A safety signal is information on a new or known adverse event that is potentially caused by a medicine and that requires further investigation. 2
Signals can be obtained from a number of sources such as spontaneous reports, non clinical data, interventional and non-interventional studies, systematic reviews and meta-analyses, clinical studies, quality data and scientific literature.3
Signal detection refers to the process of identifying signals in the context of uncertainty.
The stages of this are input, detection, output and decision. 1 Various statistical analysis methods have been produced and applied to scan large scale adverse event (AE) databases to identify disproportionality in the reporting of specific product-AE combinations against marginal distributions in the background. 4
Signal detection is now increasingly based on monitoring of large databases such as the EudraVigilance database. MAHs should regularly screen literature, internet and digital media for potential reports of suspected ADRs, which may characterise a new signal.
The detection of signals shall be based on a multidisciplinary approach. Signal detection within the large databases such as the EudraVigilance database shall be complemented by statistical analysis where appropriate.
Qualitative methods refer to clinical diagnosis, causality oriented, established knowledge and a small number of reports. The Qualitative methods (data mining) involve statistical analysis, coded data and large databases (Eudravigilance). It identifies statistics of disproportionate reporting (higher volume than expected). Both methods should complement signal detection.1
Results do not mean that there is a signal and usually requires further investigation. The event could be a symptom of an underlying illness or from taking another medicine concomitantly. The evaluation of safety signals is required to establish whether or not there is a causal relationship between the medicine and the reported adverse event.
Different regions have different methods of Quantitative analysis such as Proportional Reporting Ratio (EMA), Reporting Odd Ratio (The Netherlands), Bayesian Cofidence Propagation Neural Network (WHO) and Multi-item Gamma Poisson Shrinker (FDA).
Proportional Reporting Ratio is a measure of disproportionality of reporting and makes the assumption that a signal of disproportionate reporting involving a particular AE is identified for a product, this AE is reported relatively more frequently in association with this medicinal product than with other medicinal products. 5
The Pharmacovigilance Risk Assessment Committee (PRAC) is involved in the investigation, prioritisation and determination of validated signals, resulting in a recommendation that covers all medicines with a MA in the EU. Where a relationship is considered to be likely, the MAH will be asked to provide more information.
Regulatory action may be necessary and usually is an update of the package leaflet and summary of product characteristics (SmPC).6
Signal detection is now increasingly based on monitoring of large databases such as the EudraVigilance database. MAHs should regularly screen literature, internet and digital media for potential reports of suspected ADRs, which may characterise a new signal
...(download the rest of the essay above)