Hacker News Reader

Comment by RcouF1uZ4gsC | Hacker News Reader

I know Derek Lowe is an expert. However, sometimes, experts think that because things have always been a certain way throughout their career, they will always be like that. Drug development has been very conservative especially in the US for the past several decades. The FDA is all about making sure they never have a thalidomide incident and would rather lose a lot of good drugs than approve a bad drug. This is the steady state.

However, a global pandemic is more akin to wartime footing. During WWII design and production ramped up to an unprecedented degree. There is a possibility that Covid19 will spark such a global response.

View on HN

`...would rather lose a lot of good drugs than approve a bad drug.`

Thanks for saying this, and rightfully so! Approving a bad drug in the world where distribution channels are very well set up could have disastrous effects.

We can point to things like the opioid crisis in the USA right now - a drug that is so addictive it should have never been put into use the way it was.

I don't understand the "rightfully so!" There is a real cost to not approving good drugs in order to avoid approving bad ones.

It is a matter of diminishing returns already somewhat accounted for in the approval process. Cancer drugs are under far more lax standards given the alternative is death by cancer as even an ironically carcinogenic cancer treatment which gives it again later is preferred to dying of it now.

As opposed to a slightly better allergy pill which more caution may be afforded. Not perfect by any means but it isn't totally blind.

> There is a real cost to not approving good drugs in order to avoid approving bad ones.

Do you have any evidence for this? Seems to me like a classic shibboleth of the pharma guild.

Not sure what sort of evidence you're looking for. There are certainly approved life-saving and other valuable drugs. If they had been approved sooner they would have had a greater benefit. There is an institutional bias towards avoiding costs due to approving a bad drug versus delaying approval and incurring unseen costs.

This seems like a difficult thing to test maybe even unethical. In order to detect the rate of false negatives you would have to re-test drugs that were shut down for being too risky.

of course there is, it's just a lesser cost than the cost of approving bad ones

How can you possibly know that?

It's about psychology, not numbers.

the parent post referenced Thalidomide. a disaster like would seriously shake public confidence in medicine, across the board

The public seems able to understand that there is a cost/benefit tradeoff in driving cars. They could probably understand that idea in the case of drug approval also.

People accept the risk of driving because they feel like they're in control and can prevent a bad outcome. The same is not even remotely true about a drug.

I agree. I think this is why self driving cars will be difficult for society to accept. It's an odd topic because most people would never want to drive their own plane, yet willingly get on airplanes with some % chance they will crash.

The notion that there is some % of cars that will crash seems more palatable when you're the one driving rather than some robot. Or maybe that's just because I pretend I'm a better driver than the self driving cars. This sort of "I could do it better..." psychology is probably what drives these behaviors.

I can't build a better drug; .'. I trust that medical professionals will keep me safe. I can't fly the plane in a safer manner; .'. I trust that pilots will keep me safe. I might be able to drive in a safer manner; .'. I don't trust a self driving car.

The anti-vaxx movement would like a word...

>Thanks for saying this, and rightfully so! Approving a bad drug in the world where distribution channels are very well set up could have disastrous effects.

Denialism's Michael Specter disagrees. Vioxx was one solution to the pain problem but 2.4% of participants showed heart problems when using it, so it was pulled off shelves. That's like saying no, 97.6% of the population can't have penicillin because 2.4% are allergic to it.

That’s about the current death rate of Coronavirus and there is panic so yeah

Also see thalidomide in Europe

Wow looks like there are still people around who suffered deformities because of the mistakes in approving thalidomide.

https://www.bbc.com/news/magazine-15536544

It's also still used successfully as a medicine for a number of uses.

I sure they take more precautions like make sure women who take the medications are not pregnant.

Yes, exactly. Seems better than a blanket ban. As pointed out elsewhere, the FDA is basically trying to keep dramatic risks/outcomes low, to maintain confidence in approved drugs, but tends to have a high false negative rate. What's interesting is that Vioxx was approved, but later appeared to have huge negative impact when it was given to large numbers of people (and apparently, it was also evident in the raw clinical trials data which was submitted to FDA).

It's an entirely non-trivial, multi-dimensional, systemic problem.

The obvious risk here is that you're combining a rush job with limited oversight, and large-scale distribution. So any mistake that would have been caught by the existing process will instead have the potential to roll into a global disaster.

Compared to 2% chance of death from the virus? I'll take a 2% chance of blindness over 2% chance of death for example. Of course we don't know what the odds are for the negative side effects.

Just to be clear, a 2% mortality rate does not apply to everyone. It hits the sick and old way harder. As a healthy person in my early 30s, my best guess for my own mortality rate is 0.4%. However, that does mean it is extremely dangerous for the elderly or otherwise compromised.

like, for example, if the vaccine is made from chicken eggs and a non-sterile batch gets through because we scaled up the pipelines faster than what was safe...

BTW, your 2% chance of death is assuming you catch the bug (and that the 2% figure is accurate). vs a much higher chance that you take the vaccine.

Do you prefer a 0.02 * 0.8 == 0.16 chance of blindness over a 0.02 * 0.05 == 0.001 chance of death?

Nobody knows what the real numbers are, so it is pointless to ask that question.

2% might be the overall death rate if this thing keeps coming back until eventually everybody has been infected. Or maybe it mutates to be less deadly. Maybe the hypothetical vaccine is only a .2% chance of deafness not blindness - or a 10% chance of your little toe falling off...

I agree, but the death rate of COVID-19 may not be high enough to justify a huge increase in risk tolerance.

Dearh rate for sever cases is huge.

We'll get the first test results at end of march/start of april.

If I have 30% chance of dying I take anything that lowers the chance significantly.

Even if the side effects are cancer in a year? Strong, consistent headache? Organ poisening? Permanent weakening of immune system? ...

And you would have to be very old and with strong preconditions, to have a 30% death chance with corvid-19. And then any side effects would be even more fatal.

Why would I get those side effects?

Remdesivir has already gone through safety trials years ago, it's in phase 3 now.

"If I have 30% chance of dying I take anythingmthat lowers the chance significantly. "

That's arbitrary. What's the death rate for severe flu? If I define severe to be "throwing a fever of greater than 42 C" it'll be really high.

flu mortality is about 0.1% though, like Covid, varies by age group/pre-existing. Also varies by strain/year

The parent's point is about the metric of "death rate for severe cases", which is ambiguous depending on how you define severe. You could pick a definition of severe such that the death rate is 100%, but that wouldn't be super useful.

The running trial gives the definitions, it's not ambigous as you wouldn't be able to run a Phase III trial without it.

That's true. A higher risk vaccine may be worth it for folks at high risk or with high exposure.

Risk tolerance should be seen as a sliding scale, not as a binary.

You may be correct about the death rate, but the disruption that it's causing the global economy may change the calculus in the FDA.

We should risk people's lives because doing otherwise may harm the economy..?

Harming the economy is also a risk to people’s lives, though.

To what extent? Please expand because this depends on how people's lives depends on the economy.

You can take good measures that slows down the economy without endangering people's lives. Of course you can also consider social unrest to be a risk but that's another matter.

Well, for one, I depend on the economy for my food. I also depend on it for my drinking water. Oh, and in winter, for my heat.

Leaving that aside, I've also found the economy helpful in providing me with clothes and shelter.

And when I have gone off in the woods and played survivalist (and yes, I have), I depended on the economy not only for much of the gear which enabled this, but also for keeping me safe from attack by other tribes while off in the woods. No-one was fighting me for a particular foraging/hunting ground.

If people's lives don't depend on the economy, why don't we all stop working?

Wall Street and stock prices are detached from day to day life for most people.

"The economy" isn't just what happens on Wall Street.

So you’re in favor of policies that aggressively redistribute wealth to help mitigate such risks for the most vulnerable (the poor)?

Man, where have you been?

I'm not saying that, but I'm sure there are people in the Government who are taking that point when making these decisions.

I feel like I should mention SV40 now https://en.wikipedia.org/wiki/SV40 When you do mass inoculation the last think you want to do is rush something out the door that could cause even more problems. Granted SV40 did not harm much at all, depending on who you talk too.

Production of pharmaceuticals (such as a vaccine, when it's ready) and medical equipment (such as ventilators) can be ramped up by commissioning and requisitioning an arbitrary number of factories, but developing a vaccine is inherently a sequence of trials that can only be "multithreaded" between relatively few research teams trying different approaches semi-independently.

True, but we can speed it up a lot. There are a lot of researchers currently looking something else that be a great research assistant (I know you are a great cancer researcher - that means you have the biology background to be a lab assistant). Even those who don't have a medical background can do menial data entry tasks, or clean up tasks.

Ummm, not quickly. Vaccine production requires very sterile facilities which take time to build.

You do not want to be taking vaccines produced under non-sterile conditions.

Also, are you assuming that supply chains are intact? How are you getting the raw and the finished materials to build the factories?

Raw materials are not my concern - the industries that supply them can get all the people they need from others. You might find yourself demoted from programmer to some third shift factory labor job, but with a few weeks training you can do it (with okay odds that you won't lose life/limb from missing some safety item)

You are conflating raw materials with trained employees. Which brings up another barrier to scaling up production.

Raw materials: you need pipettes, chemistry hoods, etc, lots of specialized manufactured items, in order to build the production pipeline. Where do you get these? What if some of these components are manufactured in China, guess what is on hold?

Employees. Vaccine production is high-skilled work. It isn't trainable in just a few weeks even if you have reduced workplace safety standards.

Only 10-20% of the employees are highly trained. HR, finance, and other management functions are universal. There are a lot of highly trained people doing those other rolls, but they can be shifted back into the role they are trained in.

By putting 4x as many random people on the problem you can double the amount of work done, but that is about the max you can get.

China is well aware of what is needed, I expect they will let the factories making the needed specialized things run while nothing else does. N95 masks can be made avaiable to thsoe who need them which includes not just medical personal but also manufactures of pipettes. If you make something else stay home.

The article makes it very clear that he is speaking of the immediate forecast (next 3-6 months) not the longer term (6 months to 1 year or more). The virus is already all over the globe, so technically it's already edging near a pandemic. The question to ask (and the article asks it) is do we have something that will materially alter spread in hand today or at most within the month. If the answer is "No" then the virus is likely to spread. "No" is a correct reading of the facts as they stand today.

This isn't about FDA conservatism or anything like that, it's "what do we have in hand?" No matter how promising current experimental treatments might be (perhaps Remdesivir will work well), they WILL NOT stop the spread in the short term. They WILL MAYBE help with treating acute cases and lower fatality rate. Still, good confirmation is a few months out, so for March/April we're still looking at very little.

The only therapeutic that could stop the spread is a vaccine. That's even further out, they're just starting some tests, but it's not going to be ready for wide deployment for another year. You don't want to inject things into healthy people on a massive scale without doing a few controls.

All of these approaches suffer from production questions too: it takes a while to ramp, but yes that could be helped by a massive effort.

In short, wash your hands. A lot. Don't touch your face. If you're sick, quarantine. In 6 months we'll hopefully be better equipped therapeutically, but don't expect that to solve the immediate problem.

That's not the rate limiting step. It takes time to determine if the vaccine works. Getting this wrong would be disasterous.

Very true for example Mefloquine was developed by the United States Army in the 1970s

Disagree - this isnt 22000 ad its 2020 ad - while we have technology, development of vaccines is still very slow, and of course, even if we did have a much faster process to develop such a vaccine, there is of course, the financials, politics, manufacturing, quality control, distribution, monitoring, etc infrastructure must be put into place.

The Pandemrix vaccine against the latest pandemic (the swine flu) caused narcolepsy in some individuals -- seemingly with a propensity for affecting young people and possibly related to the choice of adjuvant.

We're in a similar situation now. Early vaccines may have side effects and a forced introduction of a vaccine may increase the risk.

> Drug development has been very conservative especially in the US for the past several decades

That is actually not what I hear from people who know this stuff. In fact many are very concerned that the FDA is far too quick with drug approval based on weak outcomes. (Big recommendation: Follow Vinay Prasad on twitter, he's spot on on these issues.)

> During WWII design and production ramped up to an unprecedented degree

This isn't production, this is science. There's only so much you can do. You need to do trials to know if things work. You need people for that. If you want to test a vaccine you need to have an at-risk population that you can test and you need to wait till you get the results. Yeah, there are certain ways to improve that and do things faster. But only so much. There are hard limits. You can't develop a drug or a vaccine over night.

> very concerned that the FDA is far too quick with drug approval based on weak outcomes

The current FDA requires strong evidence of safety, over a long period. Prasad etc are arguing the FDA should be stricter on efficacy. In a pandemic, it makes sense to lower your standards for safety to get new treatments out sooner.

This is required for good reason. It only makes sense to lower standards when there is going to be a realistic payoff. The general public place a lot of trust in the various parties to make the right decisions. Running with too much risk from adverse events is extremely problematic at different levels.

People in pharma and in authority (FDA and other organizations) are intimately familiar with the need to strike the right balance and have a lot more considerations than you seem to assume they do.

The FDA already allows lesser standards for efficacy for rare things. Which is why there are a lot of treatments for rare things - once you are on the market you can "wink wink" tell doctors about the more common thing you also work on.

America did a lot of science to produce a nuclear weapon quite quickly once given the resources to do so.

It didn't involve clinical trials on humans. That's the one thing that takes time in medicine and can't be optimized away. You can't do medicine without clinical trials.

You can speed up clinical trials by running larger ones at earlier stages. This risks more deaths from bad medicines, which may still be worth it in a pandemic.

Though the people dying (or losing their friends and families) from the bad medicines might not see it that way.

When people are dying in mass numbers from not being in the trail though...

Large clinical trials require large numbers of patients, which thankfully we don't have, and also what we're trying to avoid.

If things get bad enough that this is worth the risk, we will have large numbers of patients

Human trials cannot be optimized away, but they sure can be optimized. For example, you can do away of IRB reviews, which require lots of effort on the side of researchers, take lots of time and heavily restrict what researchers can do, but which ultimately aren't necessary to run a human trials. Of course, IRB reviews have their purpose, and doing away with them altogether in normal times is most likely a non-starter, but in emergency pandemic time having drug earlier might justify lots of risk of harm to trial subjects.

its time in logistics dude!

Your are making the same bad assumption that our President is making.

Let me clue you in,

Any idea how ,many test kits we need for 400 million?

400 million for one time use but there will be more than one time use...ie more than 400 million.

any idea how much money it takes to get just 40 million test kits out to Americans for free?

One Billion dollars...

Now here comes the logistics bite in the ass.. how many trained US Health Public Service workers do we have skilled in showing how to use such test kits?

Oh damn Trump cut CDC funding and public health funding...

It snot rocket science if you cut public health funding than you cut the very fabric of the future response to such an epidemic in the first place..ie cannot spend even the demo price of 7 billion to immediate fix no immediate fix is there its been cut already!

Rant mode off

By the way, the DJIA lost trillions of dollars. That's like 100 Manhattan Projects ($20 billion adjusted for inflation).

EDIT: Don't know why I'm down-voted, but my point is that an extremely aggressive development project is easily justified if it can reduce the health and economic impact. 10 Manhattan Projects worth of effort to accelerate mass treatments, if it even just halved the economic impact, would easily pay for itself.

Your statements about the FDA are way off mark. There's plenty of very dangerous drugs in use now riddled with misinformation around their safety. Fluoroquinolones and statins are some of the worst offenders with terrible irreversible cumulative side effects and the fda barely has done anything for them. Many in the know believe they're conplicit on the matter since those drugs go out to basically everyone now at some point.

Reply on news.ycombinator.com