Why I’m not listening about vaccines

I don’t claim to know the truth about vac­cines. I am not a sci­en­tist, or a doc­tor, or a researcher of any kind. So far the avail­able num­bers over­whelm­ing­ly indi­cate that they do a lot of good. Would we be bet­ter off if we brought back polio? Do I even need to write it out? No.

Yet it keeps on com­ing up, even among peo­ple I’d oth­er­wise thought of as intel­li­gent. I’m get­ting tired of fol­low­ing up on these con­ver­sa­tions, because the con­ver­sa­tions them­selves are sim­ply tire­some. There’s nev­er new infor­ma­tion pre­sent­ed, and any­one who does not imme­di­ate­ly accept the claims is dis­missed as a stooge or a shill for what­ev­er con­spir­a­cy there may be.

Because the peo­ple say­ing these things are friends and rel­a­tives and oth­er peo­ple I respect, and more impor­tant­ly because I care about learn­ing what is and isn’t true, in the past I have lis­tened, and fol­lowed links, and inves­ti­gat­ed research.

No more. Because I’ve done it too many times, and this is how it’s gone every time so far:

1. Go to link of anti-vac­cine site and col­lect the list of ref­er­ences (usu­al­ly pep­pered through the text and not list­ed in a bibliography.)
2. Fol­low those links, and the links on those pages, until I find the research that has been indi­rect­ly cit­ed. This is a huge time­suck. Most often they are blogs link­ing to blogs link­ing to blogs link­ing to blogs link­ing to orig­i­nal research. Some­times there is nev­er a link but one can find a name asso­ci­at­ed with an orig­i­nal study.
3. a. Exam­ine the research and deter­mine what, if any pat­terns can be discerned.
4. b. Search for rebuttals.

Each time, one or more of the fol­low­ing has been true:

1. The claims made in the blogs were not sup­port­ed by the research the blogs cit­ed because the author of the blog did not under­stand sta­tis­tics or basic math.
2. The claims made in the blogs were not sup­port­ed by the research because the author of the blog failed to under­stand (or lied about) the means of col­lect­ing information.
3. The claims made in the blogs were a dis­tor­tion of the char­ac­ter­i­za­tion made by a cit­ed expert, eg «there are rare risk fac­tors that peo­ple ought to be aware of» some­how becomes «researcher who invent­ed vac­cine says it should be banned».¹
4. The orig­i­nal research cit­ed sim­ply did not exist.
5. Sub­se­quent inves­ti­ga­tion showed that the researchers fraud­u­lent­ly fal­si­fied their results.

Hilar­i­ous exam­ple: not too long ago some­one whose opin­ion I used to respect told me about how Gar­dasil is «killing our daugh­ters». He help­ful­ly pro­vid­ed research and analy­sis of the research. At first glance it almost sup­port­ed the claims. The math was­n’t quite right but if you extrap­o­lat­ed the deaths from the study to the pop­u­la­tion being vac­ci­nat­ed the num­bers were quite dis­turb­ing. The analy­sis claimed 40,000 deaths², 140,000 adverse reac­tions, and cit­ed a study com­par­ing a claimed 1 in 912 fatal­i­ty rate among per­sons tak­ing Gar­dasil with a 1 in 40,000 cer­vi­cal can­cer death rate.

The cit­ed research results³ show 40 deaths out of 29,323 study par­tic­i­pants. That’s actu­al­ly 1 out of every 733. Much more than 1 in 912! Either way, With 170 mil­lion dos­es⁴ hav­ing been admin­is­tered⁵ that’s tens of thou­sands of peo­ple.⁶

How­ev­er, this fails in a sig­nif­i­cant and obvi­ous way: it count­ed all the deaths in the study, both of the peo­ple who were giv­en the vac­cine and those who were part of the con­trol group and not giv­en the vaccine.

«Well, OK,» con­cedes the friend. «So it’s only 21 of those 29 thou­sand. That’s one in about 1400.⁷ Isn’t that enough to make you angry, mis­ter statistics?»

No. Well, yes, it is enough to make me angry, but no, just com­par­ing the ones that died from one part of the study with the total of par­tic­i­pants does­n’t tell us any­thing. You have to take a look at how many peo­ple who weren’t giv­en the vac­cine died, and use that as a base­line to see how many more peo­ple died from the group that took the vac­cine. But, sure, look at the deaths per par­tic­i­pant from one group, sub­tract the deaths per par­tic­i­pant from the oth­er group and still there was a high­er death rate among the par­tic­i­pants who had the vaccine.

So what that study shows is that 21 of 15,706 peo­ple giv­en the vac­cine died, and that 19 of the 13,617 who weren’t giv­en the vac­cine at all died. That’s 1 in 748 who were vac­ci­nat­ed and 1 in 717 of those who weren’t.

Now, final­ly, the anti­vac­ci­na­tion­ist begins to think (semi)critically, because the very num­bers he was using to sup­port his argu­ment are now show­ing that get­ting the vac­cine pre­vents death. In despa­ra­tion he points out that maybe some of those deaths did­n’t have any­thing to do with the vac­cine. This is a very impor­tant piece of the puz­zle: only 13 of the 21 who received the vac­cine died of med­ical caus­es. 6 of the 19 in the con­trol group died of med­ical caus­es. 62% ver­sus 32%! That’s a huge difference!

Close. What we want to do is not deter­mine how many of the deaths were med­ical­ly-relat­ed, but how many med­ical­ly-relat­ed deaths there were in each group. The math is still pret­ty easy: 13 out of 15,706 (1 in 1208 or 0.08277%) and 6 out of 13,617 (1 in 2270 or 0.04406). That leaves us with a dif­fer­ence of 0.03871%, sug­gest­ing that an extra 1 out of 2583 peo­ple will die due to this vaccination.

Again extrap­o­late to the 57 mil­lion peo­ple who have been giv­en this vac­cine, and now we’re talk­ing about an esti­mate of 22,000 peo­ple dead. My friend right­ful­ly point­ed out that 22,000 sets of griev­ing par­ents is too many — it does­n’t have to be 40,000 for us to be alarmed.

This is where two very impor­tant ideas come into play: sta­tis­ti­cal sig­nif­i­cance, and the non-syn­ony­mous nature of cau­sa­tion and cor­re­la­tion. The med­ical deaths were almost all from dif­fer­ent caus­es: one case of pul­monary embolus/deep vein throm­bo­sis⁸ two cas­es of sep­sis, one case of pan­cre­at­ic can­cer, one case of arrhyth­mia, one case of pul­monary tuber­cu­lo­sis, one case of hyper­thy­roidism, one case of post-oper­a­tive pul­monary embolism and acute renal fail­ure, one case of trau­mat­ic brain injury/cardiac arrest, one case of sys­temic lupus ery­the­mato­sus, one case of cere­brovas­cu­lar acci­dent,⁹ one case of breast can­cer, and one case of nasopha­ryn­geal cancer.

It seems absurd to sug­gest that Gar­dasil could cause a post-oper­a­tive pul­monary embolism or a trau­mat­ic brain injury¹⁰. Account­ing for that brings us under 15,000 deaths but more impor­tant­ly illus­trates that sim­ply because some­thing hap­pens lat­er than an ear­li­er event it does­n’t mean that it was caused by the ear­li­er event.

There is a cer­tain amount of vari­abil­i­ty inher­ent in peo­ple’s lives; there will be dif­fer­ences between any groups even if you do your best to make each group match each oth­er for risk fac­tors and so on.

Nev­er­the­less, just to arrive at some con­clu­sions I’ll ignore sta­tis­ti­cal sig­nif­i­cance, if only to illus­trate why ignor­ing sta­tis­ti­cal sig­nif­i­cance is dangerous.

When inves­ti­gat­ing the cita­tions for the pub­lished claim that Gar­dis­il is «killing our daugh­ters» I found that the research data used to base that claim sup­ports the fol­low­ing conclusions:

Sub­jects giv­en the vac­cine were 8.4% more like­ly (5÷15,706 vs 4/13,617) to die in auto­mo­bile acci­dents than sub­jects not giv­en the vaccine.

Sub­jects giv­en the vac­cine were 346% less like­ly (1÷15,706 vs 3/13,617) to die of gun­shot wounds than sub­jects not giv­en the vaccine.

So you should always weigh the risks. Gar­dasil caus­es dis­eases and car acci­dents, but you have to fig­ure out whether it’s worth more than tripling your odds of dying of a gun­shot wound.

Then the per­son whose opin­ion I once respect­ed said, «I’m not giv­ing that poi­son to my kid, and the gov­ern­ment can’t make me do it.»

I’ve been through this process a half-dozen times. Each time, the research to fol­low up the claims made by some­one past­ing a URL into Face­book has tak­en four to six hours, and has proven if noth­ing else that the per­son post­ing did­n’t both­er to check any of the cita­tions in the arti­cle they post­ed before insist­ing that the rest of us were sheeple for buy­ing the lies.

Well, I’m sor­ry, but here’s the truth: if you believe every­thing you read on the Inter­net you are an idiot. If you nev­er both­er to check the facts on pages you repost to social media and insist are «the truth», you are a cred­u­lous idiot. And if you call oth­ers «blind» for not believ­ing the stuff you did­n’t both­er to research in the first place, you’re an asshole.

That’s why I’m not lis­ten­ing any more.

If you have a study or actu­al data of some kind to cite, do it. If all you have is a link so some crack­pot’s blog, well here you go. All you got was a link to this crack­pot’s blog.