Journal of Public Health and Preventive Medicine 2024;35(5):15-19

doi:10.3969/j.issn.1006-2483.2024.05.004

Identification of vaccine safety signals based on Apriori and GPS in Gansu Province in 2021 -2022

Xuefeng LIANG 1 ; Jing AN 1 ; Shuyu LIU 1 ; Yongzhuo JIAO 1 ; Na JIN 1 ; Xiaoshu ZHANG 1

Affiliations

+expand

Keywords

Pharmacovigilance; Signal identification; Empirical Bayes Gamma-Possion Shrinker; Apriori algorithm

Country

China

Language

Chinese

Abstract

Objective To identify signals of adverse events associated with vaccines in electronic medical record data. Methods Data cleaning was performed on electronic medical records and vaccination data in Gansu Province from 2020 to 2022. Signal identification was performed using Apriori algorithm and Empirical Bayes Gamma-Poisson Shrinker (GPS) with stratification by gender and age group. Results A total of 33 statistical signals were identified, with 7 signals identified by the Apriori method and 26 signals identified by the GPS method. No overlapping signals were found between the two methods at two different signal thresholds. Conclusion The vaccine safety signals identified by GPS and Apriori in electronic medical records are reliable and accurate, but further research is needed to determine the correlation between the vaccine safety signals recognized by GPS algorithms.