For studies with subjective endpoints, was the control really adequate.
amjmed.com/article/S0002-…
Always be cautious when surrogate endpoints are used.
thelancet.com/journals/lance…
Consider the “pre-study” probability that the treatment is effective.
jamanetwork.com/journals/jama/…
When the results of a trial only barely reach statistical significance, consider the fragility index.
jamanetwork.com/journals/jamas…
Remember that trials terminated early systematically overestimate the benefit of the intervention.
jamanetwork.com/journals/jama/…
Never forget about publication bias and p-hacking.
ncbi.nlm.nih.gov/pubmed/22853650