(180) External Comparator Cohort studies – clarification of terminology

Background: Though there is only one term for the gold standard of Randomized Controlled Trials, the terminology used for controlled research involving external data is diverse. Common terms include External Comparator/Control Arm study, Externally Controlled Trial, Synthetic Control study and Historical Control study.

Objectives: To clarify nomenclature when using external data by examining whether applied terminologies are exchangeable or meaningfully different, and whether some nomenclature should be considered preferable over others.

Methods: Paired terms like study/trial, control/comparator and arm/cohort as well as composite terms like External Comparator/Control Arm study, Externally Controlled Trial, Synthetic Control study and Historical Control study are appraised for meaning and implications.

Results: The term Externally Controlled Trial was recently selected by the U.S. Food and Drug Administration (FDA) and is in line with the use case of a pivotal trial. It entails pre-specification of the external dataset and its analysis in the trial protocol, which produces the highest amount of transparency. Hence, the term Externally Controlled Trial should be used if the condition of pre-specification of the comparison against the selected external dataset applies. If not, we derive the term External Comparator (Cohort) study, because due to the observational study setting there is no fully controlled research overall (hence: comparator instead of control), no pure trial setting (hence: study instead of trial) and no treatment arms as in randomized trials, but only treatment cohorts. The term Synthetic Control study should not be applied, because it is already used for another kind of design which is re-weighting units from aggregated data. The term Historical Control study is valid terminology due to the ICH E10 guideline about the choice of control groups in clinical trials but could be perceived to denote a negative connotation.

Conclusions: Those who are most familiar with conducting RCTs may use the terms like arm and trial out of habit which may be misperceived to imply an inadequate proximity to the RCT setting when using external data. Re-using RCT nomenclature in the case of not fully controlled observational settings may affect (potentially subconsciously) unrealistic expectations of the design approach. Hence, we have proposed an overall framework of nomenclature, which is generally applicable for research projects involving external data. Precise and consistent use of terminology will most effectively safeguard from unintended implications and inaccurate perceptions.