FDA provided additional details for sponsors on how they should go about designing adaptive trials for medical devices.
The guidance, released July 27, makes several updates to draft guidance that was released on May 18, 2015.
Some of the changes in the updated guidance include additional examples of modifications for prospectively-planned adaptive study designs. The agency stresses that adaptive trials focus on anticipated changes based on accumulating data.
The most important consideration is whether an adaptive design is feasible and whether an adaptive design is more appropriate than conventional designs, the FDA said. The agency provides suggestions on how to decide if an adaptive design is advantageous, and it provides more detail on various scenarios to make that determination.
The document lays out the principles for adaptive trials and how sponsors need to control the chance of erroneous conclusions and to minimize operational bias. Both of these factors can be a “significant threat” to the scientific integrity of a clinical study and can’t be overcome by statistical adjustments, the agency warns.
Different adaptive trial designs are explained in detail, including group sequential designs, sample size readjustment, Bayesian sample size adaptation, dropping a treatment arm, changing the randomization ratio, and investigating both superiority and non-inferiority adaptive enrichment.
The agency received comments from six organizations, including AdvaMed, which praised the agency for the document, and provided an eight-page attachment that highlighted line item changes it wanted, including examples of adaptive designs for diagnostic devices. In response, a new section provides more information on adaptive studies for diagnostic devices.
The agency also received comments from Cook Medical Group, which noted that the agency should clarify when trials should be stopped early for futility. It also suggested the agency provide clarity on when adaptive designs may not be feasible.
The guidance also goes into detail on special considerations such as changes to pivotal trials that are not preplanned using blinded data and unblinded data. Challenges of adaptive designs are highlighted as well, and the agency stressed the importance of data monitoring committees and IRBs, techniques to minimize operational bias and logistical challenges.