Abstract
The present study proposed a new statistical measure to find out the association between Drug-Adverse Drug Reaction (ADR) pair in the Pharmacovigilance system. The study proposed a subjective Bayesian measure corresponding to Proportional Reporting Ratio (PRR) parameter for quantitative signal detection. Classical and Bayesian inference procedure were used in analysis with four Drug-ADR pair as an example. We made prior information by expert opinion. The result of this analysis shows that Bayesian inference is more reliable as compared to classical inference. This study suggests that in case of spontaneous reporting in Pharmacovigilance system subjective Bayesian inference better to applying in case of small sample size.
Highlights
The goal of Pharmacovigilance is to identify such drug which is responsible for cause of adverse drug reaction and ensuring patient safety by using drugs
Emperical Bayes Geometric Mean (EBGM) and IC are the Bayesian measure of signal detection while Proportional Reporting Ratio (PRR) and Relative Odds Ratio (ROR) are the classical measure of signal detection
EBGM and IC are the Bayesian version of Relative Reporting Ratio (RRR), but for PRR and ROR, There is no Bayesian measure exist in literature
Summary
The goal of Pharmacovigilance is to identify such drug which is responsible for cause of adverse drug reaction and ensuring patient safety by using drugs. The association between certain drug and ADR [1] is identified by certain statistical tools This identification for certain drug-ADR pair is known as Quantitative Signal detection. It refers to computational or statistical methods used to identify drugevent pairs (higher-order combinations of drugs and events) that occur with disproportionately high frequency in large safety databases. The reason behind using such statistical methods that large size of dataset makes individual case by case assessment impossible when entering cases into the database For this purpose US-FDA using a tool Emperical Bayes Geometric Mean (EBGM). EBGM and IC are the Bayesian version of Relative Reporting Ratio (RRR), but for PRR and ROR, There is no Bayesian measure exist in literature. The main concern of this study to develop a Bayesian measure for PRR, which will computable and use the advantage of Bayesian theory
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
