Biotech Values

[iwfal]

Emil Kakkis has been a big advocate for an adaptive trial design (in particular for rare diseases where populations are small and efficacy could be small/take long to show significance). Perhaps its time to explore these/other changes to speed development?

"Adaptive trials", like Bayesian Analysis, is a buzzword that implies some secret sauce but the reality is there generally is very little there. Running a quicker trial campaign involves, for instance, less gap between trials. Adaptive trials in theory directly address this - but the reality is you generally want to change the protocol between trials in difficult-to-predict ways. And that time to analyse what change you want, determine how to roll it out to the sites, cut off old enrollees, ... all takes time. Trying to avoid that either hampers your ability to tune the final protocol or, if you do take the time you blow the supposed savings provided by the Adaptive Trial.

See SNTAs trial as a good example of the Scylla v Charybdis problems with ATs. Or see BMRN for example of how to run trials quickly without "AT" paradigm.

Finally note that AT paradigm can work if you are highly confident you will not want to change anything more than planned trial size. And perhaps there is some benefit if you are absolutely confident you know in advance all the possible trial protocol decisions you'll make at the decision point and it is a small number. Otherwise it is just a buzzword - generally used by academics or MBA types (with something to sell).

The summary point is that if a company is slow at running trials it is almost certainly more about the management or the difficulty of the indication/drug. Not something magically fixed by a different trial paradigm.