Table 1 Key features and approaches to adaptations and ethical considerations of frequent interim analysis and response-adaptive randomization for adaptive trial designs

Frequent interim analysis (FIA)

• Trial success and futility are repeatedly assessed.

• Terminate when successful

• Lower total trial size when treatment is more effective than anticipated or has little or no effect

• Ethical advantage accrues as fewer participants may be required to terminate a successful trial or negative trial.

• Provides an ethical advantage for the population of patients involved by exposing fewer participants to ineffective treatments.

• Terminate if continuation is unlikely to succeed

• Assess:

• If more participants are needed when experimental treatment appears to have marginal effectiveness, the probability of benefit will likely exceed 50:50; this potentially poses an ethical advantage for the population of participants because the chance of getting an effective treatment, even if marginal, may be greater than in a trial without FIA

• May increase sample size per prespecified rules

• May result in a larger required sample size if experimental therapy has marginal effectiveness

− evidence of treatment efficacy

• Offers ethical advantage by avoiding erroneous conclusion that a marginally effective treatment is ineffective, and obfuscating use for individuals who could benefit from treatment

− rates of adverse events

− patient safety

Response-adaptive randomization (RAR)

• Randomization proportions are varied during the course of the trial.

• Reallocation between arms to increase the fraction of participants who receive the most effective treatment

• Can decrease the required sample size by allocating patients in arms most likely to be efficacious

• Decreasing the sample size offers ethical advantage to the population of eligible participants by exposing fewer of them to ineffective treatments.

• Ethical benefits are more likely to accrue for the population of patients in the trial as more than half are likely to be enrolled in arms with probability of benefit

• Ongoing modification

• Ethical benefit accrues over the course of the trial for individual participants, who will be more likely to receive effective treatment.

• Allocate patients to different treatment arms

• Efficiency enhances ethical benefits for society by minimizing unnecessary costs and allowing precious resources to be re-allocated to other research for other effective treatments.

• As in randomized, controlled trials without RAR, participants enrolled early have no advantage for being in an arm with the more effective treatment, but enrollment later enhances probability of enrollment in the effective treatment arm

Offers ethical advantage to the population of persons in society who need the treatment under evaluation, since the trial results of an effective treatment will be known sooner, providing clinicians and patients with treatments known to offer benefit

• Allocate patients to different doses of an active agent

• May improve the statistical efficiency of the trial if there are more than two arms