What is an adaptive design?

Adaptive trial designs aim to answer research questions efficiently while balancing ethical and scientific interests. They offer controlled flexibility to change certain aspects of an ongoing trial based on emerging trial data, such as removing ineffective therapies, stopping the trial early, changing the sample size, or targeting a specific group of patients who are likely to benefit most.

Read more about adaptive designs

General information

Adaptive designs allow researchers to make changes to certain parts of a clinical trial based on what they are learning from patient data. However, researchers need to ensure that: the design is relevant to the research question and feasible; the trial is conducted and monitored correctly; data are analysed appropriately; the results are reported fully and transparently.

Read more about general practical and statistical considerations when using adaptive designs

Types of Adaptive Designs

Sample Size Re-estimation (SSR)

Changes to the initial sample size required to address research questions are made based on accrued information in response to incorrect assumptions or guesses made at the planning stage about study design parameters. Sample size can be increased or reduced accordingly.

Read more
Group Sequential Design (GSD)

Offers opportunities to stop the trial early as soon as there is sufficient evidence to make reliable conclusions about the effects of a treatment by analysing accumulating data at interim analyses. Researches can stop a trial early because they found enough evidence of benefit or lack of benefit. The criteria for early stopping are pre-specified.

Read more
Multi-Arm Multi-Stage (MAMS)

Several competing treatments are evaluated simultaneously. Based on accrued outcome data, treatment arms showing promising benefit are chosen for further testing, those unlikely to show benefit are dropped, or trial stopped early for benefit. The criteria for selecting/dropping arms or early stopping at interim analyses are pre-specified.

Read more
Adaptive Population Enrichment (APE)

Researchers can evaluate whether treatments benefit heterogeneous target patients or only patients with specific characteristics (subpopulations) in a single trial . Using predefined selection rules, subpopulations showing promising benefits are selected at interim analyses and eligibility modified to focus recruitment to selected subpopulations.

Read more
Response Adaptive Randomisation (RAR)

Offers researchers opportunities to change how patients are allocated to treatments based on some pre-specified rules capturing the emerging signals of treatment benefits. More patients can be allocated to treatments that are indicating more benefits away from those that are less efficacious or unsafe.

Read more

Our Collaborators

Find out more about our collaborators