Adaptive and Platform Clinical Trials Explained

Traditional clinical trials are usually designed with a rigid, fixed structure. A specific number of participants are enrolled, given either the experimental medicine or a comparator, and the results are only analysed once the trial concludes. If the drug fails, a new trial must be designed from scratch. In recent years, researchers have increasingly turned to adaptive and platform trial designs. These modern approaches allow for pre-planned modifications while a study is ongoing, based on accumulating data. Rather than waiting years to conclude a study before pivoting, these designs aim to make the research process more efficient, ethical, and responsive to emerging scientific evidence.

Adaptive trial designs

An adaptive clinical trial design is one where the protocol is amended based on interim data. Crucially, these changes are not arbitrary decisions made by researchers on a whim; they are pre-specified in the trial protocol before the study begins. Common adaptations include adjusting the sample size, changing the dosage, or modifying the criteria for enrolling participants.

One of the most frequent uses of an adaptive design is to drop an experimental treatment arm if early data shows it is ineffective or causing more side effects than the comparator. Another common approach is a seamless design, which combines what would traditionally be separate phases—such as Phase II (testing efficacy and safety) and Phase III (comparing to standard treatments)—into one continuous trial. By stopping an ineffective treatment early, researchers can avoid exposing more participants to a drug that offers no benefit, freeing up resources to focus on more promising options.

Platform, umbrella, and basket trials

Platform trials, sometimes called master protocols, take adaptation a step further by creating a long-running infrastructure for a particular disease. Instead of a trial having a distinct start and end, a platform trial operates indefinitely. New experimental treatments can be added as they become available, and treatments can be dropped once their effectiveness is proven or disproven. This design shares a single control group among multiple experimental arms, which significantly reduces the number of participants needed overall and speeds up the time it takes to reach a scientific conclusion.

Related to platform trials are basket and umbrella trials, which are particularly common in oncology. A basket trial tests a single targeted treatment across multiple different cancers that share a specific genetic mutation, regardless of where the cancer originated in the body. An umbrella trial, conversely, looks at a single broad disease—such as lung cancer—and tests multiple targeted treatments against different genetic mutations found within that single disease type. Both approaches rely heavily on biomarker testing to match the right patient to the right experimental arm.

Why modern trial designs became prominent

These designs gained significant traction over the last decade because they address some of the major inefficiencies of traditional clinical research. Fixed trials often require large numbers of participants and many years to complete, and if a drug fails, the process must start over entirely. Platform and adaptive designs allow researchers to answer multiple questions simultaneously and reuse the data infrastructure.

They became highly visible during the COVID-19 pandemic, where platform trials like the UK's RECOVERY trial rapidly identified effective treatments by adapting on the fly and dropping those that showed no benefit. They are also particularly useful in rare diseases, where finding enough patients for multiple traditional trials is incredibly difficult. Regulators, including the South African Health Products Regulatory Authority (SAHPRA), evaluate the robustness of the trial design when considering these medicines for registration. Consumers can later use tools like HealthSA to search for these registered medicines by active ingredient or trade name.

Statistical and ethical safeguards

A common concern with mid-study changes is the introduction of bias or an inflated chance of false-positive results. If researchers look at data too often and change the rules every time, they might inadvertently skew the statistical outcome to make a useless drug look effective. To prevent this, adaptive and platform trials employ strict statistical controls, such as alpha spending functions, which carefully manage the overall error rate of the study regardless of how many times the data is reviewed.

Additionally, these trials rely heavily on Independent Data Monitoring Committees (IDMCs). These are external panels of experts who review the unblinded data—data showing which participants received which treatment—at pre-specified intervals. Only the IDMC sees the accumulating results; the researchers running the trial remain blinded to prevent bias. If an adaptation is needed, the IDMC recommends it based on strict criteria outlined in the protocol. Regulators globally provide detailed guidance on how these trials must be structured to ensure that flexibility does not compromise scientific integrity.