Linda Malek and Jason Johnson are partners in Crowell's Health Care and Privacy & Cybersecurity Groups, and have a particular focus on advising clients on compliance issues related to clinical research and clinical trials. Stephen Holland is Senior Counsel in Crowell's Government Affairs Group and previously served as Senior Health Counsel to the U.S. House Committee on Energy and Commerce, where he advised Members of Congress and their staffs on FDA policy.
On July 25, the Food and Drug Administration (FDA) issued final guidance entitled Real-World Data: Assessing Electronic Health Records and Medical Claims Data to Support Regulatory Decision-Making for Drug and Biological Products. The guidance aims to provide drug sponsors with considerations should they wish to use real-world data (RWD) drawn from electronic health records (EHRs) or medical claims data in their clinical studies to support regulatory decisions related to the safety and efficacy of a drug.
The use of real-world evidence (RWE) in drug research has significant potential advantages, including increasing access to a significant amount of data reflecting more diverse populations, providing robust longitudinal data that may better reflect long-term outcomes of clinical interventions, and avoiding many challenges that come with recruiting and retaining clinical trial participants. However, as this guidance describes, RWD sources come with their own limitations.
The guidance includes a broad array of topics sponsors should consider when developing their study protocols. In particular, the FDA is focused on the reliability (accuracy, completeness, and traceability of data) and relevance (availability of data for key study variables such as exposures, outcomes, covariates, and sufficient numbers of representative patients for the study) of the data being used. The topics provide guidance for meeting these requirements in various areas.
The topics addressed include the following:
- Data Sources. The guidance discusses how sponsors should consider the appropriateness, relevance, and comprehensiveness of certain data sources; challenges of linking patient data and synthesizing across multiple data sources; recommendations related to assessing a common data model (CDM) across distributed data networks; and the use of large amounts of unstructured data in EHRs (i.e. doctor's notes in free text, which may not be included in specific data fields), including through using artificial intelligence (AI) to read, assess, or categorize unstructured data, though FDA notes that it does not specifically endorse any specific AI technology for this use.
Across all the discussion of data sources, FDA emphasizes transparency in study protocols and recommends that sponsors acknowledge inherent limitations in some data sources, for example, missing or incomplete data.
- Study Design Elements. The guidance also advises sponsors to clearly define relevant time periods included in the data and how to ascertain the treatment or regimen used through a data source, the duration and dose of a treatment, and the definition, ascertainment, and validation of outcomes measures.
FDA details some of the challenges that may come in assessing data through EHRs and medical claims, such as incomplete information about drug dosage adjustments in medical claims data, and missing or incomplete records related to dosing in EHRs. FDA advises sponsors to validate RWD collected through these methods, and suggests that one method of validation of a data source's capture of a treatment regimen may include the use of other studies studying the same population to estimate misclassifications in different data sources (e.g., surveys of study participants to assess the intake of a drug can help validate the completeness of a data source).
Another example of a limitation in EHR data is mortality as an outcomes measure, as death and the cause of death are often not included in EHRs in the United States, and linking mortality data creates challenges and limitations of its own. FDA advises sponsors to carefully document mortality data quality in a study protocol.
- Data Quality. The guidance describes examining the quality of data as "an ongoing process" with multiple data quality checks, cleansing, and monitoring of data quality at each stage of the process. FDA recommends that sponsors address the completeness and accuracy of data, including when the format and traceability of EHRs and medical claims data vary across health systems, and include quality assurance and quality control plans in the study protocol. Final study reports should also include an assessment of the data retrieval and transformation processes the sponsor used when conducting the study.
FDA's guidance repeatedly stresses the importance of documenting and justifying how EHRs and medical claims data are used in study protocols, and recommends addressing upfront the limitations of these RWD sources. FDA also recommends that sponsors who intend to include this data in a regulatory submission first submit their protocols and statistical analysis plan to the agency before conducting the study, so FDA can provide feedback.
Crowell Health Solutions is prepared to assist with questions about planning clinical studies, including those utilizing RWE. Please reach out with any questions.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.