Acknowledging that collecting data from a diverse set of patients is not always feasible in preapproval trials, an FDA draft guidance provides postmarket strategies for gathering data from underrepresented populations, including study design and statistical considerations for single-arm trials, randomized trials, real-world data (RWD) sources and pooled studies.

Although the FDA strongly encourages sponsors to gather data from diverse and representative patient populations early in drug development and prior to initial approval, it realizes that in certain scenarios, those data may be limited and “must be balanced within the benefit-risk framework, including whether there is unmet medical need and the importance of the product.”

Randomized trials planned at the time of NDA or BLA submission “may be revised to enrich the trial for the subpopulation(s) of interest to obtain postmarketing data,” the FDA says. In some cases, randomized trials that are ongoing at the time of submission can be modified as well, but those proposed revisions should be discussed with the agency, as changes could affect statistical analyses.

Sponsors may also stratify based on the subpopulation(s) of interest in randomized trials if those groups possess potential prognostic implications, though sponsors should collect adequate pharmacokinetic and pharmacodynamic data, as applicable, to understand differences in those groups.

Sponsors may also group by underrepresented populations of interest in randomized trials if those groups have the potential to show relevant outcomes.

For instance, “a trial can stratify based on race, ethnicity, sex, age, or a hypothesized difference in efficacy in the population of interest versus the general population, so that analyses can focus on benefits and risks in the underrepresented population,” the guidance advises.

Read the draft guidance here.