It is generally true that whenever some intervention doesn't show an effect, it is possible that the intervention was done incorrectly (wrong drug dose, wrong patient population, etc, etc.). However, the conclusion stands that the intervention "as given" was ineffective. The problem of RCT vs observational data will never be answered to everyone's satisfaction. Real world biases vs difficulty in doing RCT studies means that neither will be the perfect gold standard. We have to make judgements as to cost / benefit on a case by case basis.
It is generally true that whenever some intervention doesn't show an effect, it is possible that the intervention was done incorrectly (wrong drug dose, wrong patient population, etc, etc.). However, the conclusion stands that the intervention "as given" was ineffective. The problem of RCT vs observational data will never be answered to everyone's satisfaction. Real world biases vs difficulty in doing RCT studies means that neither will be the perfect gold standard. We have to make judgements as to cost / benefit on a case by case basis.