EM3 Robustness of External Control Arm: WHEN to Use Them


      During the COVID-19 pandemic, many aspects of traditional clinical trials have been affected worldwide. Recruitment of patients and on-site visits have been challenging when not compromised. Many groups have turned to the possibility of replacing their randomized standard of care (SOC) arm with a real-world (RW) external control arm (ECA). Unlike randomized controlled trials (RCTs) that have international guidelines, the use of an ECA is not subject to any consensus.


      The aim of this study is to guide the design of an ECA (when it is justified or recommended) from different context and data sources.


      We propose to summarize the evidence into a decisional matrix. We crossed popular data sources (RW data collected prospectively, RW data obtained from retrospective chart review, administrative or insurance data, and clinical trial data) with situations where the use of an ECA could be justified or beneficial (regulatory submission, health economics and outcomes research [HEOR] investigation, hypothesis generation). Our reflection was influenced by our consulting practice at Evidera and by United States Food and Drug Administration (FDA) guidelines on the use of RW data. We developed a framework that should help researchers to build a quality ECA. Building an ECA should be based on a clear research question that will inform: (1) design and data source(s) to be used; (2) selection of control that will limit biases; and (3) adjustment methods that allow fair comparison with the treated arm.


      Good ECAs would have a clear and accepted SOC treatment (limited changes in medical practice), standardized variables and definitions, similar outcome evaluations, and validity of the variables.


      When it comes to ECA, there is no one size fits all solution. The ECA should not replace RCTs but be considered as a complement tool to provide additional evidence to medical research.