Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Leafsnail]

Also the entire argument is based on the idea that hundreds of children are dying of vaccines every year, with nobody noticing. Even though we have bodies dedicated to monitoring people for adverse reactions.

[Shazbot]

The whole premise is flawed. "2014 America" is not a vacuum-sealed bubble, and the diseases vaccines prevent are being brought in to the country constantly through immigration and travel. The entire argument falls apart from there.

Every day that passes, you change the point of reference.  On january first, 2015, the reference point for decision would be 2015 America.

I am not suggesting it is a permanent reference point we should use forever. I'm saying that the best logical reference point for policy on Sep 4, 2014, is the situation on the ground locally on Sep 4, 2014. I'm not sure how this is remotely controversial. (Keeping in mind that I've agreed several times you should also keep one eye toward the vague future by making sure you're outside of endemic levels for eventual eradication, mind you)

You have ignored everything I said of actual importance to quibble about dates.

[GavJ]

I'm saying that you're equating the amount of data we have right now with no data at all, which completely ignores the fact that we do have data and it's not insufficient at all.

No, like I said, I think the amount of data we have right now probably guarantees something around "55-60% or higher vaccine coverage for measles" A guess based on extrapolation but with more examples of countries, I could be more confident.  Point is though it's a lot higher than zero. And lower than 70% (not the optimal rate! Just "the amount we know for sure from data the optimal rate is higher than")

Zero data would only guarantee "0% or higher"

@Leafsnail

Also the entire argument is based on the idea that hundreds of children are dying of vaccines every year, with nobody noticing. Even though we have bodies dedicated to monitoring people for adverse reactions.

Are you referring to the one that is largely oriented toward court remuneration and explicitly doesn't even try to verify causality, that the huge majority of the population hasn't heard about, that isn't mandatory for doctors, that doctors have outright said they commonly don't utilize, that only cared about a fixed window of time after the vaccine, that has huge two-way selection bias, and that has no control group? This one https://vaers.hhs.gov/index ?

It is almost universally laughed at by everyone on all sides of this issue.

@Shazbot:

You have ignored everything I said of actual importance to quibble about dates.

What else did you want me to respond to?
"If 90% is vaccinated, then 10% of the population is at risk" Well yes, by definition... However the risk for those 10% is extremely lower than at, say, 80% vaccination, and the 10% FEWER are at risk from potential vaccine complications (which do not get any less risky with changing coverage).

Did you read the thread? I mean, these are all sort of basic assumptions laid out in the OP. If you disagree with the math or something, then please specify further. But just saying one sentence that's addressed already isn't much to respond to.

Other than that, your post was just listing some vaccine-related laws regarding school and health care and stuff. Clearly, a country wouldn't explicitly aim for an 85% vaccination rate or whatever, and yet still also require 100% vaccination for schools or health insurance. Unless the optimal rate is 100%, you would simply change stuff like this according to your overall policy.

[GavJ]

As an example of a reasonable law, for instance you might be able to say (after having enough data to find the optimal rate for sure):

a) "If federally published vaccine coverage is below the target level, and has been for over 1 continuous month, then you can be denied new insurance coverage for [that disease or known symptoms/complications thereof]  for that person until either vaccinated or the rate hits target levels"

b) "If federally published vaccine coverage is at or above the target level, and has been for over 1 continuous month, then if you get vaccinated at that time, you can be denied new insurance coverage for that person for [whatever the huge studies identified as the types of vaccine reactions relevant] until the rate falls below again or ___ amount of time passes (expected period of time for most of adverse reactions)."

c) "If it changed threshold in the last month, there's a grace period and you can't be denied on these grounds"

d) "Regular existing exemptions [blah blah, e.g. medical contraindication] apply."

e) "Pharmacies/doctors can allow vaccination at any time it is available, but must legally update you on the current status, i.e. (a) vs. (b) vs. (c) above, when you come in, prior to vaccinating" and give you a receipt to that effect, etc.



(This would act as a strong incentive to keep the rate near target by the way, which responds further to one of the questions from earlier about how you enforce a specific number)

[Leafsnail]

VAERS is not the only scientific body in the world dedicated to studying adverse vaccine reactions, so pointing out flaws with that is not sufficient to demonstrate your case when many medical bodies who do not rely on it have come to the conclusion that the MMR vaccine is safe.  I think we'd notice if literally hundreds of deaths were being posted on it every year though.

[Neonivek]

By the by did you include the rate of maiming and hospitalization too?

[GavJ]

VAERS is not the only scientific body in the world dedicated to studying adverse vaccine reactions, so pointing out flaws with that is not sufficient to demonstrate your case when many medical bodies who do not rely on it have come to the conclusion that the MMR vaccine is safe.  I think we'd notice if literally hundreds of deaths were being posted on it every year though.

Which system is better? 
The Canadian one is almost worse: still voluntary, swiss cheese of a database, and only has like a couple years of data in it. It also is explicitly not requiring of verified causal links, and encourages all reporting of anything. I believe it is also only reportable by public health officials or manufacturers, not general citizens.
The Danish are I believe implementing a mandatory vaccination reports and health records database available to researchers as of this year, I want to say? Previously was similar to the other programs and not comprehensive either
The Aussies have a system with most of the same problems as VAERS: voluntary, not verified or causally controlled, etc.
New Zealand is/has rolled out a similarly voluntary and mediocre system in the last year or two.
Almost all of these have suggested windows of time to consider that are pretty short.

So what are you referring to that's better?



"You would notice hundreds of deaths on systems like these" IF they are super obvious, like anaphylactic shock, occurring seconds or minutes right after the vaccination (although even these VAERS and other databases explicitly warn are often not reported! The CDC estimates something like a 10:1 non-reporting rate or something, I think?, even for obvious vaccine reactions).

But what about some kid who gets a vaccine then a month later starts showing very mild symptoms, deteriorates and dies of kidney failure after 6 months? Almost never would that be reported to any of the above databases. Maybe he ate some rat poison while nobody was looking, or who knows? Most doctors probably wouldn't even have vaccines enter their head as a fleeting consideration for this. Yet it might, in reality, be directly caused.

Relying on "oh we would have seen the deaths" is very presumptuously assuming that doctors are some sort of hypervigilant, superhuman pattern detectors that can notice mildly correlated trends over thousands of patients just on a hunch, in order to voluntarily report to a registry (that they don't get compensated or recognized for contributing to) that only even appears relevant when considering deep potential connections over huge spans of time.

This is why these databases are effectively useless compared to clinical trials -- In a trial, you don't have to have thought to look for some specific causal connection. You just record EVERYTHING and compare EVERYTHING to baselines, and will find things above X occurrence threshold, period, without having to have any theory or be super observant, or anything of the sort.  It would catch even the most obscure, unexpected, unlikely of causal links just as easily as it will catch the ones everybody would think of.

By the by did you include the rate of maiming and hospitalization too?

No, for purposes of the thread and all of our sanity, I'm neither considering hospitalization for disease OR for vaccine reactions. This is fairly ignored on both the risk and the benefits sides of the equation. Deaths are much easier to think about and find data on for sake of argument.

And unless you have a good, solid reason to believe that the rate of death:maiming is hugely different for vaccine reactions versus disease reactions, then this is a fine simplification.

[Neonivek]

Really? because for Polio we actually got really good at making sure people survive it near the end of its run. Just that... well... polio survivors had... difficulties.

[GavJ]

I don't know what sentence or claim or whatever your "really?" is referring to.

Anyway, regarding polio, please click back through the last several pages of the thread and ctrl-F for "polio" to find previous posts on it, and its numerous and significant issues when it comes to plausible vaccine complications and logistical reporting errors for decades, etc.

It's probably the single messiest, least clear available example to work with regarding vaccine discussions.

(The reason is because it started out as the name of a generic syndrome of symptoms, not a specific antigen like most vaccine examples, a definition which then changed several times over history making all of the data inconsistent. Plus its main complication of paralysis is also a syndrome that is shared by a dozen+ other similarly prevalent diseases.)

[Shinotsa]

I'm still wondering how one would enforce an optimal vaccine rate. It'd go against the principle of justice to just publish data, as those in poverty lack the education or access to get vaccinated. Sure you might wind up with an 80% vaccination rate, but even that's pretty worthless if your main unvaccinated population is grouped together with poor access to healthcare, poor nutrition, etc. Requiring 100% of people who take advantage of public education and other such services be vaccinated allows for people who fall through the cracks or refuse these services to not significantly increase the risk of outbreak, and also does not require some to shoulder whatever potential risk there may be without getting something in return.

Also - once again I must say that we're forgetting other aspects of risk analysis. Sure a clinical trial has a certain confidence range, but when you combine the results of a clinical trial with results from post-marketing studies (Which use many thousands/millions of people) you get much more coverage.

[Reelya]

For gardasil they tracked deaths after taking gardasil. All deaths though, in a particular timeframe after the vaccine. From what I read, that would put the maximum possible deadliness of Gardasil somewhere between 1 in a million and 1 in 2 million.

For longer term measurements (unlimited years), all vaccines have a 100% death rate.

[GavJ]

I'm still wondering how one would enforce an optimal vaccine rate. It'd go against the principle of justice to just publish data, as those in poverty lack the education or access to get vaccinated.

So? If the country is currently at the optimal vaccination rate, then by definition of the optimal vaccination rate, it doesn't matter if they get vaccinated or not. The optimal risk/benefit balanced rate is defined as the point at which the risks of the vaccine exactly balance those of the disease, so those people are just as well off as everybody else.

It is not unethical to have poor people have an equal average payout as everybody else.

Also - once again I must say that we're forgetting other aspects of risk analysis. Sure a clinical trial has a certain confidence range, but when you combine the results of a clinical trial with results from post-marketing studies (Which use many thousands/millions of people) you get much more coverage.

1) Post marketing studies still do not remotely approach the numbers you'd need to verify that higher than 90% coverage is safe. Even if they were just as good as clinical trials, you're a couple of orders of magnitude short, still.
2) They're not as good as clinical studies. If run poorly, they're potentially not worth anything at all.
3) Do you have specific post-marketing studies in mind? I'm not aware of very many excellent ones that would approach the value of a clinical trial, if any. Please provide links that you are basing this opinion on.

For gardasil they tracked deaths after taking gardasil. All deaths though, in a particular timeframe after the vaccine. From what I read, that would put the maximum possible deadliness of Gardasil somewhere between 1 in a million and 1 in 2 million.

For longer term measurements (unlimited years), all vaccines have a 100% death rate.

Statistics is not the same thing as counting...

[Putnam]

1) Post marketing studies still do not remotely approach the numbers you'd need to verify that higher than 90% coverage is safe. Even if they were just as good as clinical trials, you're a couple of orders of magnitude short, still.

Please define "verify" and tell me where you're getting these numbers.

[GavJ]

1) Post marketing studies still do not remotely approach the numbers you'd need to verify that higher than 90% coverage is safe. Even if they were just as good as clinical trials, you're a couple of orders of magnitude short, still.

Please define "verify" and tell me where you're getting these numbers.

It is impossible to prove a negative, but it is possible to obtain data that supports/"verifies"/whatever (at an alpha 0.05 level) that the risk is "at or below ____ rate."

The math in the OP suggests that the benefit of measles vaccines at 90% coverage in the US is about 1/60,000,000 chance of saving your life. So in order to justify a higher vaccination rate as a government policy, we would want to show using data that the vaccine has a lower than 1/60,000,000 rate of killing you.

In order to show this, you would need obviously, at the very least, 60,000,000 people to even expect one extra death in your study above baseline (which is of course the smallest measurable number of deaths). And in reality, you need several times that to have a strong enough signal to actually show that at a p < 0.05 level. Specifically, you'd need around 300 million people.

Since the largest metastudies are around 1-2 million datapoints, that is over two orders of magnitude less than what you need to show that a higher than 90% vaccination rate is reliably beneficial. And that's assuming that the metastudies have zero flaws whatsoever and that they are perfect causal experiments (lol)

[Putnam]

So the only way to verify it is to record over 85% of the US population getting the vaccine?

Besides that, actual statistics suggest that the benefit is good enough.