Real World Evidence

Real World Evidence

Derived from data sources such as electronic health records, claims data, registries, and mobile devices, real world evidence (RWE) has the potential to bring innovative products to patients more quickly. Unlike randomized controlled trials (RCTs), which may not be representative of a general population due to strict inclusion and exclusion criteria, RWE may more closely identify how an investigational product will perform in a general population. To address the absence of widely accepted best practices for utilizing RWE, the MRCT Center collaborated with Duke-Margolis Center for Health Policy to define a framework to establish best practices for utilizing RWE for regulatory decision-making.

The MRCT Center also worked with OptumLabs to investigate sources of variability in RWE analysis as applied to replication (emulation) of RCTs. Conceptualized by the MRCT Center and OptumLabs, the Observational Patient Evidence for Regulatory Approval and uNderstanding Disease (OPERAND) Project aimed to understand the sources of variation in design, approach, methodologies, statistical analyses, and decision-making using real world evidence (RWE, specifically claims and EHR data) to emulate phase 3 clinical trials. First, RWE is used to replicate published clinical trials, using and analyzing data limited explicitly to the eligibility criteria of the trial. Later, the aperture of inclusion is broadened to determine any change in effect size by the broader population.

The OPERAND project convened a technical expert panel (TEP) comprising key stakeholders from industry, academia, and regulators, to consider the principles behind, the methodology to draw upon, and the appropriate utilization of, observational data in regulatory review and approval and comparative effectiveness research.

Brown University and Harvard Pilgrim Health Institute, were selected to replicate two trials:

• (1) the ROCKET trial for atrial fibrillation, a multicenter, randomized, double-blind, double-dummy, event-driven trial that was conducted at 1178 participating sites in 45 countries) and
• (2) the Lead-2 trial for Type 2 Diabetes control, a multi-center, randomized, double-blind, parallel assignment of 1091 participants conducted at 190 study sites across 4 continents).

The project helped validate the use of observational data to complement evidence from RCTs and used empirical data to understand methodologies and sources of variability.

Project Resources