Comment by kdkirsch | Hacker News Reader

Comment by kdkirsch | Hacker News Reader

I’m a physician. I really appreciate Tom’s enthusiasm, but I think that enthusiasm led him to overestimate the significance of the problem. Moreover he didn’t really understand for whom he was building this product. At first he thought it would be a consumer product. But he realized that consumers would not pay/subscribe and later learned that advertising wouldn’t work. Next he thought doctors would pay because “Doctors have money, right?” Again paying for this or any service is an expense. Its value has to justify the expense. We pay hundreds of dollars a year for UpToDate because it’s valuable. Cochrane is the gold standard for meta-analyses and they’re publishing for topics that are of clinical significance. Ultimately I doubt anyone is going to change their practice based on this product because there isn’t a compelling reason to do so. My major observation is that Tom tried to make a healthcare startup with negligible understanding of healthcare: the players, how payments work, how physicians practice, and what patients want/need. By not understanding the environment you’re not going to be able to understand the Problem which means your Solution will probably fail. This is a mistake repeated by most of the engineer founded health startups I’ve read about. Finally for anyone wondering I usually recommend ibuprofen 800mg every 8 hours and Tylenol 1000mg every 8 hours. This isn’t medical advice, just something you may find from a quick search.

If the COVID-19 pandemic has taught me anything about science, it's that if you do a "meta-analysis" without reading each paper carefully and critically, you can end up proving anything regardless of its veracity. You cannot replace domain-expert scientists spending huge amounts of time painstakingly going over every detail of many papers to weed out mistakes and fraud in order to write a meta-analysis, with a computer program. Well, at least not until we figure out AGI. Until then, it would be irresponsible to rely on such a program for any clinical decisions.

https://news.ycombinator.com/item?id=29249686

https://astralcodexten.substack.com/p/ivermectin-much-more-t...

https://ivmmeta.com/ (Note the long list of things in the right sidebar which the "meta-analysis" shows have huge positive effects on COVID treatment)

Yes, this is a perfect example. ivmmeta is shocking, born of either complete bad faith or madness.

ok, but... would a meta meta-analysis analysis show that on average, meta analyses do find positive benefits? get back to me.

"TOO META"

https://xkcd.com/1447/

I think the generalised lesson to be learned is that in reducing complexity (and all modelling requires reducing complexity), you get to make decisions about which bits of data are cut out, and this is a decision that can introduce bias. We can't just blindly trust reductions, especially in areas where there are competing interests. This is why transparency is so important! If I can reproduce your methodology, I can critique it for bias.

I agree. And what Tom didn't realize is that Doctors aren't financially incetivized to give better healthcare. There is no money in it for them. If anything, return visits because of a partial prior resolution make more money.

Now clearly I'm not saying that doctors try to give bad diagnoses for profit (because they don't). But beyond a certain minimum bar, and on a purely financial basis, improving diagnostics or prescription accuracy doesn't not make a medical practice more money. And the responses here illustrte that.

Now, whether that's the right/wrong incentive strucgture is a whole 'nother discussion. And I personally can't think of something better than what we currently have - but it is a truth of the current system.

He actually received a clue from 'Susan', one of the doctors he demonstrated his application to: "in many cases I’ll just prescribe what I normally do, since I’m comfortable with it"

He should have sold his tool to sales and marketing at pharmaceutical companies who would use it to convince doctors to prescribe their product.

But if you're selling a drug, aren't you going to rely on just the studies that show the efficacy of your product? If your drug _isn't_ the best according to the meta-analyses, this tool is of no help to you. And even if your drug _is_ the best according to the meta-analyses, that's because there are already multiple studies that show how good your drug is, and you can overstate your product's value by throwing out the rest of the studies.

Agreed, at that point the value proposition is to some trying to close a sale to overcome the objection that the supplied studies are cherry picked.

Its... uh.... niche.

I see this take all the time but it’s kind of misleading. Realistically things like uptodate, clinical journals and conferences mint money. Doctors are one of the prime demographics of people who do things for non monetary gains. Yes if you want to be nihilist you’ll find a minority of docs where this isn’t true, but the point is there’sa huge market to doctors willing to improve care without direct monetary gain.

This startup product failed because it wasn’t shown to improve outcomes— it just compiled study results by a non expert. At the very least the founder should have at least attempted doing a study on the tool showing using it resulted in better outcomes. But it didn’t make that link and so people didn’t know what to do with it.

The financial incentives behind care differ greatly between states and health systems, so you can't just make blanket statements like that. Look up accountable care organizations and value-based care. We're still in our early stages, but our country is trying to figure out how to properly incentivize care, I hope.

> Doctors aren't financially incetivized to give better healthcare.

I think that's an overly nihilistic and not entirely accurate viewpoint. Good doctors get more patients through referrals, and once they have enough patients they can be more selective about which patients they take on. And they make more money. Bad doctors will lose trust and patients.

What doctors aren't incentivized to do is provide marginally stronger drugs. If aleve works a little bit better but tylenol is easier on your stomach, then is Aleve really "better"? If you actually need a stronger painkiller you can just up the dose. Medicine is rarely a black and white "best" situation.

your comment makes sense from a layman's perspective of logic and economics. but in the industry you'll find most doctors actually do not want to see patients for revisits and would rather a pill or injection solve the problem in one go. unfortunately due to biology this is rarely the case, and managers/pharma reps/business dev are keen to capitalize on it- not the overworked doctors.

Let me rewrite your comment:

Doctors are incentivised to provide better healthcare. The incentives are not directly financial.

That's sad.... that you think that.

It is like saying that software engineers write bad code, so that they can get paid to fix it. (Which is really not true at all)

This was the whole thing with bundled payments. https://www.cms.gov/newsroom/fact-sheets/bundled-payments-ca.... One lump sum for a condition, and the stakeholders involved want to treat it the fastest/cheapest way possible.

Barring that aside, at least the doctors I'm related to would love nothing better than to give better care. I always hear about how they need to stay up on the latest stuff.

> But beyond a certain minimum bar, and on a purely financial basis, improving diagnostics or prescription accuracy doesn't not make a medical practice more money.

But giving better diagnostics and better prescriptions can result in improved profits if it's marketed in an efficient way. Would you go to a clinic that will give you the best diagnostic and prescription possible or to one where you are unsure about the result?

Some mediocre restaurants in tourist areas are doing well because they are vouched for and have TripAdvisor stickers on their front door. What if clinics had GlacierMD stickers on their front door? To sell GlacierMD to doctors you have first to convince the patients that is something great about it.

> If anything, return visits because of a partial prior resolution make more money.

I don't know what system you are referring to, but in the EU, for common physician services, this is (fortunately) not how it works. Physicians are by and large being paid per case, not per visit. In Germany a case is being defined as the same patient visiting the same physician in the same quarter of the year (https://www.kbv.de/tools/ebm/html/3.1_1623969609994938562151...)

I'm assuming the US as that is the case here. Each visit is billed for individually. By the case is certainly a much better model.

> If anything, return visits because of a partial prior resolution make more money.

I don't think any doctor deliberately prolongs the illness of their patients to keep the money flow coming.

Do you have any evidence to back up what you said (other than a flawed game-theory hypothesis)?

Yeah I adopted my confused face from the 'problem' statement/'killer' start-up pitch on. Have a headache and don't know which drug to buy? Come on. You have a headache and you take either acetaminophen or ibuprofen, both of which you already have. You don't buy anything.

I'm not a physician, I just take (rarely anything other than max dose) them per the label, which is different to your suggestion (1g/4h, max 4g/day) at least where I am. Ibuprofen if something seems 'inflamationy' or if I have drunk/will drink alcohol (on some sort of naïve hand-wavy basis to lighten the livery load).

I’m not a doctor but I watched a friend’s startup have the exact same experience. My personal experience was an early job selling online advertising to all categories of businesses. I quickly learned that medical offices are absolutely inundated with salespeople like pharma reps, and doctors time is already scant. Naturally they tend to want to see evidence in a way most other SMB decision makers do not. Since at time, bundled payments have gone into effect and consolidation means that the purchasing process is exponentially more complex.

Startups: don’t sell to medical offices without deep domain experience, deeper pockets, and rigorous evidence.

> By not understanding the environment you’re not going to be able to understand the Problem which means your Solution will probably fail.

Yes it boils down to politics. Having built a platform for accelerating routine histocompatibility analysis (providing customers a profit from the start) I got to see this firsthand. Risk aversion across the industry, small armies of established technicians working manually, administrators making purchasing decisions with no understanding of the science or impact, the myopic insular bent of academics, the massive gap between healthcare and modern technology. One could pay labs to save time, improve outcomes while reducing risk and they still wouldn't use your product if it doesn't map onto their political priorities.

I think this startup would have more of a chance if it was able to solve the Google/trust problem with medical advice.

From my perspective as an ignorant consumer, Googling medical things always leaves me unsatisfied. I get too much mixed advice when I poll multiple websites and my final feeling is that I've learned no new information, there was no clear information winner. Part of that is the SEO spam problem, part of it is not knowing accurate Google words for medical stuff.

I am developing a non invasive diagnosis tool that uses infrared absorption, electrical activity and a few other sensors with machine learning. As it stands there are some physicians who think it has potential. I am also using it to identify people which would make it useful in its own right. My question is, do you think a quick diagnosis tool would be helpful and would someone be prepared to pay something for it?

I’ve been a PM in a small healthcare tech company run by engineers over the past few years and this 100%.

Together? That sounds a lot..

Reply on news.ycombinator.com