Epidemiologist/Health Statistician US Food and Drug Administration Silver Spring, United States
Background: The U.S. Food and Drug Administration Biologics Effectiveness and Safety (BEST) Initiative conducts active surveillance of adverse events (AEs) after COVID-19 vaccination. During near real-time safety monitoring, we identified a signal indicating possible increased risk among the adult population for myocarditis/pericarditis (myo/peri). Medical chart review was conducted to validate the claims-based myo/peri case identification algorithm.
Objectives: To validate the performance of the claims-based algorithm used to identify myo/peri cases following COVID-19 vaccination among adults aged ≥18 years and describe characteristics of myo/peri cases.
Methods: Myo/peri cases were identified using claims from all healthcare settings through the observation of at least one relevant ICD-10 diagnosis code within 42 days post-COVID-19 vaccination among adults ≥65 years in Medicare and 18-25 years (HealthCore and CVS Health) or 18-64 years (Optum) in commercial claims databases. Cardiologists reviewed medical records to assess each case against the Brighton Collaboration’s case definition for myo/peri, and positive predictive values (PPVs) and 95% confidence intervals (CIs) were calculated. We also summarized descriptive characteristics such as time from COVID-19 vaccination to myo/peri diagnosis (i.e., time-to-event), hospitalization status, and length of hospitalization.
Results: We identified 47 cases within Medicare and 75 cases across all commercial databases (all ages; 18 in HCI, 3 in CVS Health, 54 in Optum) using the claims-based algorithm. Among adults ≥65 years in Medicare, 8 cases were classified as confirmed/probable, resulting in a PPV of 17.0% (95% CI: 8.9, 30.1). Across all commercial databases for the 18-25 years age group, 26 cases were confirmed or probable, resulting in a PPV of 74.3% (95% CI: 57.9, 85.8), with variation observed within specific databases. The PPV for Optum’s 26-64 age group was 35.0% (95% CI: 22.1, 50.5). For confirmed/probable cases, the median time-to-event was 15 days (Medicare) and 4 days (commercial databases 18-25 years), respectively. The median length of hospitalization was 6.5 days (IQR: 5-8 days) among 4 hospitalized cases in Medicare and 2 days (IQR: 1-2) among 9 hospitalized cases aged 18-25 years in commercial databases.
Conclusions: The claims-based algorithm used to identify myo/peri events post-vaccination performed well in the 18-25 age group. Among those enrolled in Medicare (≥65 years) and those aged 26-64 in Optum, the algorithm performed poorly, indicating a high degree of outcome misclassification. Younger individuals with myo/peri tended to have shorter time-to-event and hospitalization compared to older adults.
