[deltaninenine]

It is true that in science nothing can be proven but this fact doesn't apply if you think in terms of statistics. Science is broad and doesn't always give answers in terms of a boolean falsification. You can be given a ratio as the result.

This result is usually A probability established to a degree within the context of a sample size. This is the closest thing to proof that science can offer and we just go with it.

EBM goes a bit further then just a probability in the context of a sample size. There's a further degree of rigor in the qualitative sense that the probability is causal and not correlative. It's subtle but causality is even closer to proof then a correlative study.