The Masking Phenomenon: Evaluating COVID-19 Vaccine Reports and Their Impact on AE Reporting in KAERS

Background: The masking effect refers to concealment of signal due to disproportionately high reporting rates of another drug or adverse event (AE), complicating accurate detection. With widespread distribution of Coronavirus disease 2019 (COVID-19) vaccines, surge in AE reports could pose challenges to signal detection in spontaneous reporting systems.

Objectives: To investigate masking effect of specific AEs due to increase in COVID-19 vaccine reports in Korea Adverse Event Reporting System database (KAERS DB).

Methods: Using the KAERS DB(2307A0006), we arbitrarily formed two datasets from January 1, 2018, to December 31, 2022, one with all vaccine reports including COVID-19 vaccine reports and one without. First, in each dataset, we examined Relative Reporting Frequency (RRF) trends for each AE by calculating annual RRF percentage change. RRF was its reports for specific AE and period divided by total reports in same period. Percentage change was computed by subtracting initial value from final one, dividing this difference by initial value, and multiplying by 100. To focus on most reported AEs, we selected those with Medical Dictionary for Regulatory Activities (MedDRA) code representing at least 82 reports (top 5%) over five years. Second, identifying period with highest RRF change, we applied Proportional Reporting Ratio (PRR) and Masking Ratio (MR) to detect and quantify masking effects in AEs with positive percentage changes. A signal meets criteria if PRR is ≥ 2, Chi-square is ≥ 4, and ≥ 3 cases. MR was PRR of drug-AE pair excluding COVID-19 vaccine reports divided by PRR with all reports, considered as masking effect when MR > 1.

Results: Out of total of 854,865 reports within entire study period, 95.6% (N=817,283) were related to COVID-19 vaccines. Between 2018 and 2020, annual changes in RRF values on 121 AEs remained stable, ranging between 69.2% and 78.7%. However, this figure spiked to 335.2% from 2020 to 2021. Excluding reports on COVID-19 vaccines, RRF reverted to 72.6%. A total of 16 signals were identified between 2020 and 2021 that met all indicators and show positive percentage change. Among 7 AEs with percentage change in RRF of 200% or greater, chest pain showed highest average MR at 27.7, followed by purpura thrombocytopenic (3.8) and diarrhoea (3.8).

Conclusions: The study shows that starting from February 2021, with the initiation of COVID-19 vaccine distribution in Korea, there was a marked increase in number of reported AEs in 2021, along with notable changes in reporting of specific AEs. This suggests surge in vaccine-related reports may have unintentionally masked reporting of other AEs in KAERS. Our findings emphasize the need for enhanced surveillance and analytical methodologies to detect signals through disproportionality analysis following mass vaccination campaigns.