ABSTRACT
Introduction There is nothing like a little pressure to brighten up one’s day, and this
one was no exception. I had arrived as usual at the clinical pharmacology unit and was at work preparing a study design proposal when my boss walked into my office. She had run up the stairs. My office by this time (about two years
after I had hired on) was one floor up and well away from my clinical colleagues who tended to be a bit noisy and nosy. The first was no problem (get some earplugs), but the second is irritating for a working statistician. They were always stopping by for just a ‘peek’ at the data, but they were full of questions. Answer one and a dozen more pop out. After about two years, one figures out that a little distance is not a bad thing. After she had got her breath back, she told me that one of my col-
leagues from Pharmacokinetics had a bioequivalence data set that needed to be looked at ‘Stat’ (an expression the clinicians used all the time). I’m guessing that ‘Stat’ in clinician-speak means ‘run the Statistics as soon as humanly possible’. I guess they like to think that we sit around twiddling our thumbs unless they shout ‘Stat’ repeatedly. There is one certainty in drug development and statistics that one
can depend on: the data are always late. There are always reasons that someone wanted to know the findings yesterday. Sometimes it is even a good reason! Like the art of Statistics itself, after you get used to it, it does not
bother you too much. In any event, it was 10:30, and the results were needed by lunchtime.
After making sure she meant a late lunch (she did not), I hastily pulled the code you will see later in this chapter, grabbed the data, and went to work. We did have a late lunch that day, by the way. Analysis of bioequiva-
lence data is not as simple as pressing a button.
