6/7 June 2023
Following the draft guideline published in September 2020, the European Medicines Agency (EMA) published the final guideline on the use of real-world data (RWD) for benefit-risk assessment. According to the agency, the guideline addresses "the methodological, regulatory and operational aspects involved in using registry-based studies to support regulatory decision-making". In addition, the U.S. Food & Drug Administration (FDA) recently issued several draft guidelines on the use of RWD for drug and biological product submissions.
The objective of the EMA guideline is to provide recommendations on key methodological aspects that are specific to the use of patient registries by marketing authorization applicants and holders planning to conduct registry-based studies. The guideline contains information on data collection, data quality management, data elements, analysis, reporting and governance.
According to the agency, registry-based studies are investigations using the data collection infrastructure or patient population of one or more patient registries. A patient registry is an organized system "that collects uniform data (clinical and other) to identify specified outcomes for a population defined by a particular disease, condition or exposure". Patient registries may have several purposes, such as to monitor the clinical status, quality of life, comorbidities and treatments of patients over time. They are an important data source for registry-based studies on healthcare practices, utilization of medicines and medical devices, and outcomes of treatments. In particular, they may represent a data source on rare diseases and patients treated with advanced therapy medicinal products (ATMPs).
The guideline includes an Annex on considerations on patient registries plus the following three Appendices:
For more information please see EMA´s Guideline on registry-based studies.
The FDA recently issued several draft guidelines on benefit-risk assessment and the use of RWD:
For studies that require combining data from multiple data sources, FDA recommends demonstrating whether and how data from different sources can be integrated with acceptable quality.
According to the FDA, artificial intelligence (AI) may permit more rapid processing of "unstructured" parts of electronic health records (EHRs). Advances include natural language processing, machine learning, and particularly deep learning to extract data elements from unstructured parts in addition to structured ones in EHRs; develop computer algorithms that identify outcomes; or evaluate images or laboratory results. However, FDA does not endorse any specific AI technology. In addition, all of these computer-assisted methods require currently a significant amount of human-aided curation and decision-making. Furthermore, the FDA recognizes the challenges involved in standardizing study data derived from RWD sources. Thus, the agency plans to further update the guidance on data standards (i.e. a set of rules about how a particular type of data should be structured, defined, formatted, or exchanged between computer systems) that are derived from RWD sources.