The FDA has published further draft guidance on real-world evidence (RWE), this time outlining regulatory considerations for observational studies that use real-world data (RWD).

The agency reminds sponsors that they must still meet the regulatory requirements for IRBs and the protection of human subjects. The guidance recommends that sponsors consult with experts on data privacy, when it’s appropriate, to inform the protocol during the study design process, as they may aid in identifying and handling data privacy and security concerns that arise when accessing healthcare data.

Sponsors should connect with the agency early on when designing a noninterventional study to support a marketing application, such as by requesting a Type C meeting to learn the agency’s expectations for the design and conduct of their trial, the guidance explains. Draft versions of the proposed protocol and statistical analysis plan (SAP) should be given to the FDA for review and comment prior to being finalized and conducting study analyses.

In addition, sponsors should communicate with the agency early in the design stage to learn its expectations for RWD access. Sponsors need to ensure that they will be capable of submitting patient-level data for any RWD analyzed as part of the study when required to do so. If certain RWD are owned and controlled by third parties, agreements should be in place with them to ensure this patient-level data can be given to the FDA and that source data for verifying RWD will be available.

All RWD and associated programming codes/algorithms should be documented, well-annotated and complete when submitted to the FDA, the guidance says.

Because the agency needs to be assured that the sponsor didn’t pick data sources and databases and conduct analyses to favor a certain conclusion, the protocol and SAP should be finalized before the prespecified analyses listed in them are carried out. It’s expected that the sponsor will provide evidence that the protocol and SAP were finalized before the outcome data of a study are reviewed and prior to doing the prespecified analyses. Further, any protocol revisions should be dated and explained, according to the guidance.

In the study protocol, describe all the data sources accessed when designing the study, in addition to results from feasibility evaluations or exploratory analyses of those sources. The agency expects sponsors to list its reasoning for including or excluding data sources, and it recommends generating audit trails in the study’s datasets to track access to and analyses conducted on data sources.

In addition, sponsors are expected to document every analysis performed on the data during the study design phase, including the feasibility evaluations and exploratory analyses. They’re also expected to show that their choice of the final analytic dataset and the performance of the final analyses align with the research question rather than favoring particular study findings.

The patient characteristics of both the source population (the population from which the study population is drawn) and the study population (the population undergoing analyses) should be described, and sponsors should outline any differences between them that could have an effect on study results.

Read the full guidance here: https://bit.ly/3EBDe9M.