Hacker News new | past | comments | ask | show | jobs | submit | more wangyjx's comments login

This was a head-on blow for MS?


no matter MS is good enough or not, this is a good thing


A stupid accusation. The current situation is by 2017, a successful new drug costs 1.5B $ and 14 years in average, at the same time, more than 95% drug researches fail at different stages. In addition, the patent period counts far before the clinical trials ( begin from the molecular compound finding), usually the patent will only protect its drug less than 10 years. The drugmakers have to earn enough money in less than 10 years to cover all the cost spending on its own research and other failed researchs.

Yes, they can make enough profit. Without that, what else can support them to continue research new drugs? by you?


I'm not sure where that $1.5B figure came from but the caveat I've always heard that came along with it is that drug companies pay less than 10% of the actual research costs. The rest is made up from government grants and university research. They mention the total costs to make themselves sound like victims.


That is simply not true. The $1.5B is cost spent to get a single drug approved and includes cost of failure, i.e. Spend on drugs that end up not getting approved. The data that supports this comes from a sampling of pharma company r&d spend

The NIH funds about $30B in research each year, NSF about $7B. The top 15 pharma companies spend ~$70B on r&d a year. VCs invest $10-15B / year as of the last few years. Globally all pharma companies spend over $200B on r&d


Is it known why that costs so much and that the cost isn't artificially inflated to justify the patents or other reasons?


This is the seminal and oft-cited source of those estimates for cost of drug development. The data comes from internal financial records at pharma companies. This is a 2010 paper and costs have since increased:

https://www.nature.com/articles/nrd3078

Bottom line is that it is expensive to develop drugs because most of them fail. You need to spend millions of dollars on dozens of drugs that don't work in order to get one drug approved.

The single biggest cost driver is cost of Phase 2 failure. Basically when a drug that works in animals doesn't work in people. It costs $40-50M+ and takes 5-7 years to get through phase 2 and over 60% of phase 2 studies fail

This is because biology is hard. So the fundamental reason drugs are expensive is because biology is hard

You can also look at high level financial metrics compared to new drug approvals to get an orthogonal way of assessing r&d productivity. The top 15 pharma companies spend a combined $70B on r&d per year. In 2017 only 49 new drugs were approved and only like 40% of those were developed by big pharma (the rest were developed by startups that were acquired by big pharma, or developed by mid size pharma). Even if all 49 of those drugs go on to make $1B a year, that will still not return a profit to even just the top 15 pharma companies

Also most of those drugs are developed for cancer or rare disease bc the economics of developing drugs for those conditions actually work. For things like cardiovascular disease, stroke, or psychiatric disease, where the economics don't work out, you see very few new drugs despite massive need for better treatments. This has nothing to do with patents and everything to do with the cost of r&d. If pharma companies could develop a new heart drug for cheap and get some lame patent protection to give them a monopoly they would do it


The linked article is behind the paywall. From you comment I still can't take out _why_ this costs so much. It sounds like development cost X because phase of development cost Y. I find it strange that you can say that something called "phase 2" costs precisely $40-50M+. What does it involve that it cannot cost $1M or $100k? What's so hard about biology that makes it expensive? What I am seeing is just words but no substance or maybe I don't understand something? I would love get more insight.


I wrote a blog post on the drug development basics that contains some of the charts from that paper

https://www.baybridgebio.com/blog/drug_dev_process.html

It actually costs an average $40M (as of 2010) to do a phase 2 study. That's an average, not a precise cost. The cost entails designing a study, getting regulatory approval, recruiting hundreds of patients, paying for multiple doctor visits for each patient, manufacturing the drug, process engineering, administrative costs, getting hospitals and healthcare facilities onboard to treat patients, lab tests for all the patients, development of custom biological assays to measure the drugs activity, payments to lawyers and consultants, plus paying dozens of employees for a few years

There is an additional $30-40M needed just to get to phase 2 plus the $40m or so for the phase 2 study.

The why is biology hard question is beyond the scope of this post, and I'm not a biologist (though i work with them). But I think a lot of it has to do with the fact that biologists can't directly study human biology -- you can't just delete a gene in people and see what happens, or stick a bunch of random chemicals in somebody. Even if it wasn't ethically, financially and logistically challenging to do that, you couldn't always even see what you want to see. You can't watch changes in live human organs at the level of watching specific molecules, you can't see where all the molecules go and how they impact all other systems in the body in real time. You just have snapshots of specific things, like analysis of blood samples, or imperfect models, like animal or cell models that seek to replicate specific biological phenomena

Contrast that with something like computer science, where you can directly play around with the thing you want to study -- computers. Or chemistry -- you can be in a lab and work directly with the chemical matter you want to study. Not that those are easy, but you at least can poke around directly at the stuff you want to learn about


Clinical trials up to FDA standards are expensive.



I would love to see a break-down of those 1.5B. How can one company even organize the spending of that kind of money?


You can go to the SEC EDGAR website and search for 10-K filings for large pharma companies -- Roche, Novartis, Pfizer, Merck etc. they summarize their financials including r&d spend. Most big pharma companies spend $5-12B / year on r&d and it can take 10 years to get approval


Seems like we're running into incrementally diminishing return with the current approach in drug research.


They'll spend more than $1.5B on advertising for that wonder drug.


you 4-6 years to recoup a couple billion in cost and have to make a profit. how do you do that without any marketing and sales?


nice


microsoft not forbids manufacturers to install OS other than windows, that's why you can buy computers with linux. There are always options, there is nothing to blame microsoft for this. If you don't want to buy Windows, you can choose a model without windows. If you cannot find a satisfied model without windows, go to blame your PC manufacturer


Microsoft used to threaten PC manufacturers that bundled any other OS other than Windows or no OS at all. They claimed that not bundling an OS was the express purpose of installing a pirated copy of Windows.


But lots of Office-like apps have implmented OpenXML read/write without MS libraries


Why MS cares your private repositories? give a reason? Maybe using your code to train their programming robot, lol


This is a much more concrete possibility than people might realize: https://www.microsoft.com/en-us/research/blog/deep-learning-...


public repositories contain enough data to learn, private repositories only a small part


Consider applying for YC's Summer 2025 batch! Applications are open till May 13

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: