Bioequivalence and Statistics in Clinical Pharmacology

Scott D. Patterson, Byron Jones

• Publisher: Chapman and Hall/CRC
• Year: 2017
• ISBN:
  9781315356600
• Edition: 2
• Topic: Mathematics
• Subtopic: Probability & Statistics
• Index: Mathematics

Maintaining a practical perspective, Bioequivalence and Statistics in Clinical Pharmacology, Second Edition explores statistics used in day-to-day clinical pharmacology work. The book is a starting point for those involved in such research and covers the methods needed to design, analyze, and interpret bioequivalence trials; explores when, how, and why these studies are performed as part of drug development; and demonstrates the methods using real world examples.

Drawing on knowledge gained directly from working in the pharmaceutical industry, the authors set the stage by describing the general role of statistics. Once the foundation of clinical pharmacology drug development, regulatory applications, and the design and analysis of bioequivalence trials are established, including recent regulatory changes in design and analysis and in particular sample-size adaptation, they move on to related topics in clinical pharmacology involving the use of cross-over designs. These include, but are not limited to, safety studies in Phase I, dose-response trials, drug interaction trials, food-effect and combination trials, QTc and other pharmacodynamic equivalence trials, proof-of-concept trials, dose-proportionality trials, and vaccines trials.

This second edition addresses several recent developments in the field, including new chapters on adaptive bioequivalence studies, scaled average bioequivalence testing, and vaccine trials.

Purposefully designed to be instantly applicable, Bioequivalence and Statistics in Clinical Pharmacology, Second Edition provides examples of SAS and R code so that the analyses described can be immediately implemented. The authors have made extensive use of the proc mixed procedures available in SAS.

Part III

Clinical Pharmacology

8

Clinical Pharmacology Safety Studies

Introduction

One day, out of seemingly nowhere, I received a very strange request from a clinical scientist. We will call her Betty, and she asked if I could round off a confidence interval. My immediate response was, ‘No. Why would anyone want to do that?’

In essence, we had derived an upper bound in a drug interaction trial of 1.2538 for AUC. Evaluation of this value relative to the acceptance level of 1.25 showed that it was higher than 1.25. We could not conclude the two treatments were equivalent. Pretty elementary. Betty wanted to round it off, so she could claim equivalence had been demonstrated.

I told her no, and left it at that. Such would misrepresent the data, and the statistics underlying the upper bound could not support “rounding it off.” Clearly, as the value was higher than 1.25, the null hypothesis had not been rejected, and it was out of the realm of possibility. To my mind it was also a matter of professional integrity, and I was a bit surprised that anyone would ask such a thing. The less I said, the better off we would both be.

However, I was still new on the job, and did not know that some people will not take no for an answer, even if it is a matter of professional integrity. So began one of my most important “learning experiences” on the job. “Learning experiences” are a business euphemism for an experience no one in their right mind wants any part of, but you are stuck with it because you work there.

Rounding off turned out to be really, very important to Betty and the physician for whom she worked, and a major disagreement at the company developed. Peoples’ egos became involved, and everyone who had even only a nebulous stake in this (or a potentially related) issue felt compelled to comment. Academic experts were paid and consulted. Opinions were sought from the FDA on the topic. Many internal meetings on the topic were held, and (despite their best efforts to avoid it) several senior vice presidents had to be consulted and in the end backed us up: “No rounding.”

Years later FDA guidance [373] was issued saying the same thing, but, as is often the case, such business precedes regulatory guidance by many years.

Guess who was at the center of this argument? It was a rough experience (for what I still feel was a ridiculous request), but I learned a lot from interacting with such people on such a thing and from watching how they and many other people behaved. If I had it to do over again, I would have followed a different approach to dealing with such people. I call it the “Nurse” approach in honor of the people whom I saw do it.

We had a drug intended for the treatment of hypertension (high blood pressure) which caused migraines if given at high doses. We discovered this in the first study in man (which is designed for this purpose, see Section 8.1), and carefully worked out at which dose the problem started. These were bad migraines — the throwing-up kind. The study team wanted to stop the study, but a chief medic said to continue. The rationale was that they wanted to explore more doses before going to the next study.

There was no point in continu...