An Interview with Dr. Bob Nelson; Atlanta GA; How Doctors Think August 15, 2013

I think you are correct as to the influences that shape doctor’s thinking.  Koch’s postulates of scientific reasoning and causes disease are very important.  More importantly, where does all the didactic teaching leads to as far as the thought process? 

When evaluating a treatment of new drug for example... most doctors want to know the treatment or intervention interacts or influences normal physiology, as well as how it impacts or alters the disease state.  What they are told or read, must jive with the current accepted understanding of the disease process.  The result of the treatment must be measurable with a verifiable outcome.  The outcome can take many manifestations, but mostly we want to see a measurable advantage in a patient’s health and well-being the “acid test” of efficacy.  The more anecdotal verifications of data that we receive in the form of the positive patient feedback, then eventually we validate the data as “real”.  Equally important is the therapeutic index, which is essentially the ratio of a good outcome compared to unfavorable or side effects; expressed as the toxicity dose 50 / effective dose 50.  The higher the ratio the better. 

 

The other issue that I (and many doctors) look at is the absolute vs. the relative effect of an intervention.  To determine these values, it is essential to know the prevalence of a disease or condition in the population you are studying.  This is a very important concept because most pharmaceutical data is presented as “relative risk reduction/benefit increase” as opposed to absolute risk reduction or benefit.  Knowing the absolute risk/benefit numbers is essential to determining the idea of the “number needed to treat”, which is essentially the number of patients you need to apply a treatment or invention to in order to have one good outcome  (or prevent one bad outcome).  For example, the relative treatment benefit of two different treatments for unrelated conditions might be 50%.  However, if the prevalence of condition “A” is 1% and the prevalence of condition “B” is 10% then I only have to treat 20 people with suspected “B” to benefit one person.  But, I would have to treat 200 people to benefit one person if trying to intervene in condition “A”.    So, if the prevalence of the condition is very low in the population, even a treatment with a very high relative benefit or relative risk reduction could require you to treat a very large number of patients in order to benefit one person.  This must be balance, of course, against the severity of the condition you are treating: mild, moderate, life-altering, or life-threatening, etc...