Some perspective on the VAERS database.

[varkam]

Orac has a good post on a recent Medscape article. Within that post, though, he has some good stuff on the VAERS database:

Finally, Gandey fails to give any sort of context for the reports of 9,700 adverse reactions reported to the Vaccine Adverse Event Reporting System (VAERS). What people don't understand about VAERS is that it's not a good epidemiological for establishing strong evidence of causation, and it's dubious even for correlation. Indeed, such was never its intent when it was designed. VAERS was originally intended as a "canary in the coalmine," so to speak, in that anyone can report problems that occur in close temporal proximity to vaccines. It's an early warning system, not a rigorously administered database. "Adverse events" reported don't even have to be something that looks suspicious as having been caused by vaccines. Indeed, there is the infamous story of how Dr. Jim Laidler once reported to VAERS a most unusual complication from vaccination:

The chief problem with the VAERS data is that reports can be entered by anyone and are not routinely verified. To demonstrate this, a few years ago I entered a report that an influenza vaccine had turned me into The Hulk. The report was accepted and entered into the database.

Because the reported adverse event was so... unusual, a representative of VAERS contacted me. After a discussion of the VAERS database and its limitations, they asked for my permission to delete the record, which I granted. If I had not agreed, the record would be there still, showing that any claim can become part of the database, no matter how outrageous or improbable.

More after the jump. There's also a discussion down a few paragraphs about the likelihood of certain events being reported to VAERS as being litigation-driven. Good food for thought.

[trotsky]

1. More good stuff from Orac.

It's pretty obvious in this forum how many people are clueless about the nature of data in the VAERS.

[Liberal Veteran]

2. It does however provide a valuable resource...

...to the anti-vax kooks and crusaders.

[trotsky]

3. I wonder if the cognitive dissonance ever kicks in for any of them?

I mean the CDC/NIH/FDA are all corrupt and in cahoots with "big pharma," yet they provide this database and let anyone write anything in it. Seems to me that if these gov't agencies were all about covering up and protecting for industry, this database never would have been created let alone made completely public for researchers.

Oh well, no one ever said the anti-vax movement was based on common sense or logic.

[Liberal Veteran]

5. Not only that, but we are also to believe every country's health oversight is corrupt...

Apparently, the agencies responsible for public health in France, Denmark, Canada, Germany, Australia, Sweden, Iceland, etc are exactly the same as the Congo, Mexico, Thailand, Ethiopia and Chile.

It's a GLOBAL conspiracy to poison people!

[xchrom]

4. recommend

[dropkickpa]

6. LMAO

The Hulk! I can just imagine the number of kids who wopuld line up for flu vax if they could be the Hulk!

[izzybeans]

7. To be fair, the typical pattern of discovery for technological

risks starts with the end user, which is why VAERS exists. But also part of that pattern is an initial denial by the experts and other financial stakeholders that "the canary in the coal mine" actually exists. Claims that data in this database are "litigation driven" might also be litigation driven, or an expert merely protecting his or her professional turf.

It's healthy to remain skeptical of the claims, but not do dismiss them out of hand. It's reactionary, in fact.

Epidemiology doesn't establish causation either; though some epidemiologists naively think so (i.e. by confusing statistical prediction with causation). It establishes trends and true prevalences (e.g. the level x in the population is Z; Y is more likely to exhibit x by a ratio of 1.5 than Q and 3.4 times more likely than T, but is .5 as likely as B). The VAERS, if valid, establishes the treated prevalence of known cases, which can't be used to extrapolate the "true prevalence" existing in the population. It is as they say "a canary in a coal mine". There are typically reasons why the miners trust the canary.

[varkam]

8. The take-home point, though, is that

one should not mine data directly from VAERS and expect to draw valid conclusions as it is entirely too easy to manipulate the data that you get to support your hypothesis (which is exactly what Mark and David Geier did in their autism research).