Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The FDA's perpetual process machine

paulromer.net

Post reply on HN
Print view View original

51–60 of 61 posts

Re: The FDA's perpetual process machine

#51

Post by refurb »

Symmetry wrote: post #40
refurb wrote: post #32

Earlier quoted context omitted.

The EMA is? The EMA typically takes longer on average to approve a drug than the FDA.[1] And if you want to look up an EU scandal, check out the Servier and Mediator. Hundreds of deaths reports and it took France almost 20 years to pull it off the market.[2] [1] https://news.yale.edu/2017/04/05/fda-approves-drugs-more-qui... [2] https://www.fiercepharma.com/pharma/servier-faces-courtroom-... .

Sure but I'm thinking of different metrics here.Europe has a variety of competing EpiPen equivalents so no price gouging. They had beta-blockers approved a decade before they were available in the US, with a small risk of liver failure for that first drug true, but many, many more lives were saved than lost. They had sunscreen that stopped UV A a decade before the US because the FDA's test only looked at skin burning…

Valid point for EpiPen, but a part of the reason for the US price gouging was physician prescribing - there were/are non-equivalent EpiPens available in the US (different, but still effective). Doctor choose or don’t know to prescribe it.

The beta-blockers is also valid, but back then (mid-20th century) the EMA (or constituent countries) were consistently approving faster than the FDA. It wasn’t until the 80’a when the FDA was revamped to be more like the EMA.

Regardless, even though some may prefer failures like Mediator or Thalidomide, I just don’t think there is the political appetite to tell folks “yeah, the drug turned out to hurt people, but that’s the price we pay for having faster approvals”.

Logically it makes sense, but when it’s your child who is hurt by a drug, people get very risk adverse very quickly.

Re: The FDA's perpetual process machine

#52

Post by refurb »

amanaplanacanal wrote: post #50
refurb wrote: post #10

Earlier quoted context omitted.

Higher blood pressure is linked to increased mortality. It's not like they are blocking the drug for fun.

You really need to look at all cause mortality. You can’t just say A is associated with B, therefore we should keep A low. We make that mistake all the time as far as I can tell.

My understanding is that connection has already been made through several robust trials.

There may be drugs that raise blood pressure, but not mortality, but so far we haven’t actual data on one of them.

Re: The FDA's perpetual process machine

#53

Post by gwern »

refurb wrote: post #9
jschwartzi wrote: post #4

Earlier quoted context omitted.

The FDA is largely composed of Type B people. They are very risk averse and will often delay a new technology simply because there might be unknown risks. There are technologies widely used in European ventilators such as automatic alarm control or adaptive alarm limits which the FDA views as unsafe based on their fear that the alarm system might fail somehow. It doesn’t surprise me that they tried to slow the approv…

I work with the FDA as a part of my job and I disagree they are overly risk adverse. When we previewed our development plan with the FDA, they actually asked us "can't you move faster?". When a safety signal showed up, they were more than willing to work with us to develop a protocol to reduce the risk while still providing patients with drugs. I don't disagree that the FDA is very concerned about safety, but in situ…

It's nice that they offered to slightly tweak the protocol for you. How very generous and magnanimous of them.

On the other hand: coronavirus.

Re: The FDA's perpetual process machine

#54

Post by pinkfoot »

jschwartzi wrote: post #4
causality0 wrote: post #2

The sorry state of the FDA and drug development and deployment as a whole is in my opinion due to a preponderance of two types of people: Type A will happily bankrupt their customers while giving them heart failure in order to reduce allergy symptoms by 10%, and Type B would rather let ten thousand people die in unmedicated screaming agony than have a drug show up in a news story saying it killed three of them, regar…

The FDA is largely composed of Type B people. They are very risk averse and will often delay a new technology simply because there might be unknown risks. There are technologies widely used in European ventilators such as automatic alarm control or adaptive alarm limits which the FDA views as unsafe based on their fear that the alarm system might fail somehow. It doesn’t surprise me that they tried to slow the approv…

How would you classify Frances Kesley who prevented thalidomide from being sold in the USA. And how would you rank the German authorities who allowed it to be sold over the counter?

Re: The FDA's perpetual process machine

#55

Post by joe_the_user »

skybrian wrote: post #48
joe_the_user wrote: post #28

Earlier quoted context omitted.

Thing about these problems (and others) is that bottle necks and red tape generally are problems in just about any country. Political leadership in this situation involves noting the regulatory and bureaucratic hurdles to action and using your authority to supersede them when a pressing situation requires it. Someone has to sit down and say "what needs to happen, what's standing in its way? What orders have to be iss…

While it certainly would have helped, ideally things would be set up so that political leadership isn't needed to override everything in a crisis.

It's not a matter of overriding everything or nothing. Any crisis is going to involve some level of standard operating procedures not working, even with the best planning (though the US through out planning beforehand). Effective action requires a chain of events all happening and some central institution has to look at the process as a whole because 9 out of 10 things happening well can easily not be enough.

Re: The FDA's perpetual process machine

#56

Post by beagle3 »

LatteLazy wrote: post #41

I find this interesting as the FDA basically fell over themselves to sign off in treatments. They gave chloroquine the ok based on less than nothing and other drugs have been given breakthrough treatment status without trials. I can see an arguement to "try anything" in an emergency. But how did the fda end up being BOTH to lose and too tight with their regulation?

Chloroquine and hyrdorchloroquine have been in use for decades (70 years for chloroquine!) with millions taking them daily and a very well known safety profile. It was already FDA approved, could have been used by any doctor for COVID as an off-label treatment already. Everything about CQ/HCQ has been ridiculously politicized to the point that the facts have become very hard to find.

In fact, given the huge number of people taking them for lupus and other diseases (a significant fraction of a percent of the population, having autoimmune disorders and poor health and thus at greater COVID risk), we've basically had a biased but ultra-informative experiemnt already: What percentage of them got covid and what severity, as compared to general population, and compared to other comparable populations (autoimmune that isn't treated by CQ/HCQ)? It's not perfect, but it would have been very informative. CDC has this info. No ethical issues about it.

Re: The FDA's perpetual process machine

#57

Post by JediWing »

BurningFrog wrote: post #45
refurb wrote: post #32

Earlier quoted context omitted.

The EMA is? The EMA typically takes longer on average to approve a drug than the FDA.[1] And if you want to look up an EU scandal, check out the Servier and Mediator. Hundreds of deaths reports and it took France almost 20 years to pull it off the market.[2] [1] https://news.yale.edu/2017/04/05/fda-approves-drugs-more-qui... [2] https://www.fiercepharma.com/pharma/servier-faces-courtroom-... .

There are two types of risks in this field: 1. Approve something too soon. This kills/hurts patients using it. 2. Wait too long approving something. This kills/hurts sick people who are deprived of good treatment. The classic big FDA flaw is to heavily lean towards type 2 errors, because type 1 deaths lead to bad PR, while no one notices type 2 errors. So that EMA apparently made a type 1 error here is possibly a goo…

It's worth noting that in medical culture, knowingly implementing a process that does not aim to eliminate all type #1 errors as much as possible would violate the spirit of the Hippocratic oath, whereas #2 would not.

Re: The FDA's perpetual process machine

#58

Post by refurb »

gwern wrote: post #53
refurb wrote: post #9

Earlier quoted context omitted.

I work with the FDA as a part of my job and I disagree they are overly risk adverse. When we previewed our development plan with the FDA, they actually asked us "can't you move faster?". When a safety signal showed up, they were more than willing to work with us to develop a protocol to reduce the risk while still providing patients with drugs. I don't disagree that the FDA is very concerned about safety, but in situ…

It's nice that they offered to slightly tweak the protocol for you. How very generous and magnanimous of them. On the other hand: coronavirus.

I honestly don't think it's fair to crap all over the FDA for Covid. One day we read that people become immune, the next day they lose it. One day hydroxychloroquine works, then next day it doesn't.

It's hard enough managing drug approvals when you have all the data and plenty of time.

Re: The FDA's perpetual process machine

#59

Post by skissane »

seehafer wrote: post #3

There are number of conflated regulatory problems in this piece: 1) FDA needs to move its data ingestion into the 21st century (note: large industry can be as resistant to changing current process as bureaucracy). In no rational universe should we have to mail data in 2020 2) In vitro diagnostic regulation is a mess and needs to be revamped top to bottom. Either exempt all Lab Developed Tests, rewrite the law to make…

> 1) FDA needs to move its data ingestion into the 21st century (note: large industry can be as resistant to changing current process as bureaucracy). In no rational universe should we have to mail data in 2020

I find it fascinating that the FDA wants data in SAS 5 transport files [1][2], a file format with very obvious mainframe heritage – 80 byte records and floating point values in IBM hexadecimal floating point instead of IEEE [3].

[1] https://www.fda.gov/media/80146/download

[2] https://www.fda.gov/media/88173/download

[3] http://support.sas.com/techsup/technote/ts140.pdf

Re: The FDA's perpetual process machine

#60

Post by zaroth »

cycomanic wrote: post #39
zaroth wrote: post #30

Earlier quoted context omitted.

There’s no reason to play partisan politics here. If we want we can say Democrats were too busy trying to impeach Trump and calling Wuhan a “manufactured crisis” back when Trump ordered the borders closed Jan 30. The simple fact is non-partisan non-evil, simply incompetent bureaucrats at the State and Federal levels all failed to adequately prepare (PPE stockpiles) and had actively harmful policies set in place years…

So you are equating democratic legislators taking part in the impeachment hearings to the trump administration dismantling the CDC pandemic response team? Let's not even get into the "corona will just disappear... or it's just like the flu..." These arguments about "let's not make this a partisan issue" completely ignores the reality. The same about how news and politics are so divisive at the moment, yes they are bu…

So much for not playing politics, the only reply I get being one that falsely calls Trump a Nazi sympathizer.

> yes they are but let's not forget one side of politics literally said people carrying swastikas are very fine people as well

I'll assume you're not being sarcastic, and perhaps just inadvertently making my case. What Trump said about Charlottesville was literally the opposite of what the media reported him saying. Watch the actual video of his remarks, or read the full transcript.

If you're willing to examine the actual primary source evidence in that case, perhaps it will help you take a fresh look at how COVID is reported w.r.t Trump as well.

As to your question, no I was not equating that. I am saying just one of many, many flubs was when Democrats called COVID a "manufactured crisis" in the early stages as it was unfolding during impeachment. Instead of pushing the Administration to do me, COVID response was made out to be a red herring or a con used as a distraction. Before we even knew was SARS-CoV-2 was, the response was already being irreparably politicized.

Post reply on HN