You are here

Contemporary Legends - How To Lie With Statistics I

Daily News Navigator

Here we go again.  HRT, knee surgery, Vioxx ,Celebrex.  Every day, it seems, there is another revelation about an accepted medical practice which has been based on poorly designed, industry-generated research. The average person can't help but wonder who to trust, how to know whether or not the recommendations their doctors are making are based on valid science or drug company hype.

Unfortunately, there is neither a magic bullet available to cure our many ills nor a bible available to answer our numerous health questions. Unfortunately, there is no substitute for learning to challenge the information we are fed.  Nothing short of learning to question and analyze it ourselves, of reading between the lines, will do.

May we at least rely on our doctors to sort it all out for us?  Clearly not entirely, given the institutionalized influence of  possibly conflicted groups like the American Academy of Pediatrics and obviously conflicted ones like the drug companies. (There is even a website challenging doctors to "Just Say No To Drug Reps".)

The truth is, we must learn to rely on ourselves in health matters, whenever and wherever possible.

Thus, the first order of the day is to understand how easy it is to "lie with statistics". 

Just this week, for instance, was an excellent example of how something might be passed on and accepted as fact which on further scrutiny doesn't pass muster.  And how easily such things can slip by even the most jaded and critical eye.

Suggesting that many parents might not choose to accept even the allegedly low risks associated with vaccination, were they aware of them, Christine Colebeck, in her eloquent and sorrowful testament to the loss of her first-born child, "Death by Lethal Injection: From the Heart of One Mother to Another",  rightly challenges the underlying meaning of official statistics on DPT vaccine injury and death when she asks:

"If the odds are 1 in 500,000 for death, 1 in 100,000 for permanent brain injury, 1 in 1,700 for seizures and convulsions or one in 100 for adverse reaction, are you willing to take that chance?  Are any odds acceptable enough to convince you to gamble with your child’s life?"

Hers is a powerful argument for full disclosure and choice, i.e., "informed consent".  As the National Vaccine Information Center has long stated, "When it happens to your child, the risks are 100%". 

Notwithstanding the known problems with under-reporting of adverse drug events and the fact that the vaccine manufacturers control the research agenda, however, beyond Christine's salient point lies a more insidious problem, that is, how statistics, accurate or not, are represented. 

Quite simply, there are compelling reasons to believe statistics such as these represent, at best, underestimates of real injury, and at worst, deliberate attempts to mislead.  For instance, one tricky way in which the risks of vaccination can be incorrectly estimated is by using the number of doses of vaccine which appear to result in injury or death, rather than the number of children that are administered however many doses it takes to injure or kill them.  Since nearly all vaccines are currently recommended to be administered in multiple doses, using doses rather than children can result in gross underestimation of actual risk. (We also cannot ignore confounding which occurs when high-risk children are eliminated, either by not receiving the vaccine in question at all, or by receiving fewer than the recommended number of doses.)

Take as an example convulsions resulting from whooping cough vaccine vs. those resulting from whooping cough itself.  According to the CDC and  the National Library of Medicine, 1 in 1750 vaccinations may result in convulsions while according to the CDC  "as many as 2 out of 100" children who get the disease may have convulsions.  (According to WHO, it is estimated that 1 out of 12,500 doses result in convulsions.)  If the first figure is correct, and we divide the 1750 vaccinations by the between 4 and 5 doses children are required to get, the result is 1/350 to 1/438 children getting convulsions after whooping cough, not nearly as dramatic a difference.  If the latter one is accurate, it's between 1/2500 and 1/3125.

If we further try to factor in the impact of underreporting of adverse reactions, the actual incidence of which is unknown but presumed to be significant, it becomes clear that there may be no difference, and in fact, that it is possible convulsions are more likely to result after vaccination than after disease. 

The fact is, however, that we do not know the true incidence of vaccine adverse reactions (the wildly differing official estimates of pertussis vaccine-related convulsions alone providing stark evidence of that), of whooping cough itself, of convulsions after whooping cough or many other relevant and critical factors, including the actual number of children receiving a vaccine once high-risk children have been removed.  We should simply admit it and set about trying to learn what we can.  We should not, however, be issuing reassuring assessments of vaccine risk.

And yet we are continually being issued reassuring assessments of vaccine risk.  Reassurances based on faulty data and misleading analysis, however, are the stuff of which legends are made.

Shouldn't we begin to question and challenge what we are being told?

(For more on the issue of the underreporting of vaccine-associated adverse reactions, click here.)

Sandy Gottstein  

Date: 8-16-2002