Ability of a meta-analysis to prevent redundant research: systematic review of studies on pain from propofol injection