On June 17, the Food and Drug Administration (FDA) continued its efforts to mitigate COVID-19’s disrupting impact on clinical trials by issuing guidance on statistical considerations for changes to trial conduct (FDA previously relaxed restrictions on protocol modifications). As expected, public health measures designed to control COVID-19’s rapid emergence as a global pandemic—social distancing, travel restrictions, site closures, etc.—have impeded drug and device industry innovation. Many clinical trials have been delayed; others cannot get up and running. Meanwhile, participant enrollment continues to present a challenge due to fear of infection, and logistics firms are working tirelessly to smooth unprecedented supply chain interruptions. The expected downstream effect is delayed access to important medical products. FDA’s latest guidance therefore aims to protect the continuity of clinical trials and equip sponsors to adapt to these challenging circumstances without compromising trial integrity and safety. Likewise, the guidance is not limited to trials for the development of COVID-19 treatments; the recommendations apply equally to other diseases

FDA’s statistical considerations for modifying clinical trials to address the impact of COVID-19 begin with three guiding principles:

  • Proactive planning: Sponsors should update statistical analysis plans prior to modifying primary or secondary endpoints.
  • Bias avoidance: Sponsors should not propose trial modifications that have the potential to bias the interpretation of findings and should implement safeguards such as a data monitoring committee.
  • Appropriate data selection: While making trial modifications, sponsors should focus on summaries of participant-specific data, such as null values, treatment discontinuation, treatment interruptions, trial withdrawal, and endpoints.

FDA’s guidance concludes by applying the above principles to scenarios that sponsors commonly encounter in the COVID-19 context. In each case, FDA offers detailed statistical considerations for trial design and analysis strategies. For example, FDA stresses the importance of performing sensitivity analyses to evaluate the impact of any change in endpoint definition or ascertainment.