Your source for your last point has the conclusion of:

“Although these biases did not significantly inflate estimates of drug efficacy, reporting biases led to significant increases in the number of positive findings in the literature.”

Which is literally the opposite of what you’re citing on efficacy.