Tuesday, February 19, 2019

Massaging numbers in a Faustian bargain.

Roundup, the most ubiquitous chemical in agriculture does not cause cancer

 until it does-- shockingly elevating risks of lymphoma by 41%?!!

 

 

AGRICULTURAL HEALTH STUDY (AHS)



Is this the truth?





This? 







OOOOPS. Glyphosate increases risks of lymphoma by 41% ?!


Two dueling studies. AHS vs Meta-Analysis (based on AHS)





Want the truth? 



Radical idea: read the two studies pivotal in the "bellwether" case!

And rather carefully since Monsanto deemed the AHS the most valuable epidemiological study in a class action lawsuit involving thousands of plaintiffs and billions of dollars in damages riding on it.


But don't worry about what's riding on it and who wrote it--science is often simply about doing basic arithmetic, as I've pointed out about here, and here with a different "meta-analysis" It is fascinating how the AHS can be whitewashed in a meta-analysis to lead to an opposite desired conclusion. 



Paul, read the study!





The National Cancer Institute occupational cohort study (AHS)  enrolled two cohorts --applicants for commercial and private pesticide licenses. It also enrolled private pesticide licensee spouses following the cohorts through four phases. 

The latest installment, phase IV was released to great fanfare last year and cited widely by pro-glyphosate advocates as the most comprehensive study on 54,000 (??)  "farmers" proving decisively that Roundup does not cause cancer. 

Commercial pesticide applicants could theoretically be operating out of a mansion in Bel Air with a crew of pesticide "handlers" doing the actual spraying. Private pesticide license holders are primarily conventional farmers likely spraying their fields themselves, and so the degree of exposure to Roundup between the two groups are presumably somewhat different. 

No true control group without any occupational exposure-organic farmers not using Roundup were recruited- there was no unexposed group comparison. And since pesticide licenses are only required for the restricted use of pesticides, and Roundup is not a restricted use pesticide, home gardeners or farmers using a chemical marketed as safe as salt, coffee, and sunshine,  who didn't get safety training are likely exposed to much higher levels of Roundup than anyone in either cohort--but are not represented in this study. 


So flaws are already evident in the basic study design in spite of "expansiveness" of the study enrolling over 89,651 "farmers". The method of original selection is akin to recruiting groups to measure effects of cigarettes by comparing cancer rates of half a pack a day smokers to a pack a day smokers while excluding nonsmokers and three packs a day smokers. 


But forget those methodological flaws- let's just look at the arithmetic! 




"Phase 4 was completed by 60% of the enrolled private applicators 

and 61% of the enrolled spouses"



Please whip your calculators out and check the percentages of people followed up. Let's start with the first row--the private pesticide applicators

33456/52394 x100=63,85%
18599/52394x100=35.50%
17869/52394x100=34.11%
24170/52394x100=46,13%

24145/52394x100=48.08%  -phase IV







HUH!! WHY IS 46% reported as 60%?? 

Maybe it is just a solitary goof up, an innocent arithmetical error?
Let's repeat it in the next row, the spouses of private pesticide license applicants followed up. Because attrition of the male farmers might be somewhat compensated if the spouses were followed up. 

23796/32345x100=73,75%
14391/32345x100=44.49%
15385/32345x100=47.57%
19950/32345x100=61,17%

18186/32345x100=56.23%   

Nope! There it is again! 56% of spouses reported as 61%, the same as in the summary! 

What is so special about 60%? 
Once is an error, twice is an unlikely coincidence, but that weird 60% again, four separate times, National Cancer Institute-- inflating the numbers of people followed up?  It's not a sporadic mistake, it is weird shenanigans deliberately misleading and fraudulent. 




Wonder why?   Let's learn a bit about fundamental guidelines on follow up in cohort studies. 






The study that Monsanto deemed it's strongest scientific evidence--not once, not twice, but on four separate occasions published fraudulent data because epidemiologists consider studies with less than 60% retention rates of limited validity.  


Instead of admitting that losses to follow up introduced significant bias if they were due to Roundup exposure resulting in lymphoma, for instance, they fudged the numbers to underestimate risks of lymphoma by blatantly posting misleading, outright wrong percentages, to avoid raising suspicions. 

If this is not fraud, I do not know what is. 



And it gets better. Epidemiologists understand that nonrandom losses to follow up can magnify bias further, such that even losses of 10% of the cohorts are considered unacceptable if they are differential. Lo and behold, an entire cohort disappears in stages 3 and 4 of the study, compounding bias. 


The missing data and gaps one can drive a Mack truck through are mitigated by "imputation" as the student who wrote the second study, the "explosive" meta-analysis, based on the AHS study mentions, which Carey does not notice. 

Imputation is simply a fancy word for estimates. 


I personally generate fairly tight imputations for my work as a vet and I am often off by about 20% but anyone who's hired a contractor would be thrilled with a 20% discrepancy.  Estimates under the best of circumstances performed in good faith can be done well or very badly, in other words.  






MASSAGING NUMBERS 

Carey's Guardian article manipulates the public using the very same strategy often employed by Pharma to oversell and over-hype ineffective or unsafe drugs--highlighting relative risk without putting it into the context of absolute risk.


97% of average readers never ever read past the headline, 2% read the entire article, while 0.9% click on the original study's abstract, and 0.1% read the actual peer-reviewed study. How many people do you suppose cross-checked the complex computer-generated statistical "imputations" or the statistical veracity of the meta-analysis or understood the sleigh of hand Carey used to misrepresent the meta-analysis?
It's a safe bet that a number of readers are approximately zero-- a statistical black box, as naturally comprehensible, intuitive and accessible as the one holding Schrodinger's dead and alive cat.


Will these shennanigans sway the judge in the first "bellwether" Round Up Lymphoma case? 


I don't think sooo.

The buzz of Carey's obvious PR campaign









Can AHS that blatantly lies about simple arithmetic and easily verifiable percentages be trusted with complicated statistical gymnastics we cannot independently calculate ourselves, be extrapolated further in a meta-analysis, losing accuracy with every layer of assumptions and approximations? 

No one seemed to have asked. They all seem to believe that science is mixing together rotten meat (the AHS study) with fresh peas, carrots, potatoes, apples, and oranges  (the smaller case-control studies) in a crock-pot transforms the spoiled bits of AHS into a fresh gourmet meal--an accurate meta-analysis. 





Sorry to inform you, but our Rachel Carson award-winning book author, our "truth-telling veteran investigative journalist" and "research director" at US Right to Know, with the article in the Guardian-shared over 20,000 times obviously never read the AHS or the meta-analysis incorporating the fraudulent AHS, nor did her partisan "journalist" friends--Michael Balter, Paul Thacker, GM Watch, GMO-Free USA, Food Babe, Food Democracy Now--not to mention  the anonymous trolls recruiting "soldiers" into this "food movement" / US Right to Know Army

Not sure they would understand either if they did, apart from the click-baity bits. 








Sifting through AHS dumpster for nuggets-Hypothyroidism

This journalist ally of Carey's with Thacker in science defense has 40 years of experience in science reporting on fossils-- which in case you missed it, have been dead for thousands of years--bizarrely believes himself qualified in medicine of the living as he comments on hypothyroidism veterinarians see on a daily basis. 






Hilariously he has no idea that hyperthyroidism and hypothyroidism--the most common endocrinopathies seen by most veterinary hospitals on a daily to weekly basis--are polar opposites of each other. 





He clearly couldn't care less if this information ever reaches medical professionals. If it was really about hypothyroidism or lymphoma to advance animal or human health, he'd certainly be reporting on it in JAMA, JAVMA, Journal of Veterinary Internal Medicine, Lancet, New England Journal of Medicine, where endocrinopathies are customarily discussed by medical professionals who diagnose and treat living breathing patients with this all too common disease. 

For Carey's team, there is no difference between science and politics. The AHS gave linking Roundup with hypothyroidism and leukemia so our "science"- defending journalism team will be happy to trade off and compromise on lymphoma and roll the rotting fraudulent AHS over into a meta-analysis that undergoes a spontaneous metamorphosis into a dazzling edifice of scientific certainty that is supposed to sway a jury and a judge that Round Up caused lymphoma. These folks are not Einsteins, methinks.    


None could tell you in person what's in the AHS or the "meta-analysis" --even if they read it--these veteran science journalists failed elementary school math!! 


Striking a Faustian Bargain


The "meta-analysis" is a mash-up of heterogenous case-control studies with a prospective cohort (AHS) study in which the authors emphasizing several points.--the highly unusual methodology,  discarding all data except for the highest levels of occupational exposure to farmers and listing pages upon pages of limitations which Carey obviously missed. 


A meta-analysis is as good as the sum of its individual studies.  No matter how fancy the statistics, you can not build a robust meta-analysis based on studies with weak evidence any more than erecting a house with weight bearing beams and joists of rotting wood on a warped foundation. 

As I contend,  the AHS study was fraudulent, but suppose that it was not.  The authors discarded all the AHS data except for the highest occupational exposures, including in a lengthy study just one sentence Carey wanted for perfect sensationalized click-bait: glyphosate raises (relative) risks of lymphoma by 41%!!  


MAKING SENSE OF RELATIVE AND ABSOLUTE RISKS


The lifetime risk of lymphoma in 2019 is 1/42 in men and 1/54 in women. Which means that according to Carey's team, if a farmer is using glyphosate at a maximum level for a maximum period of time, and has none of the risk factors recognized by cancer experts--their lifetime risk rises from 2.3% to 3.25% in a lifetime.  Epic win for the Anti-Roundup Team? Certainly not a tabloid journalism clickbaity byline. The self-appointed science guardians, unfortunately, appear to lack traits one seeks in a science journalist- curiosity, skepticism, and competence in science starting with elementary school arithmetic. Because figuring out when science is used to manipulate rather than inform does not get much easier than this. 

And for a team this incompetent in science, what are the odds of them being good at judicial chess? 







The approximated risks in the hyped "meta-analysis" only apply to extreme occupational exposure and is absolutely irrelevant to the general public. 

You see why pro-GMO, pro-glyphosate people like it? 

 If a plaintiff has no accepted risk factors for lymphoma- his lifetime risk rose by less than one percent. Brilliant! 





Does Roundup Cause Lymphoma?

The International Agency for Research on Cancer concluded there is sufficient evidence of carcinogenicity in animals and limited in humans classifying it as a probable carcinogen.

Studies generally show a positive correlation between multiple pesticides used in farming, especially and worryingly in children! But the very fact that multiple chemicals are used and the long lag time between exposure and disease makes it very challenging to parse out the effects of a single pesticide considering the complex formulations and cocktails.  

. 
My opinion is, that neither the AHS nor the Carey-hyped "meta-analysis" offer credible evidence that Roundup does or does not cause lymphoma in humans. Neither got us any closer to the truth.  


And an astute science minded judge will see right through the manipulation.  So, in the bellwether case on farmer lymphoma,  the judge unsurprisingly tosses Carey's entire idea of "science and truth" to the curb by bifurcating the case causing spasms of consternation for depriving the jury of "evidence".

 He directs the plaintiffs to first prove causation of lymphoma before anything in Carey's  science-free book of anecdotes and chapters of insinuations called Whitewash that by the way,  garnered a Rachel Carson award from American Society for Environmental Journalism,  documenting in tedious detail the "industry playbook" (and is just one of dozens in that genre  that includes such science luminaries as the Food Babe) is deemed relevant to whether hard science supports the claim that Round Up actually caused  human cancer! 

And when the judge restricts plaintiffs suing for lymphoma to hard science on causation alone deeming your "nuggets of truth" and your book completely irrelevant to the case--you do what your singular specialty has been all along-- you smear the judge's integrity with insinuations. 


Because when you are science incompetent and your career is vested in the outcome of your Public Relations campaign--  insinuations are your entire playbook. 









SF jury awards Sonoma County man $80 Million in Monsanto weed killer cancer case












3 comments:

  1. Interesting. I'm not sure where it's going...

    I suspect Carey Gillam is, suspect RFK Jr is, and know for sure I am, intensely motivated to get glyphosate banned for personal reasons. I suspect their strategy is to grab that eye-catching word "cancer" and wave it about even though they have to know that it's almost impossible to prove anything about cancer.

    This is possible to prove: Exposure to glyphosate makes a majority of living things ill--usually in subtle ways that humans may dismiss as age, tiredness, allergies, etc., but sometimes in dramatic ways that send us to the hospital.

    The more exposure, the more damage. I tried for years to persuade a relative that he "got the lazies" when and only when he poisoned his garden. (I was in his house to see him "getting the lazies" when I "got the sniffles," his wife lost ground in recovering from a stroke, their children had "their problems," etc.) Then one day he poisoned his yard the same day the railroad company sprayed poison on the railroad, and he dozed off in glyphosate-induced narcolepsy and never woke up.

    I myself was brought up feeling the idea of spraying poison is abhorrent, but never bothered about the names of 'cides or what others chose to do until it became obvious that glyphosate was making me much sicker than "allergies."

    Why don't Gillam, RFK and others want to focus on what's provable? More controversy may sell more books & lawsuits but there's also a feeling that the proven effects of glyphosate won't scare enough people enough to get it banned.

    ReplyDelete
  2. Keep pets looking their Best Dog Grooming Tools , supplies, and equipment. We offer a full line of professional quality grooming tools and accessories including brushes, andis dematting tool , clippers, shears and dryers for your convenience.

    ReplyDelete
  3. if you need any kind of anti dog bark app, look no further than the linked app for Android users.

    ReplyDelete