FDA Suggests Statistical Strategies for Mitigating Data Loss During the Pandemic
Sponsors of trials impacted by the COVID-19 pandemic have several statistical analysis options for mitigating the effect of missing data, according to a final guidance from the FDA that outlines strategies the agency considers acceptable.
The guidance notes that multiple strategies may be needed to adequately address the COVID-19 impact. Strategies include collecting information on discontinuation of treatment, withdrawal of a patient from the trial, use of alternative or rescue treatments, and missed or alternative endpoint assessments to incorporate into the statistical analysis plan.
Sponsors that consider stopping a trial early and conducting final analysis may perform a blinded power assessment to determine what effect the smaller data pool might have on the validity of statistical conclusions. The guidance cautions that any trial that is stopped earlier than planned must avoid changing the planned analysis in a way that would introduce bias.
It also may be reasonable to exclude patients from the data pool if they were not able to receive the full benefit of the treatment being studied.
Trials also may plan to gather additional data by enrolling additional participants and extending the trial period beyond the end of the pandemic or by extending the patient follow-up period after treatment ends.
Modifications of trial endpoints and statistical analysis plans should be discussed with the FDA, the guidance says.
Read the guidance here: https://bit.ly/3dor6uB.