Collaborative Platform Trials to Fight COVID‐19: Methodological and Regulatory Considerations for a Better Societal Outcome

Abstract

For the development of coronavirus disease 2019 (COVID-19) drugs during the ongoing pandemic, speed is of essence whereas quality of evidence is of paramount importance. Although thousands of COVID-19 trials were rapidly started, many are unlikely to provide robust statistical evidence and meet regulatory standards (e.g., because of lack of randomization or insufficient power). This has led to an inefficient use of time and resources. With more coordination, the sheer number of patients in these trials might have generated convincing data for several investigational treatments. Collaborative platform trials, comparing several drugs to a shared control arm, are an attractive solution. Those trials can utilize a variety of adaptive design features in order to accelerate the finding of life-saving treatments. In this paper, we discuss several possible designs, illustrate them via simulations, and also discuss challenges, such as the heterogeneity of the target population, time-varying standard of care, and the potentially high number of false hypothesis rejections in phase II and phase III trials. We provide corresponding regulatory perspectives on approval and reimbursement, and note that the optimal design of a platform trial will differ with our societal objective and by stakeholder. Hasty approvals may delay the development of better alternatives, whereas searching relentlessly for the single most efficacious treatment may indirectly diminish the number of lives saved as time is lost. We point out the need for incentivizing developers to participate in collaborative evidence-generation initiatives when a positive return on investment is not met.

Effect of anti-interleukin drugs in patients with COVID-19 and signs of cytokine release syndrome (COV-AID): a factorial, randomised, controlled trial

“…With lower mortality due to deeper understanding of the disease and improvements in standard of care, it becomes increasingly more difficult to show a beneficial effect. In addition to event rate, differences in trial design or type of outcome measure could also affect outcome, 12 , 31 , 37 or IL-6 blockade could be most effective in populations at very high risk of death. However, the baseline characteristics of patients in RECOVERY were similar to COV-AID, including age, duration of symptoms and hospital stay, CRP concentration, and type of respiratory support, whereas those of REMAP-CAP only differed in terms of higher rate of mechanical ventilation.…”

On model-based time trend adjustments in platform trials with non-concurrent controls

“…Thus, concurrent control patients have a positive allocation probability of being randomised to this experimental arm. Unadjusted treatment-control comparisons that pool concurrent with non-concurrent controls can be biased if there are time trends in the control data, e.g., due to a change in standard of care, change in the patient population, or other external changes such as seasonal effects or a pandemic [4,5,6,7,8]. Lee and Wason [9] investigated linear regression models for continuous data that include a factor corresponding to time to adjust for potential time trends.…”

Multiplicity adjustments in parallel-group multi-arm trials sharing a control group: Clear guidance is needed

“…Some 20% of superiority trials registered in 2010–2012 had more than two groups [ 6 ]. However, there appears no consensus, across stakeholders such as regulators and scientific journals, on the necessity to control for a potentially inflated type 1 error rate when comparing distinct treatments to a shared control group in confirmatory parallel-group multi-arm trials [ 7 ] and this has become the subject of much recent debate among statisticians and trialists [ 5 , [8] , [9] , [10] , [11] ]. Many guidelines on multiplicity do not refer specifically to multi-arm trials [ [1] , [2] , [3] ], resulting in inconsistencies in the application of adjustment methods in published articles.…”

“…When multiple drugs are successful, the FDR has the advantage that it represents the expected proportion of inefficacious drugs among the successful drugs. Wason et al recommend that sponsors and trialists consider use of the FDR for multi-arm trials testing distinct treatment arms [ 22 ], while others suggest the FDR as an appropriate control measure in the context of trials with a large number of treatment arms [ 24 ]. However, trials with many treatments may be less likely to have distinct treatments.…”

“…We have focussed on the parallel-group multi-arm trial design, but various perspectives on the need for multiplicity adjustments in more complex designs such as basket, umbrella and platform trials are also being debated [ 14 , [24] , [25] , [26] , [27] , [28] ]. Such new and adaptive designs are increasingly important, especially in the case of COVID-19 where adaptive, cost-effective and rapid trials are critical [ 24 ]. However, as discussed by Collignon et al [ 25 ], these complex designs have raised challenging statistical questions around the need for control of multiple testing when adding arms or drugs over time.…”