Perhaps your definition of "experiment" is broader than mine, but to me, an experiment is a specific kind of activity involving tightly controlled conditions where only very specific variables can be adjusted. However, as I mentioned, you generally can't do such experiments on people. The core of research on human health involve other kinds of study designs (look at the whole field of epidemiology), and the analysis of those results also includes hypothesis testing. My point is not about the necessity of having a falsifiable hypothesis, but on how to go about obtaining the evidence to disprove these hypotheses.

So, either we have a definition of science that includes other methods of investigation besides a strict experimental design, or we concede that much of human medical knowledge is not based on "science," which perhaps is a view you subscribe to. (Although unorthodox, a case can certainly be made for that view on a number of points but that's another discussion...)

I agree that hypothesis testing is at the core of the scientific method, and we both concede that scientists can be wrong. There are certainly many people who complain about how researchers flip-flop about various topics when it comes to human health, such as various dietary guidelines. Eggs are good. No, eggs are bad. No, wait, eggs are good. Butter? Margarine? Soy, good? Soy, bad? Hormone replacement therapy is good. No, hormone replacement is harmful. Mammograms every year starting at 40. No, mammograms not necessary until 50, and then only every other year.

To the dismay of average folks, that happens quite a lot with human-related research and one reason is because we can't conduct experiments on people like we can with lab rats. But that's just the meandering path that science takes on its quest for Truth.