Most of us are primed to think about the design of adaptive clinical trials as a narrow set of techniques applied to a specific set of problems. If you’re worried about the power of your study, for example, you can turn to your toolkit of adaptive methods and find a suitable use for sample size re-estimation. If the concern is getting the best possible dose, a multi-arm study which drops doses after an interim look seems like the perfect solution.
Clinical development teams often look to adaptive designs, only after identifying specific objectives that a conventional trial may struggle to resolve. However, this approach has its limitations. An adaptive strategy might improve a trial design, even when a conventional strategy supplies a reasonable, (though less efficient,) alternative.
There is, of course, an alternative approach to adaptive trial design. One could begin with the whole toolkit of conventional and adaptive methods, and simply aim to build the best possible trial given the circumstances of the study. We see this mentality reflected in the design of the Thrasos trial, a cardiorenal study which completed interim analysis last September [1].
Thrasos is a biopharmaceutical firm whose drug, THR-184, responds to the market need for a new therapy for acute kidney injury following cardiac surgery. As THR-184 responds to a lacuna in the current market, Thrasos aimed for a trial design which would obtain FDA Fast Track designation. It obtained this designation in February 2014 [2] after defending the following design:
- The Phase 2 clinical study begins with 4-arms and based on all the information collected in this and previous studies, chooses the best candidate dose(s) with which to continue after interim analysis (For example, the DMC eventually decided to continue with one of the original doses, and then to add an an additional dose.) [1]
- During interim analysis, the iDMC determines whether the trial would benefit from a sample size re-estimation [1]
What we may notice here is that whether or not a sample size re-estimation benefits the study, does not necessarily depend on any pre-study anxieties about power or sample size. Rather, once there is an interim look already scheduled (in this case, so that Thrasos can move forward with the right dose without having to run three Phase 2 trials), sponsors may as well consider a sample size re-estimation to determine if it can save development costs. This move optimizes the study design, without mapping onto a specific technical problem highlighted at the start of the clinical development process.
Now it may be that sometimes, an adaptive trial is not the right trial for your study [3]. However, in order to leverage all of the possible benefits of an adaptive trial, it’s necessary to stop viewing adaptive strategies as merely functional responses to easily identified problems. Simply because a process like sample size re-estimation does not seem necessary, does not mean that it will not prove beneficial.
Related Items of Interest
[1] Thrasos Announces Successful Formal Interim Analysis and Clinical Update for Phase 2 Study of THR-184 for Acute Kidney Injury (Businesswire.com)
[2] To Adapt or Not to Adapt? 10 Simple Steps to Deciding Whether Your Next Trial Should be Adaptive (The Cytel Blog)
