Comment by skunkworker | Hacker News Reader

Comment by skunkworker | Hacker News Reader

Parent Full thread

If I read the linked PDF correctly (https://www.sirillp.com/wp-content/uploads/2021/11/020-Secon... PDF WARNING)

The issue stems from the number of total pages (329,000) and the proposed record release schedule of 500 pages per month

"FDA proposes to process and produce the non-exempt portions of responsive records at a rate of 500 pages per month. This rate is consistent with processing schedules entered by courts across the country in FOIA cases"

"FDA’s proposed schedule of 500 pages per month is consistent with schedules set by courts across the country, including in cases where the underlying records were of national significance"

"The FDA has proposed to produce 500 pages per month which, based on its calculated number of pages, would mean it would complete its production in nearly 55 years – the year 2076. "

To me this seems like it is more about bureaucracy rather than maliciousness. And in my view a fair criticism of the slow pace that FOIA documents are released.

Imho the relevant thing here is that if this is allowed to stand, by inflating the size of data the goverment can effectively hide any data it wants and block the public from accessing it while retaining a veil of "nothing wrong done here".

Additionally I would say that foi requests are generally fulfilled more by document categorisation than by page, especially if the case is not about military secrets where much more careful screening might be required.

My first thought is that there are 329,000 pages of data related to the Pfizer vaccine alone and we are supposed to believe that the FDA read and evaluated them before granting authorization.

Why would it all need to be read?

I'd expect that a lot of it is data that is gathered because it might be useful if certain things came to pass that did not come to pass.

For example, suppose in phase 3 they had found that the vaccine provides excellent COVID protection but it also has some terrible side effect in too many people to approve it for general use.

If they have extensive medical records on all the people in the trial they might be able to see some way to tell who is likely to have the bad reaction and who is not. Then they can try to find some way to mitigate that, or go for an approval for the vaccine for people other than those who are likely to react badly.

If things go smoothly on the other hand, there may be no need for anyone to look at much of that data.

getting approval for a brand new drug intended for millions of people and you don't think its important to read all submitted data? reminds me of Nancy Pelosi "we have to pass the bill so you can find out what's in it"

it's not ALL first-order data, man. have you ever seen an official document of any size? 99% of it is description of the 0.25% which actually matters, and the remaining 0.75% (AT LEAST) is whitespace.

That 0.25% is mostly known to subject matter experts. The rest is there to meet documentation requirements.

it is definitely NOT 329k pages of absolutely vital facts that can't be skipped.

besides, it can be split up among many people over the course of time.

experts know how to do this stuff.

I get the importance to allow certain DOD orgs to sift through the documents to ensure there are no national security secrets revealed to avoid any national threats.

However with this they should be fully transparent with any and all vaccine related data as they are already suffering from credibility concerns when it comes to vaccine.

there are only credibility concerns among people who don't understand vaccines and who don't trust people with decades of experience to know what they're doing.

this group is an EXTREMELY vocal micro-minority.

Exactly.

Here's literally the law on what goes into a New Drug Application: https://www.law.cornell.edu/cfr/text/21/314.50

The idea is that it serves as the "whole story" of a drug, and includes everything from preclinical data demonstrating how the drug works on a molecular level, up through the plans for actually manufacturing and administering it. Thus, it includes a lot of things "for the record". For example....

If the application relies on material published in another language, then you are required to submit the original and a translated copy. It is very unlikely that the FDA is going to retranslate each article.

Similarly, you need to include individual case reports for any trial participants who die/drop out from adverse events even if in the placebo group and the cause is unrelated to the drug. If Mrs Jones from the control group got eaten by a shark after getting the placebo, a bunch of her medical info will be in there. This a) bulks up the application and b) slows down redaction but c) takes a reviewer about two seconds to realize they can ignore that entire section.

>it can be split up among many people over the course of time

Oh, yes.

>experts know how to do this stuff

https://www.americanbar.org/groups/judicial/publications/jud...

"One way to test the effectiveness of a redacted PDF document is by using a technique that nearly all word processing users know--copy and paste."

...nearly...all...

don't confuse usability and bad software with domain expertise. those are wholly separate things, and are not related in any way; they're completely independent skill sets.

Not any more ridiculous than believing Pfizer produced them.

I guess the data exists in digital format, because no human reads 329,000 pages of such a report. If it is data, it muss be electronically processed just as it was electronically gathered in the first place. Inefficient bureaucracy isn't really a sensible excuse. You could hide any data with stuffed documents like taht.

Reply on news.ycombinator.com