what could go wrong?, page-117

Neo I have a question.

They changed the power of the substudy part way through so there must have been a review of the data at some point?

Would they have looked at the main study blinded data at the same time?

To repower the study would they do a new trial exactly the same and add the studies together?


I was concerned early on about the 400 not a larger size of the study but was told a whole heap about study design that went over my head but what stuck was they were doing it with a specific patient target group in mind and that it was enough. It was the core target group that would provide the majority of the profitable income from actual drug sales based on their market research. Other large companies with unlimited budgets do massive studies that cover a wider group of patients so they get power and success with every patient group be that ethnic, child , male , female , age and so on so their end drug can be used and marketed in a wider group of patients. Great if you have the budget.

Until the unblinded data is processed I guess you don't know if it is underpowered overall or in a particular subset of recruits.

Like you I would hate to see a trial fail as it was XX patients short but that is why you employ the advisers, trial designers and statisticians in the first place. Wish I new more about just how much of the data is available to the people running the study and ethics committee.

Trial design is a minefield and even seeing FDA pre-approval get withdrawn on another company makes the system a expensive one and all you can take is advice.

At least we employed a reputable trial company and their reputation depends on successfully designed and run trials so they wouldn't accept a design with underpowered study number lightly as it would affect their reputation. No idea if there is insurance against bad trail design etc but probably one of those that you could never make a successful claim on anyway. We don't want insurance we want a result.