These are then utilized by government, health professional organizations, and insurance companies to develop "Clinical Parameters" which are considered the "gold standard" for organizational and individual practice.
There is no question that critical health care decisions must be based on valid evidence. However, it is well known that clinical trials, and hence evidence based parameters can be flawed. As a Principal Investigator for many clinical drug trials, I have observed that "Parameter Selection" and "Type 2 Error" are major factors creating these flaws.
Parameter selection involves deciding which "dependent" factors may change favorably as a result of an "independent" factor" i.e. new therapy or drug. For instance, when studying asthma one might select:
1) lung function
2) symptoms
3) quality of life indicators
4) use of rescue medication
5) frequency and
6) severity of attacks
7) major attacks leading to ICU or ventilator management, or
8) disastrous outcomes such as death.
In this scenario a certain drug may show statistically significant improvement in the first five or six parameters but not the last two or three. These last parameters are relatively infrequent events and hence require a greater number of subjects and/or a longer study to measure either positive or negative outcomes. This is the concept of "Type 2 Error" - not seeing an effect due to inadequate sampling volume.
The problem here is that a drug may be approved based on data showing significant clinical trial benefit in some parameters expected by the manufacturer to be positive, yet show no difference in other parameters which may be more significant or be negatively impacted if the sampling size was greater.
We often see this effect after a medication is released and used by many more patients, i.e. "Post Marketing Experience" or when a drug is subjected to a much larger trial i.e. 30000 subjects instead of 300. That is when "black box warnings" of death or permanent impairment are posted, or a drug is taken off the market.
