Science versus medicine

Skewed statistical links
     Do people who wear seat belts have a higher incidence of lung cancer deaths?
     Because people who wear seat belts might be more careful generally and less likely to, for example, smoke, their data set might have less incidence of lung cancer deaths.  Then it would seem statistically that seat belts prevent cancer.
     Or rather because some of the people who do not wear seat belts die from car accidents instead of getting cancer their incidence of lung cancer deaths might be lower.  Then it would seem statistically that seat belts cause cancer.
     Or a study might by chance have a balance of those two and show the rate of lung cancer deaths is the same whether seat belts are worn or not, that there is no statistical link.  And while seat belts might prevent deaths from car accidents they have no effect on lung cancer.
     So in the third case we might have the correct answer, that there is no statistical link betwen seat belts and lung cancer, but we arrived at it by chance, not science.
     In fact scientific experimentation requires isolating a cause, knowingly removing all the other possible causes, in order to establish a scientific link to the effect.
     In can be established that heat causes water to boil.  In order to do this you take the same water, apply extra heat to half of it in one container and no extra heat to the other half in another identical container.  The half that got heated boils.  It might be argued that there was something else in one half of the water that failed to be distributed to the other half of the water, that the two halves were not really the same.  Salt for example has an effect on the boiling point of water.  Pressure has an effect on the boiling point of water.  However those other causes have been successfully isolated out of the experiments.  It is a fairly simple matter to ensure that the salt content and pressure are essentially the same.  Then a very specific amount of heat "causes" water to boil.
     A persistent problem with medicine is isolating causes.  With many medical questions there are far too many possible causes to ever isolate even most of them out.
     As simple a question as whether aspirin is an effective pain reliever is impossible to answer with anything that can be described as "scientific" evidence.  Rather what you have is "statistical" evidence.  There are countless factors that simply cannot be isolated out.  You can hope that those factors balance themselves out as they might in the seat belt example, but that is not science.  What you will likely find is that the effect of aspirin is so small that the "noise level" of other factors can easily hide it.
     When all you have is "statistical" evidence basing a decision on it can be sound practice, but a hazard lately is that too many people believe they have scientific certainty when that is quite far from the truth.
     You probably know that studies recently released say that cancer screening might not be as useful as previously believed.  In reaction there have been some rather odd claims.  One lady doctor said she "knows" she saved lives.  That is not possible.  She strongly believes she saved lives and maybe she did save lives, but she really doesn't know it, can't know it, not about cancer.  There is no such science.
Other extremes
     While it is true that science cannot "prove" whether seat belts cause cancer, there is no need to abandon the common sense of it.  Seat belts likely have no effect on cancer.  We can say that without the science.
     My best anecdote on the matter is about a small wound on the side of my arm at the elbow; that is, where it was likely caused by the bite or sting of an insect or spider.  It happened when I was a small boy.  I only noticed it at all because it itched later and I scratched it and it bled for a moment.  As soon as it quit bleeding I quit paying attention to it.  That might be the nature of boys.  But much later the same thing happened.  And again even much later.  The time between occurences was so long that I did not notice it was the very same spot, the very same wound, till at least the third round.
     It startled me that it wasn't really healing so I burned the suspicious denser portion of it with my soldering iron, which I cannot recommend to others.  For one thing it requires some ability to ignore pain.  As an adult I believe that a doctor should have been consulted.  Luckily, and I can only say luckily, the wound healed once and for all, and without a noticeable scar.
     I can't say that I prevented a more serious problem from arising later, I can't know that.  I can't be certain it won't come back again even after all these years.  I took no samples, conducted no tests.  Perhaps freezing the tissue would have caused even less scarring, but I didn't have access to liquid nitrogen.
     Recent health care reform perhaps intended by Obama to restore "science" to its proper role has failed to grasp the mathematics in insurance or depend on any other science at all.  I am happy that my wound finally healed.  If I had it to do over I would probably expect a doctor to remove it and with the least amount of scarring for a reasonable fee that I would arrange to pay.  But I can't see making it a law that anyone must do that or pay for others to do that.  It just isn't scientific enough.
The necessity of practicality
     When the temperture outside is over 95 degrees I prefer to work in the shade, but that isn't always practical.   To be practical there are chances that have to be taken, roads we must go down, and all the while knowing that there might be some small chance of dying from it.  Those chances are everywhere, all around all of us, not just with cancer.  We can abandon some things; the silly, the inconsiderate, the invasive, but some things cannot be avoided altogether in a practical way.