1 Prospective, rather than retrospective
design
2 • Experimental rather than observational
3 Controlled, rather than uncontrolled
4 Contemporaneous control groups
5 Internal control groups rather than external ones
6 Allocation concealment
7 Randomized assignment of patients
8 Blinding of patients
9 Large enough sample size to detect true
treatment effects
10 Minimal patient drop-outs or loss