactly as you did.

Somehow the ability to spend money scales pretty well relative to the ability to raise it.

Anyone working in microfluidics would've been skeptical at the time and is probably not surprised now. Microfluidics is amazingly hard to make work properly, and the history of the field is littered with great ideas which proved to be much harder to execute reliably.

Palantir is not making binary claims which can be proven right or wrong. Palantir is making tools which can be used and interpreted different ways. They are much more like vitamins.

This is a very important point in favor of Palantir. While there is much to be said for the placebo effect in medicine, Palantir's customers can more readily convince themselves Palantir works. On the other hand, you've either got herpes or you don't.

Agreed! And confidence in theories is needed to persevere to get true evidence. But if the evidence disproves the theory - adapt.

To ignore this is hubris.

Should Bill Gates have spent a little more time actually writing an operating system before selling one to IBM?

Fake it 'till you make it is often an important part of success.

I think the point is you can't fake it till you make it in medical science, and if you can, the outcomes usually aren't great.

But he did license them an OS that like worked, not a promissory note that may be like near Star Trek technology.

Reply on news.ycombinator.com