n=5, success rate 60%, not blinded; but it does appear that they already had reason to test the treatment, so it's not a green jellybeans situation.