(287) A Replication of the Incidence Rate Estimates of Selected Adverse Events of Special Interest (AESIs) from the vACCine covid-19 monitoring readinESS (ACCESS) in the EU with US Real World Data

Publication Number: 618

Background: The reproducibility of results from studies with Real World Data (RWD) is essential to evaluating the strength of the reported evidence. However, replications of results from studies with RWD are often unavailable, even for studies that could inform clinician practice or public health decision making. To address this gap, we conducted a replication of selected results from vACCine covid-19 monitoring readinESS (ACCESS) in the EU with data from Electronic health Records (EHR) in the US.

Objectives: Using data from the Optum EHR in the US from 2017 through 2019, we aimed to reproduce the incidence rates (IRs) from EU ACCESS for the following Adverse Events of Special Interest (AESIs): Guillain-Barre syndrome (GBS), Multisystem Inflammatory Syndrome (MIS), stress cardiomyopathy, and myocarditis alone and with pericarditis.

Methods: AESI definitions from 9 databases across 7 EU countries were abstracted from the online ACCESS documentation (https://zenodo.org/). IRs were estimated with patient data from the Optum EHR using the definitions from the ACCESS ‘narrow concepts.’ In the Optum EHR and ACCESS, a one-year washout period after database entry was required to establish an index date for follow up and qualify a patient for incident events. IRs from the Optum EHR were compared to those reported in ACCESS for all ages and each year from 2017 to 2020. IR point estimates from Optum EHR that fell within the range of the point estimates from ACCESS were interpreted as evidence of a successful replication.

Results: All of the AESI IRs (per 10^5 person years [PYs]) estimated with Optum EHR data were within the range of the IRs reported in EU ACCESS. Specifically, support was found for a successful replication of: GBS (Optum EHR = 1.77, ACCESS [0, 4.71]), MIS (Optum EHR = 2.48, ACCESS [0.2, 6.94]), stress cardiomyopathy (Optum EHR = 6.24, ACCESS [0.05, 7.97]), myocarditis alone (Optum EHR = 5.97, ACCESS [0.67, 23.68]), and myocarditis with pericarditis (Optum =22.52, ACCESS [1.54, 34.66]).

Conclusions: The IRs for GBS, MIS, stress cardiomyopathy, and myocarditis alone and with pericarditis estimated with Optum EHR data in the US were within the range of IRs reported by ACCESS in the EU. These findings suggest that the results from ACCESS can be independently replicated with RWD outside the EU, which may provide confidence to decision makers who are using these data to contextualize AESIs that are observed in post marketing RWD.