"Peer reviewed" is a better assurance of the accuracy of the observation. It can't do anything about the fact that evidence for current treatment attempts is purely based on observations that could be the result of hope and the placebo effect.

Furthermore, there is a difference between observations of aggregate numbers, like the number of people on ventilators and their survival rate, and small numbers of uncontrolled drug interventions where there are obvious ways for the data to be deceiving. Not all observed data has the same reliability.