18 Chapter 2: Questions for planning trials Moral of the saMe--Cont'd prevalence. For example, knowledge of existing and potential treatments and their costs is crucial. Also important is the natural history of the disease. For example, chronic illnesses typically require much more treatment over a patient's lifetime than acute ones. Also important is patent status of current treatments, informing on how soon cheaper generic treatments may be available. Then there are politics to be considered. When one proj- ect gets more resources, others will get fewer. This is likely to make some people unhappy. And unhappy people can be problematic for an organization, particularly if they are important people. Focusing on the scientific aspects of CTC-11 implies man- agement's decision would be independent of the drug's overall development plan. Well, it cannot be. In deciding whether to go ahead with an early-stage trial, management must also consider the steps required beyond this initial trial. And while development plans are imperfect tools, surprises being the rule in the development process, companies must start their planning somewhere. So, in fact, you were proba- bly designing (theoretical) future trials long before you heard management's decision on the first. No clinical trial can be considered in isolation. I thus suggest that when reading this book you also be aware of what is not mentioned. Indeed, I propose that you exercise your clinical trial mind by raising issues beyond those discussed and seeking solutions for them. It is a worthwhile exercise. In routine product development you will be asked to provide many answers and to juggle multiple factors throughout. So if you have ever dreamed of joining the circus, welcome. I began Chapter 1 with a scientific presentation to management because one needs to start somewhere. I could have just as easily chosen another point in the devel- opment process. But while "the beginning" is to some degree an arbitrary concept, there are better places to start than others. And the question of sample size--that most associated with my own profession--is typically not one of these. Over the years I have been asked for sample size estimates very early on in the trial planning process. More often than not I find that these requests had come much earlier than relevant. While I cannot be sure why this is, I have my suspicions. First, estimating the number of subjects needed for a trial goes a long way in estimating overall length and cost. Thus, it is certainly reasonable for anyone thinking about a study to want to know something about the efforts required early on. Additionally, there may be some expectation that the statistician and his mathematically driven profession will provide some definite answer at a stage where so few are to be had. Or perhaps it is simply a matter of involving another individual from beyond the pale when no handy answers are available in one's immediate professional circle. Be that as it may, R&D in the industry typically raise the issue of sample size long before there can be a meaningful answer for it. You see, I can only compute the number of subjects needed after knowing something about the product, the trial's design, endpoints, and expected effects. And obtaining all of these usually requires a great deal of preliminary study and discussion. Still, being asked the question early on gives me the opportunity to stick my nose into nonstatistical issues as well. I like this. More importantly, if you