Explain it to me.

Also known as Umbrella trials, or Basket trials (but Platform trial sounds cooler), they focus on the disease of interest, and are based on a master protocol which allows for varying groups for analysis against a control group.

Technically, they are a type of Adaptive, Multi-arm, Multi-stage trial (but even better)

• Adaptive trial: a trial which undergoes regular interim analysis to adjust the sample size and allocation weighting of the treatment group(s)
• Multi-arm, multi-stage trial: a trial with multiple treatment arms compared against the same control group, with the option of dropping underperforming arm(s)
• Platform trial: per the above 2, but also has the option of adding arms, and adjusting the size of the control group

Advantages

• DISEASE FOCUSSED rather than intervention focussed
• EFFICIENT use of study resources (cf running a separate trial for each intervention)
  • avoids redundant duplication of design & infrastructure
• EASIER TO COMPARE INTERVENTIONS, because all interventions are in the same trial, therefore have the same trial methodologies.
• RESISTANT TO OBSOLESCENCE: novel developments can be incorporated into the study protocol
• PERPETUAL: the trial never actually needs to end. new arms can be added and old arms can be dropped, all under the same master protocol
• MINIMISES CONTROL GROUP: ethically advantageous as the trial treatment is provided to the maximum number of potential beneficiaries
• CAN DETERMINE SUPERIORITY: a platform trial CAN declare one treatment superior to others. Don’t need to bother with non-inferiority
• PREVENTS UNDERPOWERED RESULTS by regular interim analyses and sample size adjustments
• DECREASED RISK OF PARTICIPATION: any patient enrolled has a reduced risk of being in the control group (smaller cf treatment groups) and reduced risk of getting ineffective treatments as these groups will be dropped from the study

Disadvantages

• DATA COMPLEXITY: harder to manage:
  • adaptive randomisation
  • multiple stages of interim analyses
  • multiple arms
  • design of master protocol with longevity
• ADMINISTRATIVE COMPLEXITY: each intervention may have its own:
  • corporate sponsor
  • additional consent requirements
  • follow up
• HARDER TO FUND: no fixed end date
• TAKES LONGER TO GET STARTED: unsuitable for interventions which need to be tested immediately
• NON-CONCURRENT CONTROL GROUP risks bias due to changing practice over time

Examples

REMAP-CAP

RECOVERY

ASCOT

I-SPY COVID