MedDRA and pharmacovigilance: a complex and little-evaluated tool.

MedDRA (Medical Dictionary tor Regulatory Activities) is a standardised medical terminology, published by the International Council for Harmonisation, used in particular for coding cases of adverse effects in clinical study reports and pharmacovigilance databases, and to facilitate searches in these databases. MedDRA has a 5-level, hierarchical structure. Some levels are used to code adverse effects, and the others are designed to group together related terms for database searches. The links between these terms are hierarchical and complex. In addition to searches for terms, broad or narrow searches of MedDRA-coded databases can be performed using predefined "standardised Med-DRA queries", of which about a hundred are available in early 2016. A French study on 4 adverse effects recorded in France's national pharma- covigilance database showed that the effectiveness of MedDRA and stan- dardised MedDRA queries is highly variable, depending on the adverse effect examined. In some cases, less than half of the cases retrieved were relevant. A study on the adverse effects most frequently reported in clinical trials of 10 randomly selected drugs showed that a given adverse effect could be described using between 4 and several hundred different codes. One risk of fragmenting the reporting of adverse effects across multiple terms is that their true frequency will be underestimated. The development of a harmonised international terminology is a worth-while endeavour. But the complexity of MedDRA and its poorly evaluated performance make it susceptible to manipulation, errors of interpretation and bias. This applies in particular to the adverse effect statistics in the clinical study reports of clinical trials. The evaluation of the effectiveness of MedDRA as a tool for identifying and quantifying the adverse effects of new drugs must continue. For healthcare professionals, the limitations of MedDRA are one more reason to recognise that the known adverse effects of drugs, and their quantification, tend to be underestimated.