Pharmacological prioritisation of signals of disproportionate reporting: proposal of an algorithm and pilot evaluation.

Abstract

PURPOSE:

Data mining in spontaneous reporting databases generates large numbers of signals of disproportionate reporting (SDRs) that need to be prioritised for assessment. The pharmacological relevance of drug-event associations is not considered in SDR prioritisation algorithms. This aimed to propose and test a pharmacological score for SDR prioritisation.

METHODS:

The Pharmacological Score for SDRs Prioritisation (PS-SP) was developed using a Delphi approach. An expert group agreed that PS-SP should include general criteria concerning SDRs and criteria concerning pharmacological relevance, and that criteria should be weighted for their risk representation. Once defined, the PS-SP was tested for prioritisation of SDRs for extrapyramidal syndrome in the French Pharmacovigilance database; the SDR classification was compared to that obtained using a traditional disproportionality approach.

RESULTS:

For a given drug, the general criteria retained were the reporting rate of the adverse drug reaction (ADR) and value of the 95% confidence interval (CI) lower boundary of the Reporting Odds Ratio (ROR). Pharmacological criteria consisted of the ADR reporting rate without concomitant at-risk drugs or those indicated for ADR treatment, and the value of the ROR 95% CI lower boundary as estimated in the subset of reports concerning drugs from the same therapeutic and then pharmacological class. Compared with traditional disproportionality, PS-SP prioritised specific drugs within congeners: metoclopramide, indoramin, and trimetazidine appeared as outliers within their classes; conventional antipsychotics had higher prioritisation than atypical antipsychotics.

CONCLUSION:

The pilot evaluation of PS-SP performed in extrapyramidal syndrome advocates for the use of pharmacological criteria in SDR prioritisation algorithms.

• Authors: Francesco Salvo, Emanuel Raschi, Ugo Moretti, Anita Chiarolanza, Annie Fourrier-Réglat, Nicholas Moore, Miriam Sturkemboom, Fabrizio De Ponti, Elisabetta Poluzzi, Antoine Pariente
• Journal: European journal of clinical pharmacology (Eur J Clin Pharmacol) Vol. 70 Issue 5 Pg. 617-25 (May 2014) ISSN: 1432-1041 [Electronic] Germany
• PMID: 24595599 (Publication Type: Journal Article)

Topics

• Adverse Drug Reaction Reporting Systems (organization & administration, statistics & numerical data)
• Algorithms
• Data Mining (methods, statistics & numerical data)
• Databases, Factual
• Drug-Related Side Effects and Adverse Reactions (epidemiology)
• Humans
• Medical Informatics Computing
• Odds Ratio
• Pilot Projects