Studies should be controlled, randomized, and blinded.

Rigorous assessment of the efficacy and safety of a new therapy requires a concurrent control group, randomization, and blinding. The control group allows comparison with a contemporaneous patient group. Randomization, properly done, avoids systematic bias between treatment groups and should balance other factors, reducing the likelihood of group differences due to patient characteristics. Blinding minimizes the risk of biases stemming from patients' and assessors' beliefs, actions, and hopes about the treatment received. This is critical given the substantial effect that such beliefs can have on outcome, often termed the "placebo effect." New and exciting technologies would probably be particularly susceptible to the placebo effect, making blinding even more critical when studying such treatments.