Increase of control - loss of relevance?

In determining the strength of evidence, the accepted progression of study strength from low to high is: Editorials, Case Studies, Case-Control Studies, Cohort Studies, Randomized Controlled Trials, and finally Systematic Reviews.  Actually, I never even thought of Editorials before I found this snazzy pyramid. As you go up the pyramid, except for the Systematic Review, the increased strength of evidence is due to increased control over how the population is sampled and the independent variables.  These two in combination increases the internal validity of the study.  That is, the increased confidence that the results found in the study were the not due to unintended bias or chance but were actually present.

However, that same control that increases internal validity decreases external validity.  It is increasingly unlikely that the population will exhibit the same characteristics as the most controlled experiment.  The results will cease to be relevant.  The rub is that each research question or problem to be solved will have a different breaking point where the lack of relevance outweighs the benefits of control.

For some research questions, the RCT is the ultimate research design, as it can be reasonably assumed that human physiology is consistent enough to make the findings externally valid to the population. However, understanding of systems or human behavior cannot be reliably reduced to a series of independent and dependent variable relationships.