FDA Guidance Allows Alternative Data Sources for External Control Arms
The FDA says trial sponsors and investigators may use patient-level data from other trials and/or real-world data (RWD) sources as an external control arm in a new draft guidance.
The guidance offers direction on design, analysis and bias considerations for trials that employ external control arms. Data sources may include medical claims, registry and electronic health record (EHR) data. It does not go into designing/analyzing natural history studies, the reliability and relevance of various RWD sources or the use of external control arms in traditional randomized trials.
Because these trials don’t randomize their participants, it’s important that the treatment and control arm groups are as similar as possible, the guidance notes, and a number of factors, including demographic characteristics/comorbidities, disease attributes, clinical observations, concomitant therapies, and start of follow-up for the investigational treatment, should be considered.
Before initiating a trial, “consider the likelihood that such a trial design would be able to distinguish the effect of a drug from other factors that impact the outcome of interest and meet regulatory requirements,” the FDA says.
Addressing potential biases in these trials is critical and best done in the design phase, the guidance says, and protocols should be finalized prior to their conduct, including the selection of the external arm and analytic methodology, rather than choosing the external arm after a single-arm trial.
The FDA recommends prespecifying certain components before conducting an externally controlled trial, including:
- Suitable trial data sources;
- Inclusion and exclusion criteria;
- Appropriate exposure definitions and windows;
- Well-defined and clinically meaningful endpoints;
- Cogent analysis plans; and
- Approaches to minimizing missing data and sources of bias.
Sponsors should prespecify their plans for measuring and analyzing data on compounding factors and sources of bias, the FDA advises, and recognize that there are limitations to identifying these confounding factors.
The guidance also cautions that suitable data on suspected confounding factors may be missing for some participants or measured differently between the external control arm and treatment arm. Before deciding whether an externally controlled trial is suitable, sponsors should verify that important prognostic characteristics can be evaluated in the data sources that will be used.
Read the guidance here: https://bit.ly/3WTlkb0.
Upcoming Events
-
21Oct