New Computer Fund

Friday, September 11, 2015

Cherry Picking your Science

With the explosion of the information age there is an explosion of "scientific" references.  Since everyone has an opinion and access to a computer there is a weird blend of scientific fact and fiction floating around.  We are becoming a nation of cybercondriacs and "natural" crusaders.  Sometimes all natural isn't what its cracked up to be and some "science" isn't all that scientific.  

The Heimlich Maneuver for drowning is a great example of "hit" science.  The good doctor Heimlich created the Maneuver to clear airways of choking victims.  He then extended the maneuver to drowning to clear airways.  When he described the maneuver for drowning he misrepresented the workings a tad.  Typically, drowning victims do not inhale lungs full of water.  Instead there is a natural involuntary spasm that closes off the airway so less than a teaspoon of water, especially salt water, can cause you to stop breathing.  An organization that provides lifeguards for water parks used the newly recommended maneuver and found it to be extremely effective, especially since it was simple enough to teach teenagers to use and didn't involve mouth to mouth reparation which can cause vomiting which is pretty gross for your average teenage first responded.  

A son of the good doctor made a lot of wild claims about his father and inspired a medical researcher to "scientifically" determine whether the drowning version of the maneuver was effective or not.  Follow this link for medical fraud info. The "scientist" found one example where the maneuver could have cause harm to a victim out of less than 30 cases where it was used.  That particular victim aspirated vomit.  Since there was a very small sample population, a bad result was "significant".  In normal science, the one fail in a small sample should have inspired the typical, "more research is indicated".  Basically there was not enough data to reach a meaningful conclusion.  However, due to the lungs full of water miscue, a lot of major organizations drop the Heimlich Maneuver for "near" drowning situations.

According to one water park lifeguard provider, the maneuver was used many thousands of times as "in water intervention" and potentially saved many hundreds of lives.  That study though merely indicated that the Heimlich Maneuver for "near" drowning "may" work but didn't appear to be harmful.  What should have been a signal for more research or no action resulted in the Heimlich Maneuver for "near" drowning victims being disapproved for basic rescue training. Follow this link for more on the controversy.

A second example is azodicarbonamide (ADA) which is an additive used in various food products with Subway breads being the big media event, though Starbucks also has it as an ingredient in some products.  ADA is a flour conditioner that increases shelf life, increases elasticity and promotes rise.  It's main use is in plastic foams like yoga mats, but the US FDA allows its use in food products.  European countries have banned its use due to indications that heavy exposure to the powder appears to cause asthmatic health issues.  The additive has been found to harm laboratory animals in concentrations of 5% and greater, but is approved as a food additive in concentrations of 45 ppm or less, 45/1000000 versus 5/100 is a huge difference.  Without any real indication that ADA is actually harmful as a flour conditioner, Subway and others are removing the additive due to popular interpretation of "science".  Ironically, ADA is an additive in many "100% whole grain" products with healthy plastered on the packaging.  

A third example is Roundup brand herbicide.  The active ingredient in Roundup is Glyphosate but the formula for Roundup has a number of proprietary ingredients including a wetting agent  polyethoxylated tallow amine (POEA).  One study aimed at genetically modified crops (GMC) also include glyphosate in different concentrations in the water of the Sprauge-Dawley rats used in the study.  One of the groups of rats given glyphosate in their water actually had less symptoms that the control group rats.  

All three of these cases represent current issues with the state of science.  When there is very little definitive evidence of harm, the noise or inaccuracy of specific testing methods can result in false positives that once linearly extrapolated appear to be more significant than they are.  The noise can be confounding factors, other real world impacts or just random chance.  

The Sprauge-Dawley Rats for example have a genetic tendency towards tumors in long term (~2 year) studies.  So a control group having more or about the same tumor growth rate as the target groups is not all that unusual.  If there is a definite harmful positive in the target groups it would have to be much more obvious or "significant" to be believable.  

A second issue is "fishing".  When an experiment is designed with several possible active elements and no predicted result, just a "let's see what happens" attitude, the results cannot ever be considered definitive or even valid.  They would indicate the possible need for future more specific testing.  

Confounding factors can sometimes be so obvious after the fact you cannot believe they were missed.  For example a study found that rural cardiac cases resulted in death more often than urban cases.  The study allowed "fishing", so diet, lifestyle and genetic traits appeared to be significant causes.  Well, emergency response time is quicker in urban setting and takes a lot longer in rural settings.  The further you live away from a hospital or emergency medical service the more likely you are to die from various emergency conditions.  

Media exposure amplifies these not quite significant findings to near catastrophic situations.  Every scientific study requires time to "prove" its worth.  There is no absolute "proof" but they can prove to be useful.  They can also be proven to be bull crap.  Almost every scientist needs recognition for their work to get funding for more work so they tend to enhance their press releases.  There is very little old fashion "science" for the sake of science because almost all scientists are not independently wealthy.  Some government and corporate grants allow for pure science but since there is money coming from somewhere some group can always accuse the scientist of being in the pocket of some entity they don't trust.

Add to that a small percentage of scientists who cheat to enhance their image or take shortcuts and another percentage that are just plain nuts, just like the general population and you end up with the old let the buyer beware caveat.  

My personal favorite though are studies that estimate the health cost saving of a particular government regulation.  Take Asthma for example.  Asthma isn't A condition it is a variety of conditions that can lead to wheezing and difficulty in breathing.  Smoking, second hand smoke, smog related to industry and transportation, rag weed, cleaning solutions, wood smoke, coal smoke, perfumes, artificial additives, natural ingredients, pine fresh scent both natural and artificial, dry air (low humidity) etc. etc. etc.all can cause of contribute to an asthmatic symptom.  Currently the government has a hard on for coal so more coal regulation will reduce asthma by so much.  Nice.  Asthma hot spots though are more closely related to diesel emission near heavily use ports and rail yards used in transporting containerized goods.  No matter what is done, a small percentage of the population will always have asthmatic reactions to perfectly natural sources.  So regulation that impact asthma will not produce zero cases, just a reduction in overall cases.  If you don't know what the baseline or "normal" rate is, you cannot accurately predict how effective the mitigation will be.  If you consider every regulation ever produced to reduce asthma with their original projects of impact, we would be in negative asthma territory. 

Flu vaccination is another good one.  Originally the vaccines were estimated to be close to 90% efficacy.  Those studies use very generic indicators of the "flu" and included the effect of the typically co-vaccination for pneumonia.  The pneumonia vaccine is very close to 90% effective but depending on the type of flu strain, the "flu" vaccine ranges from 10% to 50% effective.  In most of the studies less than 20% effective isn't statistically significant because of limits of the types of studies.  So that flu vaccine might help or might not.  If you forego the flu vaccine though you just might be labeled an "anti-vaccer".  Knowing a little bit about science in this case makes you anti-science thanks to a butt load of clueless "causers".  

The best advice I can give is "everything in moderation".  Try to enjoy life while you can so you don't succumb to Pre-Traumatic Stress Disorder or any new excuse of the day.