Interesting Applications from Three Decades of Biostatistical Consulting

Details of several interesting applications from 30 years of biostatistical consulting are presented. One application involves using a parallel design instead of a crossover design in a bioequivalence trial of several formulations. This trial also incorporated blinded sample size re-estimation—before there was a literature on the subject. Two other applications, one of a fixed combination drug for treating allergic rhinitis and the other of a dose comparison trial in duodenal ulcer, reflect using bivariate plots of two primary response measures to illustrate simultaneously drug/dose effects. The dose comparison duodenal ulcer trial also reflects logistical aspects of interim analyses procedures aimed initially at dropping the placebo arm and later dose discrimination, as well as illustrating the need to carefully define the trial objective. The last example provides an overview of assessing whether evidence exists from two Phase II trials of angina to support conducting a Phase III trial at either a b.i.d. or t.i.d. drug.