– Written by the Quality and Reusability of Real World Data project within the Real World Evidence Working Group.

The FDA guidance on assessing registry data to support regulatory decision-making has been considered by the Real World Evidence Working Group. In this blog post, we discuss some of the topics from the guidance and identify some of the challenges in applying them to our current practice. The opinions expressed in this blog post are those of the authors and should not be construed to represent the opinions of PHUSE members, respective companies/organisations or regulators’ views or policies. The content in this document should not be interpreted as a data standard and/or information required by regulatory authorities.

Registries – odd ones out among RWD sources?

A registry is an “organized system that collects clinical and other data in a standardized format for a population defined by a particular disease, condition or drug exposure”. Of the RWD sources out there, registries are unique in that they can be set up specifically to answer our questions of interest, in a similar way to how we set up clinical trials, versus the repurposing of an existing registry dataset. In either case, registries can be helpful in contexts where clinical trials may not be feasible or ethical, such as rare diseases.

Their application goes beyond that for which RWD is perhaps most well known – as an external control arm. Registries can also provide sponsors with further insights into the natural history of a disease; we can use them to inform study planning, such as sample size calculations, selection criteria and suitable endpoints. And registry data can also supplement what we collect in an interventional study.

Getting into technicalities, data can be entered directly into the registry or be linked to other data sources that characterise the registry participants, and this data may be objective or subjective (such as patient-reported) measurements.

What are the main messages in the regulatory guidance?

Since the initial draft, the guidance has made it explicit that sponsors are responsible for ensuring attributes of the registry support the collection of relevant and reliable data.

The relevance of the registry data speaks to the questions of interest we would like to address from analyses of this data. In other words, does the registry collect the key variables needed, and are there sufficient measures in place to avoid missing assessments of these variables?

When considering the reliability of the registry data, the guidance points to its accuracy, completeness and traceability, and having appropriate governance attributes and processes in place for ensuring these.

What challenges do we foresee in the application of the guidance?

In terms of process, the guidance reflects the shifting landscape of registry data usage with the 21st Century Cures Act and growing interest in leveraging RWE for decision-making across the entire product life cycle. On first sight, the proposed procedures and documentation mimic what we typically expect in a clinical trial setting, and the guidance serves as a reminder that as we, as an industry, move from collecting data ourselves in house to making use of readily available sources, strict standards still apply.

In this respect, the guidance presents various challenges:

Consenting and/or reconsenting

Obtaining the informed consent of registry participants for sponsors’ proposed use of their data may not be straightforward. In an ideal scenario, the initial informed consent provided by the participants will allow for the subsequent use of the registry data. Obtaining reconsent from registry participants at a later stage will be challenging and could introduce additional biases, if reconsent is obtained from only a selection.

Registries with linkages to other data sources

The guidance highlights additional considerations for registries with linkages to other data sources – first, with respect to informed consent, the informed consent provided by registry participants should allow for the linkages. Moreover, the transfer of the data from the external data source to the registry should be validated. This means ensuring sufficiently accurate patient-level linkages, considering any missing data, and applying predefined rules for consistency. It is also worth noting that these linkages may or may not be applicable to all registry participants. Some linkages may be jurisdiction-dependent or dependent on other characteristics of the participants, such as stage of disease. Clear documentation of the entire linkage process will be critical in understanding what inferences can be made from analyses of these data.

Using registry data in regulatory applications

When analyses of registry data are used as evidence to support a drug’s effectiveness or safety, sponsors should be prepared to make the patient-level data readily available for regulatory review – both the metadata and the source records. In the first instance, patient-level data – tabulation (i.e. SDTM) and analysis (ADaM) datasets – are required to be submitted to the NDA to allow regulatory reviewers to replicate the analysis. Second, the FDA may need access to the source data (source records), including the metadata for inspection purposes. As a Working Group, we saw the challenges in this respect at our Community Forum in January 2024 (more details here). Within the sponsor team, this requirement should be made clear upfront. When working with third-party data providers, the sponsor should, in the development phase of the study, explore whether the above is feasible. And across the RWE FDA guidance, the sponsor is invited to engage with regulators as soon as possible – on this topic and on the RWE study plans more generally.