1
2	Common Probability Models Defining the study structure One arm trial: Historical or matched controls comparison to absolute reference Two arm trial: Placebo or Active controls comparison between arms Multiple arm trial: Several treatments or doses global tests vs pairwise comparisons Regression trial: Continuous dose response test slope parameters Also need to consider randomization ratio across treatment arms (or distribution of dose levels)
3	Probability Model: Comparison Groups No comparison group Appropriate when an absolute criterion for treatment effect exists Single arm clinical trial Cohort design (Rarely do such absolute criteria exist. Instead, we are really invoking the use of results from previous investigations.)
4	Probability Model: Comparison Groups Historical controls Single arm clinical trial Compare results to Absolute criterion derived from historical trials Sample from historical clinical trial (better)
5	Probability Model: Comparison Groups Comment re use of single arm trials “There are only two types of researchers: those with a lot of enthusiasm and no controls, and those with a lot of controls and no enthusiasm.” (unknown)
6	Probability Model: Comparison Groups Internal controls Each subject serves as his/her own control Different treatments at different times Different treatments for different parts of body eye diseases, skin diseases
7	Probability Model: Comparison Groups Concurrent control group Two or more treatment arms Placebo or standard therapy Active treatments Sometimes consider equivalence Multiple levels of same treatment Stronger evidence sometimes obtained from dose-response Identifying optimal dose