Adverse event classification and signal detection of data from the customer service and pharmacovigilance of a multinational veterinary pharmaceutical company

Abstract

The pharmacovigilance of a veterinary company may include the analysis of spontaneous reports of adverse events (AE) related to its products. The present study developed an AE classification flowchart to analyze AE notified to the customer service and pharmacovigilance department of a multinational veterinary pharmaceutical company in Brazil. The product-AE binomials using the flowchart were characterized in terms of their frequencies and subsequently, three signal detection models were used: Reporting Odds Ratio, Bayesian confidence propagation neural network, and Gamma Poisson Shrinker. The signals detected with the three methods were classified according to their intensity, always with the most intense signal in the first position. Among the signals detected by the three methods, the positions of each signal were summed to obtain an aggregated classification that considered the results of the three methods and allowed a serial interpretation. Among the 531 reports, 20 types of AE; 88 product-AE binomials were identified. From the total of reports, seven were signs identified by the three methods. The classification of AE following explicit criteria and the combined use of more than one signal detection method enhances spontaneous-reports-based pharmacovigilance.

Keywords: Adverse Event; Notifications; Signal Detection; Spontaneous Reports; Veterinary Pharmacovigilance.