Moher Lancet 2001
“... (to avoid) biased estimates of treatment effect,
... to judge the reliability or the relevance of the findings”