Do you miss Dan Olmsted’s writing? He now apparently showcases his version of scientific brilliance over at Rescue Host.
Recently, he tried to pass off the Flu shots and Chinese mercury hypothesis (which I thought was David Kirby’s, but I guess I was wrong) without much more than unfounded speculation and belief.
California, of course, is ground zero as we watch autism rates keep rising — even after mercury was “removed” from childhood vaccines starting in 1999 (the situation is much more complicated than that, since more and more pregnant women and younger and younger kids are getting mercury-preserved flu shots). So if you believe as I do that autism is fundamentally an environmental illness that whacks a subgroup of susceptible kids, mercury from China — or anyplace else — is every bit as important as mercury from vaccines.
I asked him the following in the comments:
If you wouldn’t mind Mr. Olmsted, take a look at a graph of the 3-5 year-old autism caseload cohort for the past 5 years.
Such a graph would include children born at the starting point of the “removal” in 1999 you mentioned. What do you see? Does the trend look linear to you?
Do you really believe there is combined flu shot uptake and airborne mercury data that would exactly and inversely match (in dose and effect, if any) the reduction the use of thimerosal in childhood vaccines in order to produce a trendline with an R-squared value of .9954 for this time period?
You can view such a graph here.
To which he replied:
An “R-squared value of .9954” is way beyond my non-scientific expertise. All I can say is that thimerosal use has actually been increasing in by far the most vulnerable group — pregnant women — and that at least some studies suggest that greater pollution directly correlates with a greater risk of autism. If the CDC had recalled all thimerosal-containing vaccines in 1999, we’d have a genuine “natural experiment.” But we don’t. Nor will the government study autism rates in never-vaccinated kids; the survey by Generation Rescue found ominous correlations between vaccines and NDs including autism, but it’s been widely ignored.
The survey by Generation Rescue? Right. Did he just make up that part about thimerosal use increasing in pregnant women? It kind of looked like it to me, so I asked and commented as follows:
What evidence do you have that thimerosal use actually increased in pregnant women for the period immediately following the “removal” of thimerosal from childhood vaccinations? (required to make your hypothesis work)
The majority of childhood vaccines were thimerosal-free or contain only trace amounts by 2002 (more on that below). Here’s flu shot uptake estimates for pregnant women for the three years that follow:
2002 – 12.4±3.9 %
2003 – 12.8±4.4 %
2004 – 12.9±5.0 %Note: there is an increase in the estimate for 2005, but children born in 2005 and later are not old enough to be reflected in the 3-5 year-old California autism caseload cohort yet. Additionally, estimates for 2006 were back down to 12.9 percent.
Mark Blaxill and JB Handley showed up in the comments following that, and Olmsted apparently did not reply further. So the question of where Dan Olmsted might have found any data to make his Flu shots and Chinese mercury hypothesis plausible, will have to remain unanswered for now. But, while we’re on the subject of data-free gibberish, have a look at a piece of something posted by Dan Olmsted at Rescue Host on September 11th.
At some point, common sense has to prevail. For instance, let’s stipulate that better diagnosis accounts for a gargantuan 36 times more cases of bipolar disorder among kids over the past 10 years. That still would mean that the condition quadrupled in a decade — suspiciously, the same decade that autism, asthma, ADD, ADHD etc. soared out of sight. The deniers have to explain away every digit of that 40-fold number, because even a “mere” fourfold increase in the real incidence would be deeply disturbing. How can anyone be certain that one-tenth of that 40-fold increase isn’t actually real? And if they can’t be certain, why aren’t they scared to death?
Emphasis mine.
An appeal to “common sense” is a sure sign that what follows is probably not data that supports his hypothesis. Is common sense really the best way to arrive at correct answers about any subject for anyone, regardless of their background? What do statements like, “An ‘R-squared value of .9954’ is way beyond my non-scientific expertise”, tell us about the context to which Mr. Olmsted’s “common sense” might be reliably applied? Is autism epidemiology likely to be anywhere near Olmsted’s knowledge and expertise?
Sometimes common sense seems like a good way to operate, but the reality is that many things in science have quite complex answers. It’s also the case that science does not have all the answers (nor does it claim to). None of this will apparently stop Mr. Olmsted from forging ahead with assertion and anecdote in the rest of his post of course.
Did you catch this part of that paragraph above?
“…suspiciously, the same decade that autism, asthma, ADD, ADHD etc. soared out of sight.”
Did you see any real data or science whatsoever that actual autism prevalence “soared out of sight” in the past ten years? Me neither. So here you go any “Flu shot and Chinese mercury” proponents, now is your chance to post that real data or science in the comments – really, Mr. Olmsted needs your help if he’s to avoid the inevitable “You’ve got nothing!”. Either that, or get him a new hypothesis to work with (something with corroborative data preferred).
As for Mr. Olmsted’s final question in that paragraph, I’d like to answer it from one perspective.
We can’t be absolutely sure that there hasn’t been some real increase in autism prevalence, there might have been. To conclude that there has been a real increase in autism prevalence wouldn’t require much more than good data that shows it’s actually true. However, to conclude that there has been a real increase without supporting scientific evidence, but based on “stories”, is unscientific, if not a bit silly.
I can’t be absolutely 100% certain that an alien abduction has never occurred, but this lack of certainty does not translate to “therefore alien abductions are real”. I can’t be absolutely 100% certain that bigfoot doesn’t exist either, but again, that lack of certainty does not translate to “bigfoot is real”. I don’t live in fear of being abducted by aliens or encountering a hairy giant biped while on a hike with the kids, despite an abundance of “stories” about these things. I’m also not “scared to death” that there could indeed be an increase in the actual prevalence of autism. It is a possibility, but I have seen no evidence of it’s truth. I do see an increase in storytelling though.
Left Brain Right Brain 
It looks as if Dan became flustered and immediately fired off an email to his good buddy Blaxill to jump in and address the ‘epidemic deniers’ with a rare blog comment, but I guess Markie Mark missed the mark.
I think Blaxill got Olmsted involved with the whole thimerosal conspiracy theory in the first place and now he’s unemployed unless you count blogging for RxU Post.
Speaking of career change, did Mark leave BCG because they took on a few pro-vaccine clients or was he let go for the same reason?
the situation is much more complicated than that, since more and more pregnant women and younger and younger kids are getting mercury-preserved flu shots
That part is really silly, isn’t it? Sure it’s complicated, if you can’t do Math and do a few Google searchers to look for the uptake data.
“And if they can’t be certain, why aren’t they scared to death?” please! If I had to be scared to death of uncertainty, I’d be scared pretty much 24/7. I am not scared by autism, as much as I am scared by folks trying to rile people up while saying… “An “R-squared value of .9954†is way beyond my non-scientific expertise.
Do’C,
The mismatch of you with Olmstead is so obvious, that one might possibly be inclined to be sympathetic to Olmstead (if only his screed wasn’t so harmful to so many). Keep up the good work!
Kev,
Like the look of the “Kevington Post” thus far.
Joe
BTW, the 40-fold increase of BD reported in the press apparently refers to treatment, not diagnoses. I’m not sure if Olmstead et al. bothered to check the abstract of the original paper.
Increased rates of bipolar disorder diagnoses among U.S. child, adolescent, and adult inpatients, 1996-2004.
“Population-adjusted rates of hospital discharges of children with a primary diagnosis of BD increased linearly over survey years. The rate in 1996 was 1.3 per 10,000 U.S. children and climbed to 7.3 per 10,000 U.S. children in 2004.”
That’s a 6-fold increase. I can see why they wouldn’t want to emphasize that one. What you want in these cases is something that gives the impression to your audience that the sky is falling.
Also note that 7.3 in 10,000 per year is low if you consider that lifetime prevalence of BD is 1%.
I think we’re looking at the usual flaw in reasoning related to increasing prevalence of psychiatric conditions: Confusion over having a condition vs. being diagnosed with a condition.
In fact, PD, like autism, seems to be missed often. See, for example, Faedda et al. (2004) where they find that 90% of a group of adolescents with PD had not been previously diagnosed with PD.
Those bits, of course, are only for the benefit of those of us interested in facts.
_”Like the look of the “Kevington Post†thus far.”_
LOL!
I need to get some little ego-stroking photos of everyone and then I’ll be there!
At my GP’s flu shots are only given to those over 65 or with reduced immunity (not sure of the exact groups targetted). I do know that pregnant women and young children aren’t given the flu shots as a matter of course.
I can look frazzled and dilapidated for the photos, will that do? :P.
I think Olmsted and Kirby need to be called out on their fixation with Chinese mercury—the echoes of suspected “yellow peril” in their writing are lurking.
I agree Kristina. Its becoming a disturbing theme – all these middle class white males sitting around blaming dirty wombs and dirty Chinese.
The R-squared statistic is known as the coefficient of determination: an indicator of how much of the variance in the dependent variable is accounted for by a straight-line fit – essentially what is seen on the graph being referred to. Variance is a measure of how much the data actually vary around the predicted/expected values.
An R-squared value of .9954 indicates a close match to expected/predicted values.
David, I think you’d need to use smaller words for Dan!O.
🙂
And, to be fair, me too. This is one area Dan O and I are in complete agreement ;o)
I think the good R-squared fit of a straight line is temporary, though. It has to be. Autism caseload obviously can’t grow for ever at a much faster pace than population growth. If it continues to grow like that 5 years from now, I’m seriously going to suspect some sort of virus causes autism. That or diagnosticians don’t know what the heck it is they are doing.