Glossary of Terms and Definitions for Innovative Clinical Trials

The National Institutes of Health (NIH) and the U.S. Food and Drug Administration (FDA) have released a draft glossary of clinical research terms related to innovative clinical trials, including studies using Real World Data (RWD) to generate Real World Evidence (RWE). The draft glossary is intended to facilitate communication within the clinical research community by helping establish a common vocabulary to more uniformly characterize clinical research. The document is open for comments until 24 June 2024.

The terms in the glossary are grouped into the following two areas: Terms for Comment and Terms for Reference (included as Appendix A).

Terms for comment include, for example, the following proposed definitions:

• Data Harmonization
  The process of combining data from different sources and reorganizing it according to a single schema so that data are compatible and comparable. Data are combined by either identifying equivalent data elements between the sources or by applying specific transformations between the elements to derive a common data element.
• Data Transformation
  The process of converting data from one format or structure into another format or structure. It is a process of data extraction and conversion or normalization in construction of analytic datasets.
• Missing Data
  Data that would have been used in the study analysis but were not observed, collected, or accessible. This refers to information that is intended to be collected but is absent and information that is not intended to be collected and is therefore absent. There may be special considerations regarding RWD sources (e.g., electronic health records (eHRs)). Such data are generally not collected for primary research purposes and therefore may not have systematic data capture to answer a research question.
• Synthetic Data
  Data that have been created artificially (e.g., through statistical modeling, computer simulation) so that new values and/or data elements are generated. In general, synthetic data are intended to represent the structure, characteristics and relationships seen in actual patient data, except that they do not contain real or specific information about individuals.

Please see the draft glossary for more information.